DOC-8 review: say the field guide PDF is rebuilt, still v0.1.0

README and FIELD-GUIDE.md said the PDF was not rebuilt; it was, on 2026-10-17. Its content is still v0.1.0 and outdated. Rewrap the DOCS.md intro line. Text only. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DOC-8: update test_mqtt digest pins for the new manufacturer string
2026-10-17 03:27:35 +02:00 · 2026-10-17 02:16:34 +02:00 · 2026-10-17 01:53:33 +02:00 · 2026-10-17 01:09:42 +02:00 · 2026-10-16 22:31:09 +02:00 · 2026-10-16 21:48:05 +02:00
30 changed files with 1314 additions and 73 deletions
@@ -4,9 +4,9 @@ The guide technicians carry is the PDF:

 **`docs/GoodWe-RS485-Field-Guide.pdf`** — 10 pages, A4, print it double-sided.

-⚠️ **Outdated: it is v0.1.0.** Its `slew_w` row is corrected in `build_guide.py`
-and `field-guide.html`, but the PDF has not been rebuilt (see below). Where the
-guide disagrees with `goodwe_controller/DOCS.md`, DOCS.md wins.
+⚠️ **Outdated: the content is still v0.1.0.** The PDF was rebuilt on 2026-10-17
+for the EMS-port wording and the `slew_w` row only. Where the guide disagrees with
+`goodwe_controller/DOCS.md`, DOCS.md wins.

 It covers, in the order the work is done: pre-visit checks, hardware, firmware,
 add-on install and configuration, how to read the Web UI, the six commissioning
@@ -156,7 +156,7 @@ that looks hung gets deleted.
 **A GoodWe "-20" / G2 inverter cannot be driven by this.**

 Platform comes from characters 6–8 of the serial number. Only
-`ESU EMU ESA BPS BPU EMJ IJL` — platform 105, the AA55-era meter bus — are
+`ESU EMU ESA BPS BPU EMJ IJL` — platform 105, the AA55-era generation — are
 supported. SBP G2 (`SPB`/`SPN`) and ES G2 (`ESN`/`ESC`) are platform 745, a
 different protocol. The reference unit is `BPS`.

@@ -1,8 +1,8 @@
 # GoodWe RS485 Controller — Home Assistant add-on repository

 Replaces a GoodWe battery inverter's vendor controller with a local one: holds
-net grid exchange at zero by emulating the inverter's smart meter over RS485,
-and runs the monthly battery maintenance cycle the vendor box was doing.
+net grid exchange at zero by sending the inverter a grid reading over its RS485
+EMS port, and runs the monthly battery maintenance cycle the vendor box was doing.

 ```
 repository.yaml            add-on repository metadata
@@ -30,7 +30,7 @@ goodwe_controller/         the add-on
                           returns, run against its own release image
 firmware/goodwe-master.yaml  ESPHome config for the T-CAN485
 estop/rs485_log.py           optional RS485 e-stop / bus witness
-docs/GoodWe-RS485-Field-Guide.pdf  the guide techs carry (10pp, A4; outdated)
+docs/GoodWe-RS485-Field-Guide.pdf  the guide techs carry (10pp, A4; rebuilt 2026-10-17, content still v0.1.0 and outdated)
 docs/build_guide.py          its single source - rebuilds the HTML and PDF
 FIELD-GUIDE.md             how to edit and rebuild the guide
 docs/RELEASE.md            how a change reaches the battery - `master` is
@@ -118,17 +118,17 @@ the platform tag:

 | tag | platform | verdict |
 |---|---|---|
-| `ESU EMU ESA BPS BPU EMJ IJL` | 105 — ES/EM/BP, AA55-era meter bus | **supported** |
+| `ESU EMU ESA BPS BPU EMJ IJL` | 105 — ES/EM/BP, AA55-era | **supported** |
 | `SPB SPN` | 745 — SBP **G2** (`GW…-SBP-20`) | no |
 | `ESN ESC` | 745 — ES **G2** (`GW…-ES-20`) | no |
 | `ETU EHU BTU …` | 205/745/753 — ET/EH/BT hybrids | no |
 | `SDT DST MSU NSU …` | PV-only inverters — no battery to control | no |

-A "-20" or "G2" suffix means a different platform with a different meter
+A "-20" or "G2" suffix means a different platform with a different
 protocol. The tags come from the `goodwe` library's `PLATFORM_105_MODELS`, which
 is also what Home Assistant's own integration uses to pick a protocol.

- GoodWe **ES / BP family** inverters (AA55 / RS485 meter-bus generation)
+- GoodWe **ES / BP family** inverters (AA55 generation)
 - The protocol is **reverse-engineered**. There is no vendor contract, and a
  firmware change on GoodWe's side could break every installation at once. Say
  so when you sell it.
@@ -154,7 +154,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
    <table style="margin-bottom:2mm">
      <tr><th style="width:34mm">Tag</th><th>Platform</th><th style="width:30mm">Verdict</th></tr>
      <tr><td><code>ESU EMU ESA BPS<br>BPU EMJ IJL</code></td>
-          <td><b>Platform 105</b> — ES / EM / BP, AA55-era meter bus</td>
+          <td><b>Platform 105</b> — ES / EM / BP, AA55-era</td>
          <td><b>SUPPORTED</b></td></tr>
      <tr><td><code>SPB SPN</code></td><td>SBP <b>G2</b> (GW…-SBP-20) — platform 745</td><td><b>NO</b></td></tr>
      <tr><td><code>ESN ESC</code></td><td>ES <b>G2</b> (GW…-ES-20) — platform 745</td><td><b>NO</b></td></tr>
@@ -162,7 +162,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
      <tr><td><code>SDT DST MSU NSU …</code></td><td>PV-only inverters — <b>no battery to control</b></td><td><b>NO</b></td></tr>
    </table>
    <p style="margin:0"><small>Reference unit: <code>95000BPS225W0290</code> → <code>BPS</code> → supported.
-    A “-20” or “G2” suffix in the model name means a different platform with a different meter
+    A “-20” or “G2” suffix in the model name means a different platform with a different
    protocol. <b>If the tag is not on the supported list, stop — do not quote the job.</b></small></p>
  </div>

@@ -208,13 +208,13 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
 <!-- ============ PAGE 3 ============ -->
 <div class="page" id="p3">
  <h2>Steps 1–3 — Hardware</h2>
-  <div class="warn"><b>Qualified persons only.</b> Mains AC and battery DC are present. The meter bus
+  <div class="warn"><b>Qualified persons only.</b> Mains AC and battery DC are present. The RS485 EMS port
  is low-voltage but sits inside an energised installation.</div>

  <div class="step">
    <h3><span class="num">1</span> Survey and record</h3>
    <ol class="do">
-      <li><b>Photograph</b> the meter-port wiring before touching it.</li>
+      <li><b>Photograph</b> the EMS-port wiring before touching it.</li>
      <li>Identify the RS485 pair (A/B) running to the vendor controller.</li>
      <li>Record inverter model + serial, battery capacity (Ah) and nominal voltage.</li>
    </ol>
@@ -226,7 +226,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
    <ol class="do">
      <li>Bring the system to <b>0 W</b> — battery neither charging nor discharging (check the
          inverter display).</li>
-      <li>Disconnect the vendor controller from the meter port.</li>
+      <li>Disconnect the vendor controller from the inverter's EMS port.</li>
      <li>Leave it physically mounted but unplugged — that keeps the install reversible and gives
          you a rollback story.</li>
    </ol>
@@ -237,7 +237,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
  <div class="step">
    <h3><span class="num">3</span> Fit and power the T-CAN485</h3>
    <ol class="do">
-      <li>Wire <b>A→A, B→B</b> to the meter port, plus its own supply.</li>
+      <li>Wire <b>A→A, B→B</b> to the EMS port, plus its own supply.</li>
      <li>If fitting the e-stop, connect its USB-RS485 adapter to the <b>same pair, in parallel</b>.</li>
      <li>Power the T-CAN485 <b>last</b>.</li>
    </ol>
@@ -119,7 +119,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
    <table style="margin-bottom:2mm">
      <tr><th style="width:34mm">Tag</th><th>Platform</th><th style="width:30mm">Verdict</th></tr>
      <tr><td><code>ESU EMU ESA BPS<br>BPU EMJ IJL</code></td>
-          <td><b>Platform 105</b> — ES / EM / BP, AA55-era meter bus</td>
+          <td><b>Platform 105</b> — ES / EM / BP, AA55-era</td>
          <td><b>SUPPORTED</b></td></tr>
      <tr><td><code>SPB SPN</code></td><td>SBP <b>G2</b> (GW…-SBP-20) — platform 745</td><td><b>NO</b></td></tr>
      <tr><td><code>ESN ESC</code></td><td>ES <b>G2</b> (GW…-ES-20) — platform 745</td><td><b>NO</b></td></tr>
@@ -127,7 +127,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
      <tr><td><code>SDT DST MSU NSU …</code></td><td>PV-only inverters — <b>no battery to control</b></td><td><b>NO</b></td></tr>
    </table>
    <p style="margin:0"><small>Reference unit: <code>95000BPS225W0290</code> → <code>BPS</code> → supported.
-    A “-20” or “G2” suffix in the model name means a different platform with a different meter
+    A “-20” or “G2” suffix in the model name means a different platform with a different
    protocol. <b>If the tag is not on the supported list, stop — do not quote the job.</b></small></p>
  </div>

@@ -173,13 +173,13 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
 <!-- ============ PAGE 3 ============ -->
 <div class="page" id="p3">
  <h2>Steps 1–3 — Hardware</h2>
-  <div class="warn"><b>Qualified persons only.</b> Mains AC and battery DC are present. The meter bus
+  <div class="warn"><b>Qualified persons only.</b> Mains AC and battery DC are present. The RS485 EMS port
  is low-voltage but sits inside an energised installation.</div>

  <div class="step">
    <h3><span class="num">1</span> Survey and record</h3>
    <ol class="do">
-      <li><b>Photograph</b> the meter-port wiring before touching it.</li>
+      <li><b>Photograph</b> the EMS-port wiring before touching it.</li>
      <li>Identify the RS485 pair (A/B) running to the vendor controller.</li>
      <li>Record inverter model + serial, battery capacity (Ah) and nominal voltage.</li>
    </ol>
@@ -191,7 +191,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
    <ol class="do">
      <li>Bring the system to <b>0 W</b> — battery neither charging nor discharging (check the
          inverter display).</li>
-      <li>Disconnect the vendor controller from the meter port.</li>
+      <li>Disconnect the vendor controller from the inverter's EMS port.</li>
      <li>Leave it physically mounted but unplugged — that keeps the install reversible and gives
          you a rollback story.</li>
    </ol>
@@ -202,7 +202,7 @@ figcaption { font-size:9pt; color:#5b6169; margin-top:1.2mm; }
  <div class="step">
    <h3><span class="num">3</span> Fit and power the T-CAN485</h3>
    <ol class="do">
-      <li>Wire <b>A→A, B→B</b> to the meter port, plus its own supply.</li>
+      <li>Wire <b>A→A, B→B</b> to the EMS port, plus its own supply.</li>
      <li>If fitting the e-stop, connect its USB-RS485 adapter to the <b>same pair, in parallel</b>.</li>
      <li>Power the T-CAN485 <b>last</b>.</li>
    </ol>
@@ -1,5 +1,209 @@
 # Changelog

+## 0.18.2
+
+**A patch bump: one fix to how the 15-minute demand forecast sensor is read, and no option changes.**
+No option was added, removed or renamed, and no default changed (86 options).
+**Existing installs need no option change:** options that start 0.18.1 also
+start 0.18.2. Only a forecast sensor with no power unit needs a look (below).
+The change below applies with one inverter or two.
+
+**A forecast sensor with no power unit.** If your 15-minute demand forecast
+sensor has no unit, or a unit that is not power (for example kWh), the add-on
+now ignores that forecast. The log warns once, and the Home and Settings pages
+say so while it lasts. Pick a sensor that reads in W or kW.
+
+Contents of this release, in merge order since 0.18.1 (`780c1f6`):
+
+| | |
+|---|---|
+| `09a5485` | The 15-minute demand forecast sensor is now read in its own unit, W or kW, and the old rule that took any value below 50 as kW is gone. With no unit, or a unit that is not power, the forecast is not used, and the log and the pages say so; the settings page no longer says a kW sensor is taken as watts ([[SAFETY-118]], #198) |
+
+### Known issues
+
+- A false "Battery is not keeping up" can still show when the peak guard pauses a maintenance charge ([[SAFETY-92]]).
+- A battery capped at a low limit with a perfectly flat reading can still trip the stale check ([[SAFETY-74]]).
+- A battery whose reading at rest is beyond about ±175 W is still not caught when it refuses ([[SAFETY-97]]).
+- The peak guard can still back off near the end of a 15-minute period when a meter reading is late, by less than 30 s ([[SAFETY-99]]).
+- Two inverters: the maintenance charge can switch on and off while the house is above the peak aim ([[SAFETY-88]]).
+- Two inverters: the "circulating" figure counts a battery's reading at rest ([[SAFETY-98]]).
+- Two inverters: a drain that starts during a big step in house use can send the other battery to full power, and about 1 kW goes to the grid for about 10 s ([[SAFETY-104]]).
+- Two inverters: when a battery comes back during a maintenance charge, the grid can step by several kW at once ([[SAFETY-101]]).
+- Two inverters: when a battery drops out during a solar surplus, the others can swing by up to 1.5 kW in a simulation ([[SAFETY-105]]).
+- Two inverters: after a drain ends, a change in house use can go unanswered for about 9 s, because some readings are old ([[SAFETY-106]]).
+- Two inverters: when one battery's level is missing, both batteries are held at 0 W, and for about 15 s the last command stays on ([[SAFETY-108]]).
+- Two inverters: if one battery's level stays missing for 12 h, the other battery's cycle can be marked failed, with a 7-day wait and a notice, though it is healthy. This comes from reading the code; it has not been seen on a site ([[MAINT-17]]).
+- When the 15-minute demand forecast sensor counts the battery's own charge, the maintenance charge runs at about half the power the peak limit leaves, so it takes about twice as long. Seen on a site: about 1.2 kW instead of 2.5 kW, and the peak limit was kept ([[SAFETY-111]]).
+- One inverter: after a link drop, the battery's own safety stop to 0 W can be taken as a refusal, and the site stays at 0 W for up to 15 min ([[SAFETY-102]]).
+- After a Home Assistant restart, control still waits until Home Assistant has finished starting, which can take several minutes, though the readings are already there ([[SAFETY-115]]).
+- While Home Assistant is not answering, the page can list inverters as refusing commands and sensors as missing, and send you to Developer Tools, though they are fine ([[UX-37]]).
+- If Home Assistant hangs or stops, the page keeps saying "Waiting for Home Assistant" and never turns into an alarm ([[SAFETY-113]]).
+- Right after Home Assistant has started, the page can say for about 4 s that an inverter cannot be controlled, and the first command can come late ([[SAFETY-114]]).
+- After a Home Assistant restart, the controller status sensor can still say "running" while the add-on cannot control. It was seen once, and the cause is not known ([[SAFETY-116]]).
+- Battery modes on: after a planned top-up from the grid, the Batteries screen can say "Enough sun is forecast" ([[UX-32]]).
+- Battery modes on: the forecast line "we count on" can show more sun than the plan counts, because the plan takes house use off first ([[UX-33]]).
+- Battery modes on: before a planned top-up, the Today list shows a smaller "gives" amount than the plan will use; it grows as the charge happens ([[UX-34]]).
+- Two inverters, battery modes on: while the other battery is in its maintenance charge, a battery's card can say it charges from the grid while it stays at 0 W ([[OPT-18]]).
+- Two inverters, battery modes on: while a battery refuses to charge in a cheap run, the action log on the page gets about 8 lines an hour for its retries, and the page hint says "waiting for a better price" ([[UX-38]]).
+- A battery in a safety stop, or one the add-on cannot reach, can still be started for a cycle, and its drain then waits about 12 h and ends in a failed-cycle notice. This comes from reading the code; it has not been seen on a site ([[MAINT-19]]).
+- On the Settings page, the rows of the solar forecast lists are labelled "Phase 1", "Phase 2" instead of "Sensor 1", "Sensor 2" ([[UX-27]]).
+- In the solar forecast list, a blank or repeated row makes "source 2 of 2" count differently from the rows on the Settings page ([[UX-28]]).
+- The solar forecast message says "has no reading" also when the sensor shows a number that is too old or in the wrong unit ([[UX-29]]).
+- If you use only the forecast lists, two Settings notes sit on the single-sensor field, and one of them tells you to set that field ([[UX-30]]).
+- After a restart during a price outage, the "No prices" message can say "yet" when prices did come in earlier ([[UX-39]]).
+- On the new settings page, you cannot see whether a price service token is set. Check the Configuration tab ([[UX-41]]).
+- On the classic settings page, the first click on Save after editing a field can do nothing. Click again ([[UX-42]]).
+- The price chart on the Home page shows one bar per hour, the average of four 15-minute prices, while the add-on decides on the 15-minute prices ([[UX-45]]).
+
+### If you are crossing from a version older than 0.10.0
+
+`## 0.10.0` below can refuse an old config: `slew_w` above its limit, and the
+withdrawn `price_provider` ids `engie_dynamic`, `totalenergies_dynamic` and
+`eneco_eddy`. Read it before you update. Run `tools/predeploy_check.py` with the
+bare options (`jq '.data.options'`) first.
+
+## 0.18.1
+
+**A patch bump: two small fixes to the settings page, and no option changes.**
+No option was added, removed or renamed, and no default changed (86 options).
+**Existing installs need no option change:** options that start 0.18.0 also
+start 0.18.1. The changes below apply with one inverter or two.
+
+**Price rounding on 0.18.0.** On 0.18.0, the new settings page saved the fixed,
+day and night prices rounded to 2 decimals (0.34512 became 0.35); prices in time
+windows kept their decimals. If you typed more than 2 decimals in those
+fields on 0.18.0, enter them again.
+
+Contents of this release, in merge order since 0.18.0 (`a0e091c`):
+
+| | |
+|---|---|
+| `b703aaa` | The four tariff price fields on the new settings page (fixed, day, night and time windows) take 5 decimals, for example 0.34512 €/kWh, and are saved as typed. A short hint says to divide the c€/kWh on your bill by 100 ([[UX-43]], #194) |
+| `8e02a09` | The dynamic contract choice on the new settings page now says "Dynamic, every 15 minutes" ("Dynamisch, per kwartier") instead of "hourly". The choice works as before, and one line in the docs no longer calls dynamic prices hourly ([[UX-44]], #195) |
+
+### Known issues
+
+- A false "Battery is not keeping up" can still show when the peak guard pauses a maintenance charge ([[SAFETY-92]]).
+- A battery capped at a low limit with a perfectly flat reading can still trip the stale check ([[SAFETY-74]]).
+- A battery whose reading at rest is beyond about ±175 W is still not caught when it refuses ([[SAFETY-97]]).
+- The peak guard can still back off near the end of a 15-minute period when a meter reading is late, by less than 30 s ([[SAFETY-99]]).
+- Two inverters: the maintenance charge can switch on and off while the house is above the peak aim ([[SAFETY-88]]).
+- Two inverters: the "circulating" figure counts a battery's reading at rest ([[SAFETY-98]]).
+- Two inverters: a drain that starts during a big step in house use can send the other battery to full power, and about 1 kW goes to the grid for about 10 s ([[SAFETY-104]]).
+- Two inverters: when a battery comes back during a maintenance charge, the grid can step by several kW at once ([[SAFETY-101]]).
+- Two inverters: when a battery drops out during a solar surplus, the others can swing by up to 1.5 kW in a simulation ([[SAFETY-105]]).
+- Two inverters: after a drain ends, a change in house use can go unanswered for about 9 s, because some readings are old ([[SAFETY-106]]).
+- Two inverters: when one battery's level is missing, both batteries are held at 0 W, and for about 15 s the last command stays on ([[SAFETY-108]]).
+- Two inverters: if one battery's level stays missing for 12 h, the other battery's cycle can be marked failed, with a 7-day wait and a notice, though it is healthy. This comes from reading the code; it has not been seen on a site ([[MAINT-17]]).
+- When the 15-minute demand forecast sensor counts the battery's own charge, the maintenance charge runs at about half the power the peak limit leaves, so it takes about twice as long. Seen on a site: about 1.2 kW instead of 2.5 kW, and the peak limit was kept ([[SAFETY-111]]).
+- One inverter: after a link drop, the battery's own safety stop to 0 W can be taken as a refusal, and the site stays at 0 W for up to 15 min ([[SAFETY-102]]).
+- After a Home Assistant restart, control still waits until Home Assistant has finished starting, which can take several minutes, though the readings are already there ([[SAFETY-115]]).
+- While Home Assistant is not answering, the page can list inverters as refusing commands and sensors as missing, and send you to Developer Tools, though they are fine ([[UX-37]]).
+- If Home Assistant hangs or stops, the page keeps saying "Waiting for Home Assistant" and never turns into an alarm ([[SAFETY-113]]).
+- Right after Home Assistant has started, the page can say for about 4 s that an inverter cannot be controlled, and the first command can come late ([[SAFETY-114]]).
+- After a Home Assistant restart, the controller status sensor can still say "running" while the add-on cannot control. It was seen once, and the cause is not known ([[SAFETY-116]]).
+- Battery modes on: after a planned top-up from the grid, the Batteries screen can say "Enough sun is forecast" ([[UX-32]]).
+- Battery modes on: the forecast line "we count on" can show more sun than the plan counts, because the plan takes house use off first ([[UX-33]]).
+- Battery modes on: before a planned top-up, the Today list shows a smaller "gives" amount than the plan will use; it grows as the charge happens ([[UX-34]]).
+- Two inverters, battery modes on: while the other battery is in its maintenance charge, a battery's card can say it charges from the grid while it stays at 0 W ([[OPT-18]]).
+- Two inverters, battery modes on: while a battery refuses to charge in a cheap run, the action log on the page gets about 8 lines an hour for its retries, and the page hint says "waiting for a better price" ([[UX-38]]).
+- A battery in a safety stop, or one the add-on cannot reach, can still be started for a cycle, and its drain then waits about 12 h and ends in a failed-cycle notice. This comes from reading the code; it has not been seen on a site ([[MAINT-19]]).
+- On the Settings page, the rows of the solar forecast lists are labelled "Phase 1", "Phase 2" instead of "Sensor 1", "Sensor 2" ([[UX-27]]).
+- In the solar forecast list, a blank or repeated row makes "source 2 of 2" count differently from the rows on the Settings page ([[UX-28]]).
+- The solar forecast message says "has no reading" also when the sensor shows a number that is too old or in the wrong unit ([[UX-29]]).
+- If you use only the forecast lists, two Settings notes sit on the single-sensor field, and one of them tells you to set that field ([[UX-30]]).
+- After a restart during a price outage, the "No prices" message can say "yet" when prices did come in earlier ([[UX-39]]).
+- On the new settings page, you cannot see whether a price service token is set. Check the Configuration tab ([[UX-41]]).
+- On the classic settings page, the first click on Save after editing a field can do nothing. Click again ([[UX-42]]).
+- If the 15-minute demand forecast sensor reads in kW, the settings page warns that its number is taken as watts and asks for a sensor in W. The add-on does turn kW into W, so a kW sensor works ([[SAFETY-118]]).
+- The price chart on the Home page shows one bar per hour, the average of four 15-minute prices, while the add-on decides on the 15-minute prices ([[UX-45]]).
+
+### If you are crossing from a version older than 0.10.0
+
+`## 0.10.0` below can refuse an old config: `slew_w` above its limit, and the
+withdrawn `price_provider` ids `engie_dynamic`, `totalenergies_dynamic` and
+`eneco_eddy`. Read it before you update. Run `tools/predeploy_check.py` with the
+bare options (`jq '.data.options'`) first.
+
+## 0.18.0
+
+**A minor bump: new behaviour, three new options and one new value for an
+existing option.** No option was removed or renamed, and no default of an
+existing option changed (83 options -> 86). **Existing installs need no option
+change:** options that start 0.17.0 also start 0.18.0, and the new options are
+off or empty until you use them. The changes below apply with one inverter or
+two, unless a line says "Two inverters".
+
+New options, and what they do if you leave them alone:
+
+| option | default | left alone |
+|---|---|---|
+| `pv_forecast_entities`, `pv_forecast_tomorrow_entities` | `[]` | the single forecast sensors work as before ([[OPT-17]]) |
+| `price_service_token` | empty | not used; it is read only with `price_provider: price_service` ([[PRICE-01]]) |
+
+`price_provider` also accepts a new value, `price_service`. Its default is
+unchanged, so the price source stays as it is.
+
+**Price service.** The new price source `price_service` reads prices from the
+add-on's price server, with the token you were given. If it does not answer,
+prices come from energy-charts.info, then from the prices the add-on saved.
+Sites on the Home Assistant price sensors or a fixed tariff need no change. To
+move a site over, see `DOCS.md`, "Moving a site from the Home Assistant price
+sensors to `price_service`".
+
+Contents of this release, in merge order since 0.17.0 (`e36565f`):
+
+| | |
+|---|---|
+| `3050fa5` | The solar forecast can come from several sensors, added up, for example one per inverter. If one of them has no reading, that day's forecast is not used, and the status says which one. The single sensor still works ([[OPT-17]]) |
+| `e3e4ba6` | While Home Assistant is still starting, the page now says "Waiting for Home Assistant" and the log says the add-on waits, instead of looking like a fault. Control itself is unchanged: it starts when Home Assistant answers ([[SAFETY-112]]) |
+| `79dd077` | Battery modes on only: a charge in a cheap run no longer stops at the start of a new 15-minute period, and with two inverters a battery that refused to charge is tried again every 15 min. In one test, the export at the end of a charge was 48 W where the code before this fix gave 746 W ([[OPT-15]]) |
+| `0ce7d35` | Two inverters: each battery card has Start cycle and Stop cycle buttons, so you can start or stop the cycle on the battery you choose. Only one battery can cycle at a time, and a second start is refused with a short message ([[MAINT-11]]) |
+| `bf9019e` | New price source `price_service`. If it fails, prices come from energy-charts.info, then from saved prices. With no price at all and the monthly cycle on, a notice shows and the cycle waits ([[PRICE-01]]) |
+| `004f859` | Two inverters, battery modes on: the 15-minute retry of a battery that refused to charge now asks that battery for 200 W, and no longer puts the rest of the charge on another battery. In the test, without this fix, a battery in its maintenance charge took about 2,300 W for about 34 s at each retry ([[OPT-20]]) |
+
+### Known issues
+
+- A false "Battery is not keeping up" can still show when the peak guard pauses a maintenance charge ([[SAFETY-92]]).
+- A battery capped at a low limit with a perfectly flat reading can still trip the stale check ([[SAFETY-74]]).
+- A battery whose reading at rest is beyond about ±175 W is still not caught when it refuses ([[SAFETY-97]]).
+- The peak guard can still back off near the end of a 15-minute period when a meter reading is late, by less than 30 s ([[SAFETY-99]]).
+- Two inverters: the maintenance charge can switch on and off while the house is above the peak aim ([[SAFETY-88]]).
+- Two inverters: the "circulating" figure counts a battery's reading at rest ([[SAFETY-98]]).
+- Two inverters: a drain that starts during a big step in house use can send the other battery to full power, and about 1 kW goes to the grid for about 10 s ([[SAFETY-104]]).
+- Two inverters: when a battery comes back during a maintenance charge, the grid can step by several kW at once ([[SAFETY-101]]).
+- Two inverters: when a battery drops out during a solar surplus, the others can swing by up to 1.5 kW in a simulation ([[SAFETY-105]]).
+- Two inverters: after a drain ends, a change in house use can go unanswered for about 9 s, because some readings are old ([[SAFETY-106]]).
+- Two inverters: when one battery's level is missing, both batteries are held at 0 W, and for about 15 s the last command stays on ([[SAFETY-108]]).
+- Two inverters: if one battery's level stays missing for 12 h, the other battery's cycle can be marked failed, with a 7-day wait and a notice, though it is healthy. This comes from reading the code; it has not been seen on a site ([[MAINT-17]]).
+- When the 15-minute demand forecast sensor counts the battery's own charge, the maintenance charge runs at about half the power the peak limit leaves, so it takes about twice as long. Seen on a site: about 1.2 kW instead of 2.5 kW, and the peak limit was kept ([[SAFETY-111]]).
+- One inverter: after a link drop, the battery's own safety stop to 0 W can be taken as a refusal, and the site stays at 0 W for up to 15 min ([[SAFETY-102]]).
+- After a Home Assistant restart, control still waits until Home Assistant has finished starting, which can take several minutes, though the readings are already there ([[SAFETY-115]]).
+- While Home Assistant is not answering, the page can list inverters as refusing commands and sensors as missing, and send you to Developer Tools, though they are fine ([[UX-37]]).
+- If Home Assistant hangs or stops, the page keeps saying "Waiting for Home Assistant" and never turns into an alarm ([[SAFETY-113]]).
+- Right after Home Assistant has started, the page can say for about 4 s that an inverter cannot be controlled, and the first command can come late ([[SAFETY-114]]).
+- After a Home Assistant restart, the controller status sensor can still say "running" while the add-on cannot control. It was seen once, and the cause is not known ([[SAFETY-116]]).
+- Battery modes on: after a planned top-up from the grid, the Batteries screen can say "Enough sun is forecast" ([[UX-32]]).
+- Battery modes on: the forecast line "we count on" can show more sun than the plan counts, because the plan takes house use off first ([[UX-33]]).
+- Battery modes on: before a planned top-up, the Today list shows a smaller "gives" amount than the plan will use; it grows as the charge happens ([[UX-34]]).
+- Two inverters, battery modes on: while the other battery is in its maintenance charge, a battery's card can say it charges from the grid while it stays at 0 W ([[OPT-18]]).
+- Two inverters, battery modes on: while a battery refuses to charge in a cheap run, the action log on the page gets about 8 lines an hour for its retries, and the page hint says "waiting for a better price" ([[UX-38]]).
+- A battery in a safety stop, or one the add-on cannot reach, can still be started for a cycle, and its drain then waits about 12 h and ends in a failed-cycle notice. This comes from reading the code; it has not been seen on a site ([[MAINT-19]]).
+- On the Settings page, the rows of the solar forecast lists are labelled "Phase 1", "Phase 2" instead of "Sensor 1", "Sensor 2" ([[UX-27]]).
+- In the solar forecast list, a blank or repeated row makes "source 2 of 2" count differently from the rows on the Settings page ([[UX-28]]).
+- The solar forecast message says "has no reading" also when the sensor shows a number that is too old or in the wrong unit ([[UX-29]]).
+- If you use only the forecast lists, two Settings notes sit on the single-sensor field, and one of them tells you to set that field ([[UX-30]]).
+- After a restart during a price outage, the "No prices" message can say "yet" when prices did come in earlier ([[UX-39]]).
+- On the new settings page, you cannot see whether a price service token is set. Check the Configuration tab ([[UX-41]]).
+- On the classic settings page, the first click on Save after editing a field can do nothing. Click again ([[UX-42]]).
+
+### If you are crossing from a version older than 0.10.0
+
+`## 0.10.0` below can refuse an old config: `slew_w` above its limit, and the
+withdrawn `price_provider` ids `engie_dynamic`, `totalenergies_dynamic` and
+`eneco_eddy`. Read it before you update. Run `tools/predeploy_check.py` with the
+bare options (`jq '.data.options'`) first.
+
 ## 0.17.0

 **A minor bump: changed behaviour, and no option changes.** No option was
@@ -132,6 +132,23 @@ P1 is its section 6 P1, and the pull request lists what was measured.
    own, and a rise past that is the peak floor's, which ends the hold.
    A maintenance drain, hold or charge that starts during the hold ends it
    on that tick (replayed at edge + 3 s).
+- OPT-20 (TEST-23 OPT-15 g7, `test_opt20.py`): a probe asks only PROBE_W.
+  The trade was sized at the probing battery's rating (2500 W), B's pin held
+  200 W of it, and the allocator put the other 2300 W on A, in its
+  maintenance charge phase (no mode range there, so nothing kept it out):
+  grid 3100 W for 34 s. Now the ask is PROBE_W for the probing battery plus
+  what the other chargers would have had without the probe (none: 200 W
+  alone), and it carries the pin only on the ticks the pin is filed (the
+  window's first and last cycle moved another battery's share by 200 W).
+  - Chosen: the probe sits on top of a headroom-sized charge, so beside a
+    normal charger its command is the same and the grid rises 200 W (about
+    7 W on the quarter) rather than that charger giving 200 W up. A probe
+    that the headroom or the bound cannot hold at 200 W is not filed.
+  - Unchanged, same on bf9019e: with a charge ask the other batteries cannot
+    give (one sign), so an idle-looking A that gives the house stops for the
+    window and the grid rises by what it gave. The trade line now says 200 W
+    for a probe alone; its words are UX-38's. The base P4 F4 path (a normal
+    trade on a battery in maintenance charge) is not touched.

 ## P4: the solar forecast and the day plan (2026-09-29)
 - Input: a Home Assistant forecast sensor (`pv_forecast_entity`, kWh left
@@ -0,0 +1,18 @@
+# SAFETY-118 - the peak forecast is read in its own unit
+
+**Status: for review.** Tests: `test_safety118.py`.
+
+`peak_forecast_entity` was read as a bare number, and anything under 50 was
+multiplied by 1000 ("accept kW or W"). A W sensor reading 17 to 45 W, which a
+real site did 430 times in one morning, would be taken as 17 to 45 kW. The
+validator also called a kW sensor wrong, which was false for this entity.
+
+Decision: the size rule is gone. The entity is read through `pv.watts()`, the
+same unit table and reader as the solar entities, so W, kW, MW and the rest
+convert by the sensor's own `unit_of_measurement`. No unit, or a unit that is
+not power (kWh), means the forecast is not used, exactly as if the option were
+empty: the loop cap and every non-forecast peak step still apply. One WARNING
+per cause, and a status row and a settings note say the same.
+
+No size rule can be right in both units: 30 W and 30 kW are both ordinary.
+The read is still one `GET /api/states/<id>` per cycle, as `hass.number()` made.
@@ -1,8 +1,9 @@
 # GoodWe RS485 Controller

-Drives a GoodWe ES/BP battery inverter over its RS485 meter bus: holds net grid
-exchange at zero, and runs a monthly battery maintenance cycle so the BMS can
-balance cells and recalibrate its coulomb counter.
+Drives a GoodWe ES/BP battery inverter over its RS485 EMS port: sends it a grid
+reading to hold net grid exchange at zero, and runs a monthly battery
+maintenance cycle so the BMS can balance cells and recalibrate its coulomb
+counter.

 Installers: the printed field guide (`docs/GoodWe-RS485-Field-Guide.pdf`) is
 v0.1.0 and outdated; where it disagrees with this page, this page wins.
@@ -33,8 +34,8 @@ v0.1.0 and outdated; where it disagrees with this page, this page wins.

 You need:

- A GoodWe **ES / BP family** inverter (AA55 / RS485 meter-bus generation)
- The vendor's meter-emulating controller **disconnected** from the bus
+- A GoodWe **ES / BP family** inverter (AA55 generation)
+- The vendor's controller **disconnected** from the inverter's EMS port
 - A control board: a T-CAN485 (ESP32) flashed with
  `firmware/goodwe-master.yaml`
 - A grid-power sensor already working in Home Assistant, updating every ~5–10 s
@@ -210,7 +211,7 @@ The label and description each setting shows on the Configuration tab, in tab or
 | Limits | Battery cut-out level (%) | `bms_soc_floor` | The level where the battery's own management system stops giving power. Read it off your battery rather than copying a figure. |
 | Limits | Peak limit (W) | `peak_cap_w` | Your own number: the 15-minute average power the add-on keeps the grid under. Take it from your grid contract. |
 | Limits | Margin below the peak limit (W) | `peak_margin_w` | The add-on aims this much below the peak limit, so a 15-minute period that ends a little high still stays under it. |
-| Limits | 15-minute demand forecast entity (W) | `peak_forecast_entity` | Optional. A forecast of this 15-minute period's average power. Leave it empty and there is no peak limit. |
+| Limits | 15-minute demand forecast entity (W) | `peak_forecast_entity` | Optional. A forecast of this 15-minute period's average power, in W or kW. With no power unit it is not used. Leave it empty and there is no peak limit. |
 | Limits | Peak reserve (%) | `peak_soc_floor` | Peak limiting will not draw the battery below this level, but ordinary use can. Derived value, leave it unchanged. |
 | Tariff | Dynamic price source | `price_provider` | Where prices come from. Choosing a source starts the price fetch, even with trading off. |
 | Tariff | Supplier markup (EUR/kWh) | `price_markup_eur_kwh` | Your supplier's margin. It is added only when the price source does not already include it. |
@@ -1697,7 +1698,7 @@ DECISIONS/MAINT-13-full-pack-is-complete.md.

 | option | meaning |
 |---|---|
-| `peak_forecast_entity` | Quarter-hour demand forecast, for capacity-tariff markets. Empty = no cap |
+| `peak_forecast_entity` | Quarter-hour demand forecast, for capacity-tariff markets. Read in its own unit (W, kW or any power unit). No unit, or a unit that is not power, means it is not used, as if empty (SAFETY-118). Empty = no cap |
 | `peak_cap_w` | The site's capacity-tariff target |
 | `peak_margin_w` (50) | Watts the peak guard aims under `peak_cap_w` (SAFETY-79). Reports and the bill use `peak_cap_w` itself |
 | `peak_soc_floor` (20) | **The percent below which the capacity-peak guard will not discharge the battery.** ⚠️ It bounds **the peak guard only** — it is **not a pack-wide reserve**, and ordinary self-consumption discharges straight past it (SAFETY-25, see below). Not optional and not the same thing as `maintenance_soc_floor` |
@@ -3113,15 +3114,16 @@ second set of 23- and 25-hour-day bugs.
 ### Static day/night and fully-fixed tariffs (`static_profile`)

 Most Belgian households are **not** on a dynamic contract, and an EMS that only
-works under hourly pricing is an EMS for a minority. `static_profile` is how a
+works under dynamic pricing is an EMS for a minority. `static_profile` is how a
 two-period day/night tariff or a single fixed price becomes a first-class
 configuration rather than "turn the price feature off and hope".

 On the add-on page, choose **Other time windows** under Settings, Tariff and
 prices, or in Setup step 4. Add one row per time window, with its start, end,
 days and price. The page names any overlap, gap or empty price before you can
-save. You only need the JSON below for days the page cannot choose, such as
-Monday and Wednesday only.
+save. Prices take five decimals: divide the c€/kWh on your bill by 100, so
+34.512 c€/kWh is 0.34512 €/kWh. You only need the JSON below for days the page
+cannot choose, such as Monday and Wednesday only.

 It is one option, holding a JSON object, on one line:

@@ -3,7 +3,7 @@
 WHAT THIS THING DOES, in one paragraph: it reads the house's net grid power from
 Home Assistant, decides how hard the battery should charge or discharge to hold
 that at zero, and writes that figure to an ESP32 which puts it on the inverter's
-RS485 meter bus. Once a month it runs a full low->full battery cycle so the BMS
+RS485 EMS port. Once a month it runs a full low->full battery cycle so the BMS
 can balance cells and recalibrate its coulomb counter.

 ⚠️ THE ONE THING TO UNDERSTAND BEFORE CHANGING ANY OF THIS: the inverter holds
@@ -1202,6 +1202,9 @@ class Controller:
        # `grid` is P1, one meter, one connection, and belongs here.
        self.grid = None
        self.peak_fc = None
+        # SAFETY-118: the unit fault last logged for the peak forecast (None =
+        # no fault), so the warning comes once and again only if it changes.
+        self._peak_fc_fault = None
        self.cheap = False
        self._price_gated = False     # MAINT-13: is a price gate set up
        # PRICE-01. Which step of the price chain the prices come from (logged
@@ -2441,6 +2444,19 @@ class Controller:
        tick = self.mode_tick
        if tick is None:
            return False
+        if tick.action == modetrade.CHARGE:
+            # OPT-20: the ask carries PROBE_W for exactly the probes whose pin
+            # is filed this tick. `decide()` ran before the pin was (or the
+            # window closed since), so the two can differ by one cycle; left
+            # alone, the allocator takes the pin's watts out of another
+            # battery's share, or hands them to one it can reach.
+            pins = _ctl.PROBE_W * sum(
+                1 for c in claims if c.source == "refused"
+                and c.kind == "set" and c.value < 0)
+            if pins != tick.probe_w:
+                tick = self.mode_tick = dataclasses.replace(
+                    tick, ask_w=tick.ask_w + tick.probe_w - pins,
+                    probe_w=pins)
        size = modetrade.size(tick, self.loop_bound(claims)
                              if tick.action else None, self.arb_headroom_w)
        runs = self.mode_runs
@@ -3345,6 +3361,22 @@ class Controller:
        self.pv_w, self.pv_why, self.pv_parts = pv_read.total(
            list(zip(roster, states)), pv_read.inverted(self.o))

+    def _peak_fc_unit_log(self, fault) -> None:
+        """SAFETY-118. One WARNING when the peak forecast's unit cannot be read
+        in watts, and again only if the unit changes; the unit is on DEBUG."""
+        if fault == self._peak_fc_fault:
+            return
+        self._peak_fc_fault = fault
+        if fault is None:
+            _LOG.info("The 15-minute demand forecast is used again.")
+        elif fault == "":
+            _LOG.warning("The 15-minute demand forecast has no unit, so it "
+                         "is not used.")
+        else:
+            _LOG.warning("The 15-minute demand forecast is not in a power "
+                         "unit, so it is not used.")
+            _LOG.debug("The 15-minute demand forecast reads in %r.", fault)
+
    async def _read_forecast(self) -> None:
        """OPT-11 P4: the solar forecast sensors (`solarfc.read()`). A
        failed or stale read is no forecast, never a zero."""
@@ -3435,11 +3467,17 @@ class Controller:
            _reads.append(self._read_forecast())
        await asyncio.gather(*_reads)
        if o.get("peak_forecast_entity"):
-            raw = await self.hass.number(o["peak_forecast_entity"])
-            # Accept kW or W - nobody's quarter-hour forecast is 50 W.
-            self.peak_fc = None if raw is None else (raw * 1000 if abs(raw) < 50 else raw)
+            # SAFETY-118: read in its own unit, like the PV path. One state
+            # read, as `hass.number()` made. No unit or a unit that is not
+            # power: not used, as if the option were empty.
+            st = await self.hass.state(o["peak_forecast_entity"])
+            self.peak_fc, _why = pv_read.watts(st)
+            # A missing reading leaves the unit-fault state as it was.
+            if pv_read.has_reading(st):
+                self._peak_fc_unit_log(pv_read.unit_fault(st))
        else:
            self.peak_fc = None
+            self._peak_fc_fault = None

        now_e, avg_e = o.get("price_now_entity"), o.get("price_avg_entity")
        if now_e and avg_e:
@@ -7227,6 +7265,10 @@ class Controller:
            # is both is a check nobody reads.
            out.append({"ok": False, "warn": True, "id": "peak_forecast",
                        "text": messages.row_no_peak_forecast()})
+        elif getattr(self, "_peak_fc_fault", None) is not None:
+            out.append({"ok": False, "warn": True, "id": "peak_forecast",
+                        "text": messages.row_peak_forecast_unit(
+                            self._peak_fc_fault)})
        # OPT-17: a forecast summed from several entities names, by position,
        # each member with no usable reading. A caution: control goes on.
        gaps = getattr(self, "solar_fc_gaps", ())
@@ -788,6 +788,25 @@ def row_no_peak_forecast() -> str:
            "the peak limit")


+def peak_forecast_unit(unit, lang="en") -> str:
+    """SAFETY-118. The 15-minute demand forecast has a number but no power
+    unit, so it is not used. `unit` is "" for none. One sentence, for the log
+    line, the status row and the settings page. Log lines stay English."""
+    if lang == "nl":
+        return ("De kwartierprognose heeft geen eenheid, dus ze wordt niet "
+                "gebruikt." if not unit else
+                "De kwartierprognose heeft geen vermogenseenheid, dus ze "
+                "wordt niet gebruikt.")
+    return ("The 15-minute demand forecast has no unit, so it is not used."
+            if not unit else
+            "The 15-minute demand forecast is not in a power unit, so it is "
+            "not used.")
+
+
+def row_peak_forecast_unit(unit) -> str:
+    return peak_forecast_unit(unit)
+
+
 def row_peak_guard_waiting() -> str:
    return "Peak guard: no decision yet, control has not run"

@@ -1229,6 +1248,8 @@ def status_samples() -> list:
        ("row_no_estop", AMBER, row_no_estop()),
        ("row_estop_fitted", GREEN, row_estop_fitted()),
        ("row_no_peak_forecast", AMBER, row_no_peak_forecast()),
+        ("row_peak_forecast_unit", AMBER, row_peak_forecast_unit("")),
+        ("row_peak_forecast_unit", AMBER, row_peak_forecast_unit("kWh")),
        ("row_peak_guard_waiting", AMBER, row_peak_guard_waiting()),
        ("row_pv", GREEN, row_pv(3200, 2)),
        ("row_pv", GREEN, row_pv(1800, 1, True)),
@@ -47,12 +47,17 @@ Rules that keep this safe (tests: test_modes_p3.py, test_price_run_fixes.py):
    is its SAFETY-73 probe inside a cheap run (OPT-15): it is a charger then,
    pinned at PROBE_W by its own claim, so the others rest (one sign) and the
    card's "tries again every 15 minutes" holds in the run too;
+  * OPT-20: a probe asks PROBE_W for the probing battery and nothing more.
+    The rest of the ask is what the other chargers would have had without
+    the probe (none: PROBE_W alone), so a probe never places charge on a
+    battery in a maintenance phase or on one that rests;
  * no grid charge while another battery drains for its monthly cycle
    (`draining`, SAFETY-78): it would charge against the drain.
 """

 from dataclasses import dataclass

+from .control import PROBE_W
 from .packmode import (CHEAP_HOURS, GRID_CHARGE_MODES, LEFTOVER_SOLAR,
                       PLAN_CHARGE_MODES, SOLAR_FIRST)

@@ -88,6 +93,7 @@ class Tick:
    action: str | None          # "charge" | "sell" | None
    ask_w: float                # signed site ask before sizing (0 if none)
    why: str | None             # why no trade when a run is on, else None
+    probe_w: float = 0.0        # OPT-20: the part of `ask_w` that is a probe


 def _room_kwh(p: Pack, pct: float) -> float:
@@ -232,8 +238,12 @@ def decide(runs, now, packs, *, charge_ok, sell_ok, trade_why, price_now,
            ranges[p.name] = (lo, hi)

    if charging:
-        ask = -sum(abs(p.rating_w) for p in ch)
-        return Tick(ranges, roles, CHARGE, ask, None)
+        # OPT-20: a probing refuser is asked PROBE_W (its own pin holds it
+        # there); only the batteries that charge without a probe are asked
+        # for their rating.
+        probe_w = PROBE_W * sum(1 for p in ch if p.refused < 0)
+        ask = -sum(abs(p.rating_w) for p in ch if p.refused >= 0) - probe_w
+        return Tick(ranges, roles, CHARGE, ask, None, probe_w)
    if selling:
        ask = sum(abs(p.rating_w) for p in se) + max(0.0, float(house_w))
        return Tick(ranges, roles, SELL, ask, None)
@@ -250,8 +260,20 @@ def size(tick: Tick, bound, headroom_w) -> float:
        return 0.0
    lo, hi = bound
    if tick.action == CHARGE:
-        cap = min(-float(tick.ask_w), -float(lo),
-                  float(headroom_w) if headroom_w is not None else 0.0)
+        room = float(headroom_w) if headroom_w is not None else 0.0
+        pw = float(tick.probe_w)
+        if pw > 0.0:
+            # OPT-20: the others' share is sized as without the probe; the
+            # probe's PROBE_W is added on top (the grid rises by PROBE_W at
+            # most), and alone it is sized by the bound and the headroom.
+            rest = max(0.0, min(-float(tick.ask_w) - pw, -float(lo) - pw,
+                                room))
+            if rest > 0.0:
+                return -(rest + pw)
+            # PROBE_W exactly or no trade: less would void the pin and hand
+            # the remainder to whichever battery the allocator can reach.
+            return -pw if min(-float(lo), room) >= pw else 0.0
+        cap = min(-float(tick.ask_w), -float(lo), room)
        return -max(0.0, cap)
    return max(0.0, min(float(tick.ask_w), float(hi)))

@@ -29,7 +29,7 @@ _LOG = logging.getLogger("goodwe.mqtt")
 DEVICE = {
    "identifiers": ["goodwe_rs485_controller"],
    "name": "GoodWe",
-    "manufacturer": "GoodWe (via RS485 meter emulation)",
+    "manufacturer": "GoodWe (via RS485 EMS port)",
    "model": "ES/BP series",
 }

@@ -93,12 +93,12 @@ def watts(state: dict | None) -> tuple:
    unknown" are different claims and the page has to be able to make the first.

    ⚠️ THE UNIT IS READ OFF THE ENTITY AND NOT ASSUMED, AND THERE IS NO
-    MAGNITUDE HEURISTIC. `read_inputs()` guesses kW-or-W for
-    `peak_forecast_entity` with `abs(raw) < 50`, which is safe there because
-    nobody's quarter-hour demand forecast is 50 W. It is NOT safe here: 30 W of
-    real production at dusk is an ordinary reading, and that heuristic would
-    publish it as 30 kW - a fabricated number, in the channel whose entire
-    purpose is to be believed a week later. A sensor whose unit this cannot read
+    MAGNITUDE HEURISTIC. 30 W of real production at dusk is an ordinary reading,
+    and a "small number means kW" rule would publish it as 30 kW - a fabricated
+    number, in the channel whose entire purpose is to be believed a week later.
+    SAFETY-118 removed the last such rule, on `peak_forecast_entity`, for the
+    same reason (DECISIONS/SAFETY-118-peak-forecast-unit.md): the peak forecast
+    now reads through this function too. A sensor whose unit this cannot read
    is REFUSED rather than guessed at.
    """
    if state is None:
@@ -126,6 +126,31 @@ def watts(state: dict | None) -> tuple:
    return value * WATTS_PER[unit], unit


+def has_reading(state: dict | None) -> bool:
+    """SAFETY-118. True when `state` holds a number, so its unit can be judged.
+    A missing, unavailable, unknown or non-numeric state is a missing reading."""
+    if state is None:
+        return False
+    raw = str(state.get("state", "")).strip()
+    if raw.lower() in BAD_STATES:
+        return False
+    try:
+        float(raw)
+    except ValueError:
+        return False
+    return True
+
+
+def unit_fault(state: dict | None):
+    """SAFETY-118. `None` unless `state` holds a number whose unit cannot be
+    read in watts, then that unit (`""` for none). `None` also means a missing
+    reading: use `has_reading()` to tell it from a good unit."""
+    if not has_reading(state):
+        return None
+    unit = str((state.get("attributes") or {}).get(UNIT_KEY) or "").strip()
+    return None if unit in WATTS_PER else unit
+
+
 # --- TEL-12: the ROSTER, and the SITE figure over it --------------------------

 def entities(opts: dict) -> tuple:
@@ -311,7 +311,7 @@
      const days = h("select", { class: "sel", "aria-label": say(P.days) });
      for (const [v, label] of f.days || []) days.append(h("option", { value: v, selected: v === row.days }, label));
      days.addEventListener("change", () => set("days", days.value));
-      const price = h("input", { class: "num", type: "number", inputmode: "decimal", step: "0.01",
+      const price = h("input", { class: "num", type: "number", inputmode: "decimal", step: String(f.step || 0.01),
                                 value: row.price == null ? "" : row.price, "aria-label": say(P.price) });
      price.addEventListener("input", () => set("price", price.value === "" || isNaN(+price.value) ? null : +price.value));
      const cell = (label, ...kids) => h("label", { class: "pf" }, h("span", { class: "tiny muted" }, label), ...kids);
@@ -23,13 +23,13 @@
    "contract": {"name": "Your contract", "description": "How your supplier prices each kWh you take from the grid."},
    "region": {"name": "Region", "description": "Fills in grid costs, levies and VAT for the year shown, but not yet for Wallonia or Brussels, so check your bill."},
    "fallback": {"name": "Price when the feed is down", "description": "Used when the price source does not answer."},
-    "fixed_price": {"name": "Price per kWh, all in", "description": "Energy, supplier costs, grid costs, levies and VAT together."},
-    "day_price": {"name": "Day price, weekdays 07:00 to 22:00", "description": "All in, as on your bill."},
-    "night_price": {"name": "Night and weekend price", "description": "All in, as on your bill."},
-    "periods": {"name": "Time windows and prices", "description": "One price per kWh for each time window, all in, as on your bill."}
+    "fixed_price": {"name": "Price per kWh, all in", "description": "Energy, supplier costs, grid costs, levies and VAT together: divide c€/kWh from the bill by 100, so 34.512 c€/kWh is 0.34512."},
+    "day_price": {"name": "Day price, weekdays 07:00 to 22:00", "description": "All in, as on your bill: divide c€/kWh by 100, so 34.512 c€/kWh is 0.34512."},
+    "night_price": {"name": "Night and weekend price", "description": "All in, as on your bill: divide c€/kWh by 100, so 34.512 c€/kWh is 0.34512."},
+    "periods": {"name": "Time windows and prices", "description": "One price per kWh for each time window, all in, as on your bill: divide c€/kWh by 100, so 34.512 c€/kWh is 0.34512."}
  },
  "choices": {
-    "contract": {"none": "No prices", "fixed": "Fixed price", "daynight": "Day and night", "dynamic": "Dynamic, hourly", "custom": "Other time windows"},
+    "contract": {"none": "No prices", "fixed": "Fixed price", "daynight": "Day and night", "dynamic": "Dynamic, every 15 minutes", "custom": "Other time windows"},
    "fallback": {"none": "None", "fixed": "One price", "daynight": "Day and night", "custom": "Other time windows"},
    "region": {"flanders": "Flanders, Fluvius Antwerpen", "own": "I enter my own figure"},
    "meter_source": {"off": "One grid power entity", "ha_dsmr": "Import and export entities", "mqtt_p1": "Meter over MQTT", "ha_signed": "One net power entity", "homewizard_local": "HomeWizard meter on the network"},
@@ -23,13 +23,13 @@
    "contract": {"name": "Uw contract", "description": "Hoe uw leverancier elke kWh prijst die u van het net afneemt."},
    "region": {"name": "Regio", "description": "Vult nettarieven, heffingen en btw in voor het getoonde jaar, nog niet voor Wallonië of Brussel, dus kijk uw factuur na."},
    "fallback": {"name": "Prijs als de feed wegvalt", "description": "Gebruikt als de prijsbron niet antwoordt."},
-    "fixed_price": {"name": "Prijs per kWh, alles inbegrepen", "description": "Energie, kosten van de leverancier, nettarieven, heffingen en btw samen."},
-    "day_price": {"name": "Dagprijs, weekdagen 07:00 tot 22:00", "description": "Alles inbegrepen, zoals op uw factuur."},
-    "night_price": {"name": "Nacht- en weekendprijs", "description": "Alles inbegrepen, zoals op uw factuur."},
-    "periods": {"name": "Tijdvakken en prijzen", "description": "Eén prijs per kWh voor elk tijdvak, alles inbegrepen, zoals op uw factuur."}
+    "fixed_price": {"name": "Prijs per kWh, alles inbegrepen", "description": "Energie, kosten van de leverancier, nettarieven, heffingen en btw samen: deel c€/kWh van de factuur door 100, dus 34,512 c€/kWh is 0,34512."},
+    "day_price": {"name": "Dagprijs, weekdagen 07:00 tot 22:00", "description": "Alles inbegrepen, zoals op uw factuur: deel c€/kWh door 100, dus 34,512 c€/kWh is 0,34512."},
+    "night_price": {"name": "Nacht- en weekendprijs", "description": "Alles inbegrepen, zoals op uw factuur: deel c€/kWh door 100, dus 34,512 c€/kWh is 0,34512."},
+    "periods": {"name": "Tijdvakken en prijzen", "description": "Eén prijs per kWh voor elk tijdvak, alles inbegrepen, zoals op uw factuur: deel c€/kWh door 100, dus 34,512 c€/kWh is 0,34512."}
  },
  "choices": {
-    "contract": {"none": "Geen prijzen", "fixed": "Vaste prijs", "daynight": "Dag en nacht", "dynamic": "Dynamisch, per uur", "custom": "Andere tijdvakken"},
+    "contract": {"none": "Geen prijzen", "fixed": "Vaste prijs", "daynight": "Dag en nacht", "dynamic": "Dynamisch, per kwartier", "custom": "Andere tijdvakken"},
    "fallback": {"none": "Geen", "fixed": "Eén prijs", "daynight": "Dag en nacht", "custom": "Andere tijdvakken"},
    "region": {"flanders": "Vlaanderen, Fluvius Antwerpen", "own": "Ik vul zelf een bedrag in"},
    "meter_source": {"off": "Eén entiteit voor netvermogen", "ha_dsmr": "Entiteiten voor afname en injectie", "mqtt_p1": "Meter via MQTT", "ha_signed": "Eén entiteit voor nettovermogen", "homewizard_local": "HomeWizard-meter op het netwerk"},
@@ -164,13 +164,13 @@ GROUPS = (
                  wizard=4),
            Field("price_export_formula", TEXT),
            Field("ui.fallback", SEGMENTED, show=_DYNAMIC, wizard=4),
-            Field("ui.fixed_price", NUMBER, "€/kWh", step=0.01,
+            Field("ui.fixed_price", NUMBER, "€/kWh", step=0.00001,
                  show=(("ui.fallback", "==", "fixed"),), wizard=4),
-            Field("ui.day_price", NUMBER, "€/kWh", step=0.01,
+            Field("ui.day_price", NUMBER, "€/kWh", step=0.00001,
                  show=(("ui.fallback", "==", "daynight"),), wizard=4),
-            Field("ui.night_price", NUMBER, "€/kWh", step=0.01,
+            Field("ui.night_price", NUMBER, "€/kWh", step=0.00001,
                  show=(("ui.fallback", "==", "daynight"),), wizard=4),
-            Field("ui.periods", PERIOD_LIST, "€/kWh", step=0.01,
+            Field("ui.periods", PERIOD_LIST, "€/kWh", step=0.00001,
                  show=(("ui.fallback", "==", "custom"),), wizard=4),
        )),
        Part("selling", (
@@ -194,7 +194,7 @@ GROUPS = (
            Field("peak_cap_w", SLIDER, "kW", scale=1000, step=100,
                  simple=True, wizard=5, note="month_peak"),
            Field("peak_margin_w", SLIDER, "W", step=10),   # SAFETY-79
-            Field("peak_forecast_entity", ENTITY, kind="power_w", wizard=5),
+            Field("peak_forecast_entity", ENTITY, kind="power", wizard=5),
            Field("peak_soc_floor", LOCKED, "%"),
        )),
        Part("start", (
@@ -1741,11 +1741,12 @@ def _card_note(option: str, spec) -> str:


 # UX-13: entity options whose number the add-on takes as watts, with no unit
-# conversion. pv.py converts each solar entity's own unit, so those are not here.
+# conversion. pv.py converts each solar entity's own unit, so those are not here,
+# and SAFETY-118 moved the peak forecast there too.
 WATT_ENTITY_KEYS = (
    "meter_entity", "p1_net_entity", "p1_import_entity", "p1_export_entity",
    "p1_phase_net_entities", "p1_phase_import_entities",
-    "p1_phase_export_entities", "batt_entity", "peak_forecast_entity")
+    "p1_phase_export_entities", "batt_entity")


 # The meter entities each `meter_source` reads through Home Assistant.
@@ -1829,6 +1830,10 @@ def _entity_notes(notes, opts: dict, entities: dict, bad=()) -> None:
            _advise(notes, "warn", (key,),
                    f"{entity} reads in {unit}, but {S(key, where)} is "
                    "taken as watts, so pick a sensor in W.")
+        if key == "peak_forecast_entity" and unit not in pv.WATTS_PER:
+            _advise(notes, "warn", (key,),
+                    messages.peak_forecast_unit(unit) + " Pick a sensor in "
+                    "W or kW.")
        if key in FORECAST_KEYS and unit not in solarfc.KWH_PER:
            _advise(notes, "warn", (key,),
                    f"{entity} does not read in kWh, so pick a solar "
@@ -1,9 +1,9 @@
 name: GoodWe RS485 Controller
-version: "0.17.0"
+version: "0.18.2"
 slug: goodwe_controller
 description: >-
-  Drives a GoodWe ES/BP battery inverter over RS485 by emulating its smart
-  meter, holding net grid exchange at zero and running a monthly battery
+  Drives a GoodWe ES/BP battery inverter over its RS485 EMS port by sending it a
+  grid reading, holding net grid exchange at zero and running a monthly battery
  maintenance cycle.
 url: https://gittea.kammenstraatha.duckdns.org/admin/goodwe-addon
 arch:
@@ -104,7 +104,7 @@ def check(name, cond):
 # release/1.0 rather than against release/1.0.
 # ---------------------------------------------------------------------------
 _DEV = ('{"identifiers": ["goodwe_rs485_controller"], "name": "GoodWe", '
-        '"manufacturer": "GoodWe (via RS485 meter emulation)", '
+        '"manufacturer": "GoodWe (via RS485 EMS port)", '
        '"model": "ES/BP series"}')

 BASELINE = [
@@ -442,7 +442,7 @@ for c_ in unit_cfgs:
 print(f"    regression pin (sha256 of the eight payloads): {_pin}")
 check("the per-unit payloads are the ones this PR was reviewed with "
      "(REGRESSION PIN, not a correctness claim - see the comment)",
-      _pin == "eb5022b0455158d7")
+      _pin == "b4274714d8557c37")

 # ---------------------------------------------------------------------------
 print("\n4. zero overlap between A's entities and B's")
@@ -976,7 +976,7 @@ def announce_p2(units=(), omit=(), include=(), modes=(), retract=(),
 # `git show ed739d4:goodwe_controller/app/mqtt.py` and this loop.
 ROSTERS = [[], ["inverter-a"], ["inverter-a", "inverter-b"], ["a", "b", "c"],
           ["inverter-b", "Inverter B", "x", "y"], ["café"]]
-ED739D4_OFF = "501a7d93782a260b"
+ED739D4_OFF = "057be1d8ad7bb582"  # DOC-8: ed739d4 digest with only the manufacturer string changed
 _h, _n, _subs_off = hashlib.sha256(), 0, []
 for _r in ROSTERS:
    for _om in [(), ("p1_age",), ("p1_age", "grid")]:
@@ -0,0 +1,267 @@
+"""Runnable check for OPT-20.  `python3 test_opt20.py`
+
+A refusal probe (OPT-15 F3, SAFETY-73) asks the probing battery PROBE_W and
+nothing else. On TEST-23 (OPT-15 g7) the probe made the trade ask 2500 W: B
+was pinned at 200 W and the allocator put the other 2300 W on A, which sat in
+its maintenance charge phase at 0 W, with the grid at 3100 W for 34 s.
+
+1. modetrade: the ask carries PROBE_W for the probing battery, the other
+   chargers' ratings as without the probe, and `size()` adds the probe on top
+   of their share (or asks PROBE_W exactly, or nothing).
+2. main: the ask carries PROBE_W for exactly the pins filed this tick, so the
+   first and the last cycle of a window do not move another battery's share.
+3. The real Controller in the lagged replay, with B refusing and probed while
+   A is (a) in its maintenance charge phase (also with the charge shaped to
+   0 W), (b) idle, (c) a normal charger in the same run, and the OPT-15 rig's
+   g7 sequence: A's command is the same through the window as just before it,
+   B gets PROBE_W, and the commanded charge rises by PROBE_W at most.
+4. Battery modes on with no probe: byte-identical to release/1.0 bf9019e.
+
+Out of scope, and unchanged: a normal (non-probing) trade placed on a battery
+in maintenance charge (OPT-11 P4 F4) and the action-log wording (UX-38).
+Mutations: MUTATION.md says how to run them (purged copies); the 12 of
+OPT-20 are listed in the pull request.
+"""
+
+import dataclasses
+import logging
+import os
+import sys
+from types import SimpleNamespace
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, HERE)
+sys.path.insert(0, os.path.join(HERE, "tools"))
+logging.disable(logging.WARNING)
+
+import opt20_scen as sc                                        # noqa: E402
+from app import arbitrage, modetrade, packmode                 # noqa: E402
+from app import control as ctl                                 # noqa: E402
+from app import main                                           # noqa: E402
+from app.arbiter import Claim, P_SAFETY                        # noqa: E402
+
+fails = []
+
+
+def check(name, cond):
+    print(f"  {'ok  ' if cond else 'FAIL'} {name}")
+    if not cond:
+        fails.append(name)
+
+
+PW = ctl.PROBE_W
+A, B = sc.A, sc.B
+
+# ============================================================================ #
+print("1. modetrade: the ask and the size")
+_runs = SimpleNamespace(on=True, cheap=SimpleNamespace(holds=lambda n: True),
+                        charge_useful=lambda: True, give_at=lambda n: None,
+                        sell=None)
+
+
+def _dec(packs, **kw):
+    args = dict(charge_ok=True, sell_ok=True, trade_why=None, price_now=0.1,
+                headroom_w=3000.0, floor_pct=20.0, ceiling_pct=95.0,
+                house_w=800.0)
+    args.update(kw)
+    return modetrade.decide(_runs, sc.UTC0, packs, **args)
+
+
+_pb = modetrade.Pack(B, True, packmode.CHEAP_HOURS, False, 2500.0, 30.0, 10.0,
+                     refused=-1, probing=True)
+_pa = modetrade.Pack(A, True, packmode.CHEAP_HOURS, False, 2500.0, 50.0, 10.0)
+_pm = dataclasses.replace(_pa, mode=packmode.SOLAR_FIRST, idle=False)
+_pi = dataclasses.replace(_pa, mode=packmode.SOLAR_FIRST)
+
+_t_alone = _dec([_pb, _pm])
+check("a probe beside a battery in maintenance asks PROBE_W and no more",
+      _t_alone.action == modetrade.CHARGE and _t_alone.ask_w == -PW
+      and _t_alone.probe_w == PW)
+check("...beside an idle battery that cannot charge from the grid too",
+      _dec([_pb, _pi]).ask_w == -PW and _dec([_pb, _pi]).probe_w == PW)
+_t_two = _dec([_pb, _pa])
+check("...beside a normal charger the ask is that charger's rating plus "
+      "PROBE_W", _t_two.ask_w == -2500.0 - PW and _t_two.probe_w == PW)
+_t_np = _dec([dataclasses.replace(_pb, probing=False), _pa])
+check("no probe: the ask is the chargers' ratings, probe_w is 0 (as before)",
+      _t_np.ask_w == -2500.0 and _t_np.probe_w == 0.0
+      and _dec([_pa, _pm]).ask_w == -2500.0
+      and modetrade.Tick({}, {}, None, 0.0, None).probe_w == 0.0)
+
+
+def sz(ask, probe_w, lo, room, action=modetrade.CHARGE):
+    t = modetrade.Tick({}, {}, action, ask, None, probe_w)
+    return modetrade.size(t, (lo, 2500.0), room)
+
+
+check("size, a probe alone: PROBE_W, whatever the headroom above it",
+      sz(-PW, PW, -(PW), 3000.0) == -PW and sz(-PW, PW, -2700.0, 9000.0) == -PW)
+check("...exactly PROBE_W or nothing: a headroom or a bound under it files "
+      "no trade (a smaller ask voids the pin and the allocator would place "
+      "it elsewhere)",
+      sz(-PW, PW, -2700.0, PW - 1) == 0.0 and sz(-PW, PW, -(PW - 1), 3000.0)
+      == 0.0 and sz(-PW, PW, -2700.0, 0.0) == 0.0
+      and sz(-PW, PW, -2700.0, PW) == -PW)
+check("size with another charger: its share is sized as without the probe, "
+      "the probe's PROBE_W is added on top",
+      sz(-2700.0, PW, -2700.0, 3000.0) == -2700.0
+      and sz(-2700.0, PW, -2700.0, 1000.0) == -1200.0
+      and sz(-2700.0, PW, -1700.0, 3000.0) == -1700.0
+      and sz(-2700.0, PW, -2700.0, 0.0) == 0.0)
+check("size with no probe is what it was (ask, bound, headroom)",
+      sz(-2500.0, 0.0, -2700.0, 3000.0) == -2500.0
+      and sz(-2500.0, 0.0, -2700.0, 1000.0) == -1000.0
+      and sz(-2500.0, 0.0, -1800.0, 3000.0) == -1800.0
+      and sz(-2500.0, 0.0, -2700.0, None) == 0.0
+      and modetrade.size(modetrade.Tick({}, {}, None, 0.0, None),
+                         (-5.0, 5.0), 3000.0) == 0.0)
+
+# ============================================================================ #
+print("\n2. main: the ask carries the pins filed this tick")
+
+
+def trade(tick, pins, bound=(-2700.0, 2500.0), room=3000.0):
+    """The trade `_mode_trade()` files: (claims, size or None)."""
+    claims = [Claim.set("refused", P_SAFETY, -PW * 1.0, ctl.PROBING,
+                        actuator="x") for _ in range(pins)]
+    me = SimpleNamespace(
+        mode_tick=tick, loop_bound=lambda cl: bound, arb_headroom_w=room,
+        mode_runs=SimpleNamespace(
+            cheap=SimpleNamespace(start=sc.UTC0, end=sc.UTC0,
+                                  mean_eur_kwh=0.1),
+            sell=None, level_eur_kwh=0.1),
+        arb=None, _arb_action=None, log_event=lambda *a, **k: None)
+    filed = main.Controller._mode_trade(me, claims, sc.UTC0)
+    t = [c for c in claims if c.source == arbitrage.SOURCE]
+    return filed, (t[0].value if t else None), me.mode_tick
+
+
+_t0 = modetrade.Tick({}, {A: modetrade.CHARGE}, modetrade.CHARGE, -2500.0,
+                     None, 0.0)
+check("the pin is filed but decide() ran before it (the window's first "
+      "cycle): the charger keeps its whole ask and the pin is added",
+      trade(_t0, 1)[1] == -2700.0 and trade(_t0, 1)[2].probe_w == PW)
+check("the window closed after decide() (its last cycle): the probe's PROBE_W "
+      "leaves the ask, so no battery is handed it",
+      trade(_t_two, 0)[1] == -2500.0
+      and trade(_t_alone, 0) == (False, None, trade(_t_alone, 0)[2])
+      and trade(_t_alone, 0)[2].action is None)
+check("pin and ask agree: nothing changes",
+      trade(_t_two, 1)[1] == -2700.0 and trade(_t_alone, 1)[1] == -PW
+      and trade(_t0, 0)[1] == -2500.0)
+
+# ============================================================================ #
+print("\n3. the real Controller: B refuses and is probed")
+
+
+def charge(r):
+    return -sum(min(0.0, v) for v in r["shares"].values())
+
+
+def window_checks(rows, label):
+    """Per probe window: A as just before, B at PROBE_W (or 0 when nothing is
+    traded), the charge up by at most PROBE_W."""
+    win = sc.windows(rows)
+    ok_a = ok_b = ok_sum = ok_grid = True
+    for i, j in win:
+        pre = rows[i - 1]
+        lo, hi = max(0, i - 3), min(len(rows) - 1, j + 3)
+        ext = rows[lo:hi + 1]
+        base_sum = sum(min(0.0, v) for v in pre["shares"].values())
+        ok_a &= all(abs(r["shares"].get(A, 0.0) - pre["shares"].get(A, 0.0))
+                    < 1e-6 for r in ext)
+        bs = [r["shares"].get(B, 0.0) for r in ext]
+        if any(r["action"] for r in rows[i:j + 1]):
+            ok_b &= set(bs) <= {0.0, -PW} and bs.count(-PW) >= 20
+        else:       # a drain comes first: the window trades nothing
+            ok_b &= set(bs) == {0.0}
+        ok_sum &= all(charge(r) <= charge(pre) + PW + 1e-6 for r in ext)
+        ok_grid &= max(r["grid"] for r in ext) <= pre["grid"] + PW + 50.0
+    n_ch = sum(1 for r in rows if r["probing"] and r["action"])
+    print(f"    {label}: {len(win)} window(s), {n_ch} probing cycles with a "
+          f"trade, highest grid "
+          f"{max(r['grid'] for r in rows):.0f} W")
+    return win, ok_a, ok_b, ok_sum, ok_grid
+
+
+SCEN = {
+    "(a) A in its maintenance charge phase": (
+        sc.day(sc.SF_CH, a_maint=sc.MAINT_CHARGE), "same"),
+    "(a') ...with the charge shaped to 0 W": (
+        sc.day(sc.SF_CH, a_maint=sc.MAINT_CHARGE, shaped=True), "same"),
+    "(b) A idle at 0 W (Solar first, the sun covers the house)": (
+        sc.day(sc.SF_CH, house_w=-10.0), "same"),
+    "(c) A a normal charger in the same run": (sc.day(sc.CH_CH), "same"),
+    "(g7) the OPT-15 rig: drain, then maintenance charge at 0 W": (
+        sc.g7(), "same"),
+    # Not the fix's: A gives the house, and a charge ask means no other
+    # battery gives (OPT-11 owner answer 3), so A rests, as on bf9019e.
+    "(b2) A gives 800 W to the house": (sc.day(sc.SF_CH), "rests"),
+}
+for _name, (_rows, _kind) in SCEN.items():
+    win, ok_a, ok_b, ok_sum, ok_grid = window_checks(_rows, _name)
+    check(f"{_name}: a probe happens (non-vacuous)", len(win) >= 1)
+    if _kind == "same":
+        check(f"{_name}: A's command is the same through each window, first "
+              "and last cycle included", ok_a)
+        check(f"{_name}: the grid rises by PROBE_W at most (plus 50 W of "
+              "noise)", ok_grid)
+    else:
+        check(f"{_name}: A stops giving for the window (the one-sign rule, "
+              "as on bf9019e), charges nothing and gives back after it",
+              all(r["shares"].get(A, 0.0) == 0.0 for i, j in win
+                  for r in _rows[i + 1:j]) and all(
+                  _rows[j + 4]["shares"].get(A, 0.0) > 0.0 for i, j in win))
+    check(f"{_name}: B gets PROBE_W and nothing else", ok_b)
+    check(f"{_name}: the commanded charge rises by PROBE_W at most", ok_sum)
+
+# The g7 sequence in detail: the probe after the drain.
+_g7 = SCEN["(g7) the OPT-15 rig: drain, then maintenance charge at 0 W"][0]
+_w7 = sc.windows(_g7)
+_after = [w for w in _w7 if _g7[w[0]]["now"] > sc.at(18.3, sc.GUTC0)]
+_peak7 = max(r["grid"] for r in _g7 if r["now"] > sc.at(18.3, sc.GUTC0))
+print(f"    g7: probes after the drain at "
+      f"{[_g7[w[0]]['now'].astimezone(sc.L).strftime('%H:%M:%S') for w in _after]}"
+      f", grid after the drain at most {_peak7:.0f} W (TEST-23: 3100 W)")
+check("g7: the probe after A's drain asks B for 200 W with A at 0 W "
+      "(TEST-23: -2300 W on A, the grid at 3100 W)",
+      _after and all(
+          abs(r["shares"].get(A, 0.0)) < 1e-6 for i, j in _after
+          for r in _g7[i:j + 4]) and _peak7 <= 800.0 + PW + 50.0)
+check("g7: during the drain no probe trades (a drain comes first)",
+      all(r["action"] is None for r in _g7
+          if sc.at(18.0, sc.GUTC0) <= r["now"] < sc.at(18.28, sc.GUTC0)))
+_acts = [r["reason"] for i, j in _after for r in _g7[i:j + 1]
+         if r["action"]]
+check("g7: the action log's trade line names the probe's watts, 200 W",
+      _acts and all(f"{PW:.0f} W on {B}" in a for a in _acts))
+check("no opposite-direction split on any of these cycles",
+      not [r for _r, _ in SCEN.values() for r in _r
+           if any(v > 0 for k, v in r["shares"].items() if r["action"])
+           and any(v < 0 for v in r["shares"].values())])
+
+# ============================================================================ #
+print("\n4. battery modes on, no probe: byte-identical to release/1.0 bf9019e")
+# Printed by tools/opt20_scen.py on release/1.0 bf9019e (2026-10-15),
+# goodwe-addon-checks:1.
+AT_BASE = {
+    "sf_ch": "422d6578a7decf1ae3749edd1898ff3cd0541d904c95f636cdd9b1de8c1a07c9",
+    "ch_ch": "15fc25dd34186e56d34b752d37c9e8613295ba737938c92454acca08c83080fb",
+    "ch_ch_maint":
+        "e46e3c666dd63ab33b56e83c0a560c07cec7774a9a70560f55c62a6ae6f55920",
+}
+_got = sc.no_probe_digests()
+for _k in AT_BASE:
+    print(f"    {_k} {'same' if _got[_k] == AT_BASE[_k] else 'DIFFERS'} "
+          f"{_got[_k][:16]}")
+check("with no refusal every share, grid, trade sentence, role and plant "
+      "power of a scripted cheap run equals bf9019e's (A Solar first, A "
+      "charging, A in maintenance charge)", _got == AT_BASE)
+
+print()
+if fails:
+    print(f"{len(fails)} FAILED:")
+    for f in fails:
+        print(f"  - {f}")
+    sys.exit(1)
+print("all checks passed")
@@ -0,0 +1,375 @@
+"""SAFETY-118: the peak forecast is read in its own unit.  `python3 -B test_safety118.py`
+
+`peak_forecast_entity` used to be read as a bare number and anything under 50
+was multiplied by 1000, so a W sensor reading 17 W was taken as 17 kW. It is
+now read through `pv.watts()` by its own `unit_of_measurement`. No unit, or a
+unit that is not power: the forecast is not used, as if the option were empty.
+
+1. The reader through the real `read_inputs()`: W, kW, MW, no unit, kWh, bad state.
+2. The warning: once, again only when the cause changes, never every cycle.
+3. One state read per cycle, as before.
+4. The maintenance charge headroom and the charge-only clamp act on the
+   converted value (real `cycle()`).
+5. The validator, the status row and the strings.
+
+Mutating a module this file grades? Read MUTATION.md first (purge __pycache__).
+"""
+
+import asyncio
+import logging
+from pathlib import Path
+
+from app import main as _main
+from app import messages, pv, validate
+from app.main import Controller
+
+fails = []
+
+
+def check(name, cond):
+    print(f"  {'ok  ' if cond else 'FAIL'} {name}")
+    if not cond:
+        fails.append(name)
+
+
+def st(value, unit=None):
+    attrs = {} if unit is None else {"unit_of_measurement": unit}
+    return {"state": str(value), "attributes": attrs}
+
+
+class _Store:
+    def __init__(self, data=None):
+        self.data = dict(data or {})
+
+    def set(self, k, v):
+        self.data[k] = v
+
+    def get_time(self, *a):
+        return None
+
+    def set_time(self, *a):
+        pass
+
+
+class _Pub:
+    enabled = False
+
+    def publish(self, *a, **k):
+        pass
+
+    def close(self):
+        pass
+
+
+class _Hass:
+    def __init__(self, fc=None, prices=None):
+        self.fc = fc
+        self.prices = prices or {}
+        self.reads = []
+        self.writes = []
+
+    async def limits(self, entity):
+        return (-5000.0, 5000.0, 1.0)
+
+    async def state(self, entity):
+        self.reads.append(entity)
+        if entity == "sensor.fc":
+            return self.fc
+        v = {"sensor.grid": 0.0, "sensor.soc": 50.0, "sensor.batt": 0.0,
+             **self.prices}.get(entity)
+        return None if v is None else st(v, "W")
+
+    async def number(self, entity, invert=False):
+        obj = await self.state(entity)
+        try:
+            return float(obj["state"]) if obj else None
+        except ValueError:
+            return None
+
+    async def set_number(self, entity, value):
+        self.writes.append(float(value))
+        return True
+
+    async def call(self, domain, service, data):
+        return True
+
+
+OPTS = {"setpoint_entity": "number.sp", "meter_entity": "sensor.grid",
+        "soc_entity": "sensor.soc", "batt_entity": "sensor.batt",
+        "meter_source": "off", "heartbeat_s": 10, "stale_input_s": 15,
+        "max_w": 5000.0, "gain": 0.6, "slew_w": 400.0, "deadband_w": 15,
+        "step_w": 10, "target_grid_w": -10.0, "stale_input_s": 15,
+        "peak_cap_w": 5000, "peak_margin_w": 0,
+        "peak_forecast_entity": "sensor.fc"}
+
+
+class _Rec(logging.Handler):
+    def __init__(self):
+        super().__init__(logging.DEBUG)
+        self.rows = []
+
+    def emit(self, r):
+        self.rows.append((r.levelno, r.getMessage()))
+
+
+def make(fc, store=None, prices=None, **over):
+    hass = _Hass(fc, prices)
+    c = Controller(dict(OPTS, **over), hass, _Store(store), _Pub(),
+                   stop=asyncio.Event())
+    c.auto, c.dev_min, c.dev_max = True, -5000.0, 5000.0
+    return c, hass
+
+
+def reads(c):
+    asyncio.run(c.read_inputs())
+    return c.peak_fc
+
+
+def peak_of(state):
+    c, _h = make(state)
+    return reads(c)
+
+
+print("1. the reader")
+CASES = [("W 17", st(17, "W"), 17.0), ("W 820", st(820, "W"), 820.0),
+         ("kW 0.82", st(0.82, "kW"), 820.0), ("kW 0.017", st(0.017, "kW"), 17.0),
+         ("MW 0.0008", st(0.0008, "MW"), 800.0), ("W 49", st(49, "W"), 49.0),
+         ("W 50", st(50, "W"), 50.0), ("kW 30", st(30, "kW"), 30000.0),
+         ("W 0", st(0, "W"), 0.0)]
+for label, s, want in CASES:
+    got = peak_of(s)
+    check(f"{label} -> {want} W (got {got})",
+          got is not None and abs(got - want) < 1e-6)
+for label, s in [("no unit", st(820)), ("empty unit", st(820, "")),
+                 ("kWh", st(820, "kWh")), ("percent", st(20, "%")),
+                 ("unavailable", st("unavailable", "W")),
+                 ("unknown", st("unknown", "W")), ("missing entity", None),
+                 ("not a number", st("sunny", "W"))]:
+    check(f"{label} -> None (got {peak_of(s)})", peak_of(s) is None)
+c, _h = make(st(820, "W"), peak_forecast_entity="")
+check("option empty -> None and the entity is never read",
+      reads(c) is None and "sensor.fc" not in _h.reads)
+
+print("2. the warning")
+rec = _Rec()
+lg = logging.getLogger("goodwe")
+lg.addHandler(rec)
+lg.setLevel(logging.DEBUG)
+
+
+def warns():
+    return [m for lv, m in rec.rows if lv >= logging.WARNING]
+
+
+c, h = make(st(820, "kWh"))
+for _ in range(5):
+    reads(c)
+check(f"kWh: five cycles, one WARNING ({warns()})", len(warns()) == 1)
+check("...short, plain, and says it is not used",
+      warns()[0] == "The 15-minute demand forecast is not in a power unit, so "
+      "it is not used." and len(warns()[0].split()) <= 25)
+h.fc = st(820, "MWh")
+reads(c)
+check("the same cause with another unit warns again (the cause changed)",
+      len(warns()) == 2)
+h.fc = st(820)
+reads(c)
+check("no unit is a different cause and warns", len(warns()) == 3
+      and "has no unit" in warns()[2])
+reads(c)
+check("...and only once", len(warns()) == 3)
+USED = (logging.INFO, "The 15-minute demand forecast is used again.")
+
+
+def used():
+    return [r for r in rec.rows if r == USED]
+
+
+h.fc = st("unavailable")
+reads(c)
+check("unavailable adds no warning and no 'used again'",
+      len(warns()) == 3 and not used())
+h.fc = st(820, "W")
+reads(c)
+check("a good unit is read again: one INFO, no new warning",
+      c.peak_fc == 820.0 and len(warns()) == 3 and len(used()) == 1
+      and rec.rows[-1] == USED)
+n = len(warns())
+for s in (st(820, "W"), st(0.8, "kW")):
+    h.fc = s
+    reads(c)
+check("W and kW never warn", len(warns()) == n)
+rows = [r for r in c.checks() if r.get("id") == "peak_forecast"]
+check("a good forecast has no peak_forecast row", not rows)
+h.fc = st(820, "kWh")
+reads(c)
+rows = [r for r in c.checks() if r.get("id") == "peak_forecast"]
+check("the status page carries a caution row for the unit fault",
+      len(rows) == 1 and rows[0]["warn"] and not rows[0]["ok"]
+      and rows[0]["text"] == messages.row_peak_forecast_unit("kWh"))
+lg.removeHandler(rec)
+
+print("3. one state read per cycle")
+c, h = make(st(0.82, "kW"))
+reads(c)
+check(f"one read of the forecast entity per read_inputs() ({h.reads})",
+      h.reads.count("sensor.fc") == 1)
+
+print("4. maintenance headroom and the charge-only clamp use watts")
+seen = []
+_real = _main.maintenance_charge_floor
+
+
+def spy(charge_w, fc, cap):
+    out = _real(charge_w, fc, cap)
+    seen.append((charge_w, fc, cap, out))
+    return out
+
+
+_main.maintenance_charge_floor = spy
+try:
+    def cycle(fc, **kw):
+        c, h = make(fc, store={"phase": "charge"},
+                    prices={"sensor.pnow": 1.0, "sensor.pavg": 2.0},
+                    price_now_entity="sensor.pnow",
+                    price_avg_entity="sensor.pavg", **kw)
+        seen.clear()
+        asyncio.run(c.read_inputs())
+        asyncio.run(c.cycle())
+        return c, h
+
+    # cap 2500 W, charge rate 2000 W: 820 W of forecast leaves 1680 W, 17 W
+    # leaves the whole 2000 W (the size rule read 17 as 17 kW and left 0).
+    c, h = cycle(st(0.82, "kW"), peak_cap_w=2500)
+    check(f"cheap-window headroom is cap - 820 W = 1680 W ({seen})",
+          len(seen) == 1 and seen[0][1] == 820.0 and seen[0][0] == 2000.0
+          and abs(seen[0][3] - 1680.0) < 1e-6)
+    c, h = cycle(st(17, "W"), peak_cap_w=2500)
+    check(f"a 17 W forecast leaves the full 2000 W charge rate ({seen})",
+          len(seen) == 1 and seen[0][1] == 17.0 and abs(seen[0][3] - 2000.0) < 1e-6)
+    # charge-only clamp: not cheap, shaping suspended only if the forecast is
+    # over the aim.  6 kW is over 5000 W; 17 W is not (the old rule read it
+    # as 17 kW and suspended).
+    def clamp(fc):
+        c, h = make(fc, store={"phase": "charge"})
+        asyncio.run(c.read_inputs())
+        asyncio.run(c.cycle())
+        return c._maint_shape_suspended
+    check("kW 6.0 (6000 W, over the 5000 W aim) suspends the charge-only clamp",
+          clamp(st(6.0, "kW")) is True)
+    check("W 17 leaves the charge-only clamp in place",
+          clamp(st(17, "W")) is False)
+    check("kW 0.017 leaves the charge-only clamp in place",
+          clamp(st(0.017, "kW")) is False)
+    check("no unit: forecast unused, clamp in place, as with the option empty",
+          clamp(st(6000)) is False)
+finally:
+    _main.maintenance_charge_floor = _real
+
+print("5. validator and strings")
+def notes(unit, key="peak_forecast_entity"):
+    ent = {"sensor.fc": {"entity_id": "sensor.fc", "unit": unit,
+                         "state": "1"}}
+    return [n for n in validate.config_notes(
+        {key: "sensor.fc", "meter_source": "off"}, entities=ent)
+        if key in n.keys]
+
+
+check("validator: W no note", not notes("W"))
+check("validator: kW no note", not notes("kW"))
+check("validator: MW no note", not notes("MW"))
+nu, kw = notes(""), notes("kWh")
+check(f"validator: no unit warns ({nu})", len(nu) == 1 and nu[0].tier == "warn"
+      and "has no unit" in nu[0].text)
+check(f"validator: kWh warns ({kw})", len(kw) == 1 and kw[0].tier == "warn"
+      and "not in a power unit" in kw[0].text)
+check("validator: the false 'taken as watts' claim is gone for this entity",
+      not any("taken as watts" in n.text for n in nu + kw + notes("kW")))
+check("validator: another watt key keeps its claim",
+      len(notes("kW", "batt_entity")) == 1
+      and "taken as watts" in notes("kW", "batt_entity")[0].text)
+check("WATT_ENTITY_KEYS no longer holds the peak forecast",
+      "peak_forecast_entity" not in validate.WATT_ENTITY_KEYS
+      and "batt_entity" in validate.WATT_ENTITY_KEYS)
+print("   EN no unit :", messages.peak_forecast_unit(""))
+print("   EN bad unit:", messages.peak_forecast_unit("kWh"))
+print("   NL no unit :", messages.peak_forecast_unit("", "nl"))
+print("   NL bad unit:", messages.peak_forecast_unit("kWh", "nl"))
+check("the strings differ by cause and by language",
+      len({messages.peak_forecast_unit(u, l) for u in ("", "kWh")
+           for l in ("en", "nl")}) == 4)
+print("6. a missing reading leaves the unit fault as it was (F1)")
+
+
+lg.addHandler(rec)
+
+
+def prow(c):
+    return [r for r in c.checks() if r.get("id") == "peak_forecast"]
+
+
+def flap(first, gap, last):
+    """first, gap, first again (or last): the log rows and the status row
+    after each cycle."""
+    rec.rows.clear()
+    c, h = make(first)
+    out = []
+    for s in (first, gap, last):
+        h.fc = s
+        reads(c)
+        out.append((len(warns()), len(used()), len(prow(c)), c.peak_fc))
+    return out
+
+
+for label, first, gap in [
+        ("kWh -> unavailable -> kWh", st(820, "kWh"), st("unavailable", "kWh")),
+        ("kWh -> unknown -> kWh", st(820, "kWh"), st("unknown", "kWh")),
+        ("kWh -> failed read -> kWh", st(820, "kWh"), None),
+        ("kWh -> non-numeric -> kWh", st(820, "kWh"), st("sunny", "kWh")),
+        ("no unit -> failed read -> no unit", st(820), None),
+        ("no unit -> unavailable -> no unit", st(820), st("unavailable"))]:
+    r = flap(first, gap, first)
+    check(f"{label}: one WARNING, no 'used again', row on all 3 ({r})",
+          [x[0] for x in r] == [1, 1, 1] and [x[1] for x in r] == [0, 0, 0]
+          and [x[2] for x in r] == [1, 1, 1]
+          and [x[3] for x in r] == [None, None, None])
+r = flap(st(820, "kWh"), st("unavailable", "kWh"), st(820, "W"))
+check(f"kWh -> unavailable -> W: the fault clears only on the W reading ({r})",
+      [x[1] for x in r] == [0, 0, 1] and [x[2] for x in r] == [1, 1, 0]
+      and [x[0] for x in r] == [1, 1, 1] and r[2][3] == 820.0)
+rec.rows.clear()
+c, h = make(st(820, "W"))
+for s in (st(820, "W"), st("unavailable", "W"), None, st("sunny", "W")):
+    h.fc = s
+    reads(c)
+    check(f"good unit, reading {s and s['state']!r}: no row, no log",
+          not prow(c) and not warns() and not used())
+rec.rows.clear()
+c, h = make(st(820, "kWh"))
+reads(c)
+for s in (st("unavailable", "kWh"), None):
+    h.fc = s
+    reads(c)
+    check("fault row stays while the reading is missing",
+          len(prow(c)) == 1 and prow(c)[0]["text"]
+          == messages.row_peak_forecast_unit("kWh"))
+c, h = make(st(820, "W"))
+h.fc = st("unavailable", "W")
+reads(c)
+check("good-unit sensor only unavailable: the fault row stays away",
+      not prow(c))
+lg.removeHandler(rec)
+check("has_reading: numbers only",
+      pv.has_reading(st(0, "kWh")) and pv.has_reading(st("1.5"))
+      and not pv.has_reading(None) and not pv.has_reading(st("unavailable"))
+      and not pv.has_reading(st("unknown")) and not pv.has_reading(st("x")))
+check("pv.unit_fault: None for a good unit and for a missing reading (has_reading tells them apart)",
+      pv.unit_fault(st(1, "kW")) is None and pv.unit_fault(None) is None
+      and pv.unit_fault(st("unavailable")) is None
+      and pv.unit_fault(st(1)) == "" and pv.unit_fault(st(1, "kWh")) == "kWh")
+src = (Path(__file__).parent / "app" / "main.py").read_text(encoding="utf-8")
+check("the size rule is gone from main.py", "abs(raw) < 50" not in src)
+
+print()
+print("FAILED: " + "; ".join(fails) if fails else "all checks passed")
+raise SystemExit(1 if fails else 0)
@@ -286,6 +286,22 @@ check("app/uitext.py falls back to English per key and reads nl-BE as nl",
      and uitext._merge({"a": {"b": 1, "c": 2}}, {"a": {"b": 3}})
      == {"a": {"b": 3, "c": 2}})

+# ============================================================================ #
+print("\n7. the dynamic contract choice says 15 minutes, never hourly (UX-44)")
+# Belgian day-ahead prices are per 15 minutes since 2025-10-01.
+HOURLY = re.compile(r"hourly|per hour|\bhour\b|per uur|\buur\b|uurprijs",
+                    re.I)
+_dyn = {"en": EN["choices"]["contract"]["dynamic"],
+        "nl": NL["choices"]["contract"]["dynamic"]}
+check(f"the label reads 'Dynamic, every 15 minutes' / 'Dynamisch, per "
+      f"kwartier' ({_dyn})",
+      _dyn == {"en": "Dynamic, every 15 minutes",
+               "nl": "Dynamisch, per kwartier"})
+check("neither language calls the dynamic contract hourly",
+      not any(HOURLY.search(t) for t in _dyn.values()))
+check("mutation: the old labels are caught",
+      all(HOURLY.search(t) for t in ("Dynamic, hourly", "Dynamisch, per uur")))
+
 print()
 if fails:
    print(f"{len(fails)} FAILED")
@@ -327,6 +327,66 @@ check("days the editor has no choice for stay in the Configuration tab, and "
      ["ui.periods"] is None and edit(dict(SITE, static_profile=_raw), gain=0.5)
      ["static_profile"] == _raw)

+print("\n8. UX-43: tariff prices take five decimals")
+from app import uilayout as _L                                  # noqa: E402
+_PX = 0.34512
+_TARIFF = ("ui.fixed_price", "ui.day_price", "ui.night_price", "ui.periods")
+_steps = {k: _L.field(k).step for k in _TARIFF}
+check("the four tariff price fields step 0.00001 EUR/kWh",
+      sorted(_steps) == sorted(_TARIFF)
+      and all(v == 0.00001 for v in _steps.values()))
+_f5 = edit(DEFAULTS, ui__contract="fixed", ui__fixed_price=_PX)
+check("fixed price 0.34512 is saved exactly and read back exactly",
+      '"eur_per_kwh": 0.34512}' in _f5["static_profile"]
+      and U.options_to_ui(_f5)["ui.fixed_price"] == _PX
+      and trip(_f5) == _f5)
+_d5 = edit(SITE, ui__contract="daynight", ui__day_price=_PX,
+           ui__night_price=0.26789)
+check("day price 0.34512 and night price 0.26789 are saved and read back exactly",
+      '"eur_per_kwh": 0.34512' in _d5["static_profile"]
+      and '"eur_per_kwh": 0.26789' in _d5["static_profile"]
+      and U.options_to_ui(_d5)["ui.day_price"] == _PX
+      and U.options_to_ui(_d5)["ui.night_price"] == 0.26789
+      and not validate_startup_config(_d5))
+_n5 = edit(SITE, ui__contract="daynight", ui__day_price=0.3, ui__night_price=_PX)
+check("night price 0.34512 alone is saved and read back exactly",
+      U.options_to_ui(_n5)["ui.night_price"] == _PX)
+_r3 = rows(("00:00", "07:00", "mon-sun", 0.12345), ("07:00", "22:00", "mon-sun", _PX),
+           ("22:00", "24:00", "mon-sun", 0.05001))
+_p3 = periods(_r3)
+check("three windows with five-decimal prices save and read back exactly",
+      [p.eur_per_kwh for p in parse_static_profile(_p3["static_profile"]).periods]
+      == [0.12345, _PX, 0.05001]
+      and [r["price"] for r in U.options_to_ui(_p3)["ui.periods"]]
+      == [0.12345, _PX, 0.05001] and trip(_p3) == _p3
+      and "e-" not in _p3["static_profile"])
+check("a two-decimal price still works", U.options_to_ui(
+    edit(DEFAULTS, ui__contract="fixed", ui__fixed_price=0.31))
+    ["ui.fixed_price"] == 0.31)
+_old = dict(SITE, price_provider="off", static_profile=json.dumps({"periods": [
+    {"name": "day", "from": "07:00", "to": "22:00", "days": "mon-fri",
+     "eur_per_kwh": 0.34},
+    {"name": "night", "from": "22:00", "to": "07:00", "days": "mon-fri",
+     "eur_per_kwh": 0.26},
+    {"name": "night", "from": "00:00", "to": "24:00", "days": "sat,sun",
+     "eur_per_kwh": 0.26}]}))
+check("a stored profile with two-decimal prices is unchanged by load and save",
+      trip(_old) == _old)
+_js = open(os.path.join(HERE, "app", "static", "settings.js"),
+           encoding="utf-8").read()
+check("the window editor's price input takes its step from the field, not a "
+      "hard-coded 0.01", 'step: "0.01"' not in _js
+      and "step: String(f.step || 0.01)" in _js)
+for _lang in ("en", "nl"):
+    with open(os.path.join(HERE, "app", "ui_text", f"{_lang}.json"),
+              encoding="utf-8") as _fh:
+        _flds = json.load(_fh)["fields"]
+    _hint = "34.512" if _lang == "en" else "34,512"
+    _eq = "0.34512" if _lang == "en" else "0,34512"
+    check(f"[{_lang}] the c€/kWh hint sits on all four tariff price fields",
+          all(_hint in _flds[k]["description"] and _eq in _flds[k]["description"]
+              for k in ("fixed_price", "day_price", "night_price", "periods")))
+
 print()
 if fails:
    print(f"{len(fails)} FAILED")
@@ -105,6 +105,13 @@ class _Hass:
            return self.power(u)
        return None

+    async def state(self, e):
+        """SAFETY-118: the peak forecast is read as a state object; every
+        number this rig serves is in watts."""
+        v = await self.number(e)
+        return None if v is None else {
+            "state": str(v), "attributes": {"unit_of_measurement": "W"}}
+
    async def set_number(self, e, v):
        self.cmd[e[-1]] = float(v)
        return True
@@ -0,0 +1,160 @@
+"""OPT-20: the refusal-probe scenarios, on the REAL Controller in the lagged
+replay (tools/opt11p3_replay.py). Uses only the API bf9019e has, so the same
+file runs on both trees (test_opt20.py pins the digests printed on bf9019e).
+
+Battery B is Cheap hours and refuses to charge; it is probed every 15 minutes
+inside the cheap run (OPT-15). A is, per scenario, in its maintenance charge
+phase, idle (Solar first), or a normal charger in the same run.
+"""
+
+import hashlib
+import logging
+import os
+import sys
+from datetime import datetime, timedelta, timezone
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, HERE)
+sys.path.insert(0, os.path.dirname(HERE))
+logging.disable(logging.WARNING)
+
+import opt11p3_replay as rp                                    # noqa: E402
+from app import packmode                                       # noqa: E402
+from app.maintenance import CHARGE, DRAIN, MaintResult         # noqa: E402
+
+L = timezone(timedelta(hours=2))
+A, B = "inverter-a", "inverter-b"
+LAG = rp.Lag(tau_s=0.05, ramp_w_s=1e9)
+SF_CH = {"a": (True, packmode.SOLAR_FIRST, False),
+         "b": (True, packmode.CHEAP_HOURS, False)}
+CH_CH = {"a": (True, packmode.CHEAP_HOURS, False),
+         "b": (True, packmode.CHEAP_HOURS, False)}
+
+# A day of the owner's kind: cheap 02:00-04:00 local, flat otherwise.
+D0 = datetime(2026, 9, 9, 0, 0, tzinfo=L)
+UTC0 = D0.astimezone(timezone.utc)
+_PR = [0.30] * 192
+for _i in range(8, 16):
+    _PR[_i] = 0.10
+WPTS = [(UTC0 + timedelta(minutes=15 * i), 900, p) for i, p in enumerate(_PR)]
+
+# OPT-15 rig g7: prices 0 from 17:45 to 18:45 local, flat 0.06 otherwise.
+G0 = datetime(2026, 10, 13, 0, 0, tzinfo=L)
+GUTC0 = G0.astimezone(timezone.utc)
+_GP = [0.06] * 192
+for _i in range(71, 75):
+    _GP[_i] = 0.0
+for _i in range(76, 78):
+    _GP[_i] = 0.5
+G7PTS = [(GUTC0 + timedelta(minutes=15 * i), 900, p)
+         for i, p in enumerate(_GP)]
+
+
+def at(h, base=UTC0):
+    return base + timedelta(hours=h)
+
+
+def _rows(rows):
+    def cb(c, h, now, pr):
+        sh = dict(c.allocation.shares) if c.allocation is not None \
+            and c.allocation.shares else {}
+        rows.append(dict(
+            now=now, grid=h.grid(), shares=sh,
+            action=getattr(c.arb, "action", None),
+            reason=getattr(c.arb, "reason", None) or "",
+            probing=c.inverters[1].probe_from is not None,
+            roles=dict(c.mode_tick.roles) if c.mode_tick else {},
+            p={u: p.p for u, p in h.plants.items()}))
+    return cb
+
+
+def day(choice, *, a_maint=None, shaped=False, refuse=("b",), hours=0.5,
+        house_w=800.0):
+    """02:00..: a cheap run, `house_w` house. `a_maint` is A's forced maintenance
+    result from 5 minutes in (B's refusal stands by then); `shaped` closes the
+    maintenance charge's own watts (charge-only shaping, hi = 0)."""
+    rows = []
+
+    def patch(c):
+        if shaped:
+            c._charge_headroom = lambda floor, claims: 0.0
+        if a_maint is None:
+            return
+        m, real = c.maints[0], c.maints[0].tick
+
+        def tick(now, soc, blocked_by=None, gate_open=True):
+            if now >= at(1.9 + 5 / 60):
+                return a_maint
+            return real(now, soc, blocked_by=blocked_by, gate_open=gate_open)
+        m.tick = tick
+    rp.run("p3_ch_sell", [(UTC0.timestamp() - 10, house_w)], WPTS, at(1.9),
+           3600 * hours, check=_rows(rows), lag=LAG, n=2,
+           soc0={"a": 50.0, "b": 30.0}, choice=choice, refuse=refuse,
+           patch=patch, extra={"cycle_cost_threshold_eur": 0.10})
+    return rows
+
+
+def g7():
+    """The OPT-15 rig's g7: B refuses from the start of the cheap run, A
+    drains 18:00 to 18:17:04 and then sits in its maintenance charge phase
+    (charge-only) at 0 W. The probe after the drain used to put the trade's
+    2500 W on A."""
+    rows = []
+    t_end = at(18 + 17 / 60 + 4 / 3600, GUTC0)
+
+    def patch(c):
+        m, real = c.maints[0], c.maints[0].tick
+
+        def tick(now, soc, blocked_by=None, gate_open=True):
+            if at(18.0, GUTC0) <= now < t_end:
+                return MaintResult(DRAIN, True, 2000.0, False, [])
+            if now >= t_end:
+                return MaintResult(CHARGE, False, None, True, [])
+            return real(now, soc, blocked_by=blocked_by, gate_open=gate_open)
+        m.tick = tick
+    rp.run("p3_ch_sell", [(GUTC0.timestamp() - 10, 800.0)], G7PTS,
+           at(17.6, GUTC0), 3600 * 1.1, check=_rows(rows), lag=LAG, n=2,
+           soc0={"a": 17.0, "b": 73.0}, choice=SF_CH, refuse=("b",),
+           patch=patch, extra={"cycle_cost_threshold_eur": 0.10})
+    return rows
+
+
+MAINT_CHARGE = MaintResult(CHARGE, False, None, True, [])
+
+
+def windows(rows):
+    """[(first, last)] row indexes of each probe window."""
+    out, i = [], 0
+    while i < len(rows):
+        if rows[i]["probing"]:
+            j = i
+            while j + 1 < len(rows) and rows[j + 1]["probing"]:
+                j += 1
+            out.append((i, j))
+            i = j + 1
+        else:
+            i += 1
+    return out
+
+
+def digest(rows) -> str:
+    keep = [(r["now"].isoformat(), sorted(r["shares"].items()),
+             round(r["grid"], 3), r["action"], r["reason"],
+             sorted(r["roles"].items()), sorted(r["p"].items()))
+            for r in rows]
+    return hashlib.sha256(repr(keep).encode()).hexdigest()
+
+
+def no_probe_digests() -> dict:
+    """Battery modes on, no refusal, so no probe: a scripted cheap run."""
+    return {
+        "sf_ch": digest(day(SF_CH, refuse=(), hours=1.2)),
+        "ch_ch": digest(day(CH_CH, refuse=(), hours=1.2)),
+        "ch_ch_maint": digest(day(CH_CH, a_maint=MAINT_CHARGE, refuse=(),
+                                  hours=1.2)),
+    }
+
+
+if __name__ == "__main__":
+    for k, v in no_probe_digests().items():
+        print(k, v)
@@ -165,7 +165,7 @@ configuration:
    description: "The add-on aims this much below the peak limit, so a 15-minute period that ends a little high still stays under it."
  peak_forecast_entity:
    name: "Limits · 15-minute demand forecast entity (W)"
-    description: "Optional. A forecast of this 15-minute period's average power. Leave it empty and there is no peak limit."
+    description: "Optional. A forecast of this 15-minute period's average power, in W or kW. With no power unit it is not used. Leave it empty and there is no peak limit."
  peak_soc_floor:
    name: "Limits · Peak reserve (%)"
    description: "Peak limiting will not draw the battery below this level, but ordinary use can. Derived value, leave it unchanged."
@@ -146,7 +146,7 @@ configuration:
    description: "De add-on mikt zoveel onder de piekgrens, zodat een kwartier dat iets hoger uitvalt er toch onder blijft."
  peak_forecast_entity:
    name: "Grenzen · Entiteit kwartierprognose (W)"
-    description: "Optioneel. Een prognose van het gemiddelde vermogen in het lopende kwartier. Laat u dit leeg, dan is er geen piekgrens."
+    description: "Optioneel. Een prognose van het gemiddelde vermogen in het lopende kwartier, in W of kW. Zonder vermogenseenheid wordt ze niet gebruikt. Laat u dit leeg, dan is er geen piekgrens."
  peak_soc_floor:
    name: "Grenzen · Piekreserve (%)"
    description: "De piekbeperking laat de batterij niet onder dit niveau zakken, gewoon verbruik wel. Afgeleide waarde, laat ze ongewijzigd."
Author	SHA1	Message	Date
glenn schrooyenandClaude Opus 5.5	a4a7ff695e	DOC-8 review: say the field guide PDF is rebuilt, still v0.1.0 checks / unit (pull_request) Successful in 40m29s Details infra-pin / fresh (pull_request) Successful in 1s Details README and FIELD-GUIDE.md said the PDF was not rebuilt; it was, on 2026-10-17. Its content is still v0.1.0 and outdated. Rewrap the DOCS.md intro line. Text only. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-17 03:27:35 +02:00
glenn schrooyenandClaude	b8b357b552	DOC-8: update test_mqtt digest pins for the new manufacturer string infra-pin / fresh (pull_request) Successful in 2s Details checks / unit (pull_request) Successful in 40m23s Details Both digests hash the discovery payloads, which carry the manufacturer. Values regenerated; nothing else changes. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-17 02:16:34 +02:00
glenn schrooyenandClaude	4584502adb	DOC-8: HA manufacturer says EMS port; rebuild field guide PDF checks / unit (pull_request) Failing after 41m1s Details infra-pin / fresh (pull_request) Successful in 1s Details Device id (identifiers) is unchanged; only the display text moves. test_mqtt baseline updated for that one string. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-17 01:53:33 +02:00
glenn schrooyenandClaude	fd08090360	DOC-8: say EMS port, not meter bus checks / unit (pull_request) Successful in 40m22s Details infra-pin / fresh (pull_request) Successful in 1s Details The add-on is wired to the inverter's RS485 port labelled EMS. Reword README, DOCS, config description, field guide source and HTML. The HA manufacturer string is left unchanged: test_mqtt pins it byte for byte. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-17 01:09:42 +02:00
admin	c67d0cdb67	Merge pull request 'Release 0.18.2' (#199 ) from release-0.18.2 into release/1.0 infra-pin / fresh (push) Successful in 1s Details checks / unit (pull_request) Successful in 40m25s Details infra-pin / fresh (pull_request) Successful in 1s Details	2026-10-16 22:31:09 +02:00
glenn schrooyenandClaude Opus 5.5	ea77c63552	CHANGELOG 0.18.2: say how a no-unit forecast shows; who needs a look checks / unit (pull_request) Successful in 40m19s Details infra-pin / fresh (pull_request) Successful in 1s Details Review of PR #199. "Warns once" left out the Home row and the Settings note, which stay while the fault lasts. "No option change" now points to the one case that needs a look. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-16 21:48:05 +02:00
glenn schrooyenandClaude	75f4359aac	Release 0.18.2 checks / unit (pull_request) Successful in 40m23s Details infra-pin / fresh (pull_request) Successful in 2s Details Version bump to 0.18.2 and a CHANGELOG section for SAFETY-118 (#198). Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-16 20:54:29 +02:00
admin	09a54859f2	Merge pull request 'SAFETY-118: read the peak forecast in its own unit' (#198 ) from SAFETY-118 into release/1.0 infra-pin / fresh (push) Successful in 1s Details	2026-10-16 19:53:54 +02:00
glenn schrooyenandClaude	7d52dc1c0f	SAFETY-118: a missing peak forecast reading keeps the unit-fault state infra-pin / fresh (pull_request) Successful in 1s Details checks / unit (pull_request) Successful in 40m19s Details A missing reading (unavailable, unknown, failed read, non-numeric) no longer clears the fault: no false 'used again' INFO, no repeated WARNING, the status row stays. Adds pv.has_reading(). DECISIONS wording fixed; tests added. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-16 18:19:44 +02:00
glenn schrooyenandClaude	f0b4e68e49	SAFETY-118: read the peak forecast in its own unit checks / unit (pull_request) Successful in 40m20s Details infra-pin / fresh (pull_request) Successful in 1s Details peak_forecast_entity is read through pv.watts() by its unit_of_measurement; the abs(raw) < 50 size rule is gone. No unit or a non-power unit means the forecast is not used, with one warning per cause. The validator no longer claims a kW sensor is taken as watts for this entity. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-16 16:00:34 +02:00
admin	780c1f67bb	Merge pull request 'Release 0.18.1' (#196 ) from release-0.18.1 into release/1.0 infra-pin / fresh (push) Successful in 1s Details checks / unit (pull_request) Successful in 40m21s Details infra-pin / fresh (pull_request) Successful in 1s Details Release 0.18.1	2026-10-16 13:46:24 +02:00
glenn schrooyenandClaude Opus 5.5	0ee3c3468b	Release 0.18.1 review: CHANGELOG text fixes checks / unit (pull_request) Successful in 40m18s Details infra-pin / fresh (pull_request) Successful in 1s Details SAFETY-118 line: the add-on already turns a kW forecast into W (main.py read_inputs, abs(raw) < 50), so only the settings-page warning is wrong. Rounding note and rows: the new settings page; UX-44 also changed one DOCS line. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-16 13:03:37 +02:00
glenn schrooyenandClaude	dec8e7b69f	Release 0.18.1 checks / unit (pull_request) Successful in 40m39s Details infra-pin / fresh (pull_request) Successful in 2s Details Version 0.18.0 -> 0.18.1 (patch, no option changes) and a closed 0.18.1 CHANGELOG section. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-16 12:08:58 +02:00
admin	8e02a0928b	Merge pull request 'UX-44: contract label says every 15 minutes, not hourly' (#195 ) from UX-44 into release/1.0 infra-pin / fresh (push) Successful in 2s Details	2026-10-16 06:59:19 +02:00
glenn schrooyenandClaude	1bd2ba1397	UX-44: contract label says every 15 minutes, not hourly checks / unit (pull_request) Successful in 40m37s Details infra-pin / fresh (pull_request) Successful in 2s Details The settings page contract choice read "Dynamic, hourly" / "Dynamisch, per uur". Belgian day-ahead prices are per 15 minutes since 2025-10-01. Now "Dynamic, every 15 minutes" / "Dynamisch, per kwartier". One DOCS line that called dynamic pricing "hourly pricing" now says dynamic. test_ui_text.py part 7 fails if either label says hourly again. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-16 05:26:32 +02:00
admin	b703aaa6d5	UX-43: tariff prices take five decimals (#194 ) infra-pin / fresh (push) Successful in 2s Details	2026-10-16 04:30:09 +02:00
UX-43 review	92a3d3400a	UX-43 review: plain wording for the DOCS line on five-decimal prices checks / unit (pull_request) Successful in 40m23s Details infra-pin / fresh (pull_request) Successful in 1s Details Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-16 03:47:28 +02:00
glenn schrooyenandClaude	8c954acb36	UX-43: keep each field description to one sentence (STYLE) checks / unit (pull_request) Successful in 40m31s Details infra-pin / fresh (pull_request) Successful in 1s Details Co-Authored-By: Claude <noreply@anthropic.com> Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Bdqr7PiaHcBnzPituu2yTC	2026-10-16 02:29:31 +02:00
glenn schrooyenandClaude	8f72ab00d7	UX-43: tariff price fields take five decimals ui.fixed_price, ui.day_price, ui.night_price and ui.periods step 0.00001 EUR/kWh (the period editor's price input took a hard-coded 0.01). A short c-EUR/kWh hint is added to the four field descriptions, EN and NL. Co-Authored-By: Claude <noreply@anthropic.com> Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Bdqr7PiaHcBnzPituu2yTC	2026-10-16 01:40:40 +02:00
admin	a0e091cca3	Release 0.18.0 (#192 ) infra-pin / fresh (push) Successful in 1s Details checks / unit (pull_request) Successful in 40m22s Details infra-pin / fresh (pull_request) Successful in 1s Details	2026-10-15 23:17:10 +02:00
glenn schrooyenandClaude Opus 5.5	44cd1fde2a	Release 0.18.0: review fixes to the CHANGELOG section checks / unit (pull_request) Successful in 40m17s Details infra-pin / fresh (pull_request) Successful in 2s Details OPT-15 and OPT-20 rows: the retry is two inverters only, and OPT-20 no longer claims no other battery moves. SAFETY-111 added to known issues. MAINT-19 marked as from code reading. Wording fixes in the SAFETY-112 and MAINT-11 rows, UX-38 and the price service note. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-15 22:35:47 +02:00
Claude	e7c83d3a57	Release 0.18.0: version bump and CHANGELOG section checks / unit (pull_request) Successful in 40m26s Details infra-pin / fresh (pull_request) Successful in 1s Details config.yaml version 0.17.0 -> 0.18.0 and a closed ## 0.18.0 CHANGELOG section for OPT-17, SAFETY-112, OPT-15, MAINT-11, PRICE-01 and OPT-20. No code change. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-15 21:37:17 +02:00
adminandClaude Opus 5.5	004f859ec7	Merge PR #191 : OPT-20 a refusal probe asks only PROBE_W, not the whole trade infra-pin / fresh (push) Successful in 1s Details Reviewed at `ccdb9dc`: full suite 78/78, CI run 1289 green, identity equal to `bf9019e`. Review: PR #191 comment 2185. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-10-15 14:56:17 +02:00
glenn schrooyenandClaude	ccdb9dc17a	OPT-20: test_opt20 points at MUTATION.md (test_bytecode part 5) checks / unit (pull_request) Successful in 40m25s Details infra-pin / fresh (pull_request) Successful in 1s Details Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-15 13:12:22 +02:00
glenn schrooyenandClaude	06339a2446	OPT-20: a refusal probe asks only PROBE_W, not the whole trade checks / unit (pull_request) Failing after 40m47s Details infra-pin / fresh (pull_request) Successful in 1s Details The probe's trade was sized at the probing battery's rating. Its SAFETY-73 pin held 200 W of that, and the allocator put the rest on a battery with no mode range (one in a maintenance phase): TEST-23 OPT-15 g7, -2300 W on A in its maintenance charge for 34 s, grid 3100 W. modetrade.decide() now asks PROBE_W for the probing battery plus what the other chargers would have had without the probe; size() adds the probe on top of their share, or files PROBE_W exactly, or nothing. _mode_trade() carries the pin only on the ticks it is filed, so the window's first and last cycle do not move another battery's share either. Modes off and modes on with no probe are unchanged (`bf9019e` digests). Tests: test_opt20.py. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-15 13:08:59 +02:00
admin	bf9019e70d	Merge PR #190 : PRICE-01 price_service provider, fallback chain and gate infra-pin / fresh (push) Successful in 1s Details	2026-10-14 22:47:43 +02:00