Update data fields of the “energy” and “voltage” types

[Guzz-T]

• Change the scaling factor of 'ID_Waermemenge' parameters from 0.1 to 0.01
• Change the scaling factor of the voltage data type to 0.01

Relates to #198

[github-actions]

Coverage Report

File	Stmts	Miss	Cover	Missing
luxtronik
__main__.py	25	25	0%	3–66
datatypes.py	454	1	99%	147
discover.py	58	1	98%	63
luxtronik/cfi
interface.py	175	3	98%	167–168, 256
luxtronik/definitions
__init__.py	208	1	99%	216
luxtronik/scripts
__init__.py	35	5	86%	7–16, 51
discover_shi.py	60	60	0%	8–98
dump_cfi.py	14	6	57%	18–25, 29
dump_shi.py	15	6	60%	21–28, 32
performance_shi.py	63	63	0%	7–89
update_screen.py	105	82	22%	6, 20–41, 46–48, 51–61, 64, 70–75, 78, 81–85, 88–92, 95–97, 100–103, 106–111, 115–116, 119–126, 129–136
watch_cfi.py	32	16	50%	36–47, 50–56, 60
watch_shi.py	33	16	52%	38–49, 52–58, 62
luxtronik/shi
interface.py	337	3	99%	438, 519, 527
TOTAL	2436	288	88%

Tests	Skipped	Failures	Errors	Time
381	0 💤	0 ❌	0 🔥	17.971s ⏱️

[Guzz-T]

Can any of you check the values on your heat pump?

[kbabioch]

Code itself looks fine to me, but let's give this some time to verify whether it fixes the issue in the real world :-).

[Guzz-T]

Are there any findings yet?

[rhammen]

@Guzz-T sorry for the late reply. I checked the registers in this PR against the diagnostics dumps collected in the BenPru/luxtronik integration (30 heat pumps from public issue reports plus my own; firmware V1.73–V1.90, V2.88–V2.90, V3.86–V3.92). Result: the 0.01 kWh scale is confirmed for everything here except 1059, which turns out to be firmware-series dependent.

Confirmed 0.01 kWh — parameter equals the matching calculation (kWh/10) × 100, exactly

Parameter	Calculation	Units with data	Example pairs (raw ↔ kWh)
852 ID_Waermemenge_Seit	154 ID_WEB_WMZ_Seit	9 of 9	3449447 ↔ 34494.4, 17613162 ↔ 176131.6, 835590 ↔ 8355.9
854 ID_Waermemenge_Hz	151 ID_WEB_WMZ_Heizung	26 of 27	11960251 ↔ 119602.5, 14545611 ↔ 145456.1, 6449260 ↔ 64492.6 (V1.73)
878 ID_Waermemenge_BW	152 ID_WEB_WMZ_Brauchwasser	24 of 24	2615400 ↔ 26154.0, 3004080 ↔ 30040.8, 1337833 ↔ 13378.3
879 ID_Waermemenge_SW	153 ID_WEB_WMZ_Schwimmbad	5 of 5	1565690 ↔ 15656.9, 436410 ↔ 4364.1, 104840 ↔ 1048.4

(The one exception for 854 is a ratio of 99.4 from a dump where the two blocks were evidently read a few minutes apart.) This holds on all three series. Note that on V3.88+ units calculation 154 reads 0 while 852 keeps counting — the total is apparently no longer mirrored to the calculation there.

• 1136 / 1137 / 1139 (HEAT/DHW/COOLING_ENERGY_INPUT): verified against the controller's own display in "Additional heat generator energy" is off by factor 10 BenPru/luxtronik#752: 1136 = 543197 vs 5431.9 kWh, 1137 = 222335 vs 2223.3 kWh (LD9, V2.88.3); 1139 = 83272 vs 832.7 kWh (LD5 REV, V2.90.0). Independently, in chore(sensor): verify energy-input scaling (1136/1137/1139) against measured data BenPru/luxtronik#734 a regression of 1136/1137 against 151/152 over 19 snapshots gives a marginal COP of 6.4/3.8 at /100 but 0.64/0.38 at /10.
• 1060 ID_Waermemenge_Reset: same magnitude as 852 and identical to it on 10 of 18 units where it is non-zero (differs in both directions on the other 8, purpose unclear). 0.01 kWh by analogy; no direct verification.

1059 ID_Waermemenge_ZWE — scale depends on the controller series

• Series 2 (V2.x): 0.01 kWh as in this PR. An LD9 on V2.88.3 read 147140 against 1471.4 kWh on the controller's web page ("Additional heat generator energy" is off by factor 10 BenPru/luxtronik#752).
• Series 3 (V3.x): 0.1 kWh. On five inverter units (MSW4-16, MSW2-6S, MSW2-9S, LW SEC 2; V3.79–V3.92.3) the ZWE energy divided by the ZWE run time equals the rated element power in parameter 1025 only at 0.1 kWh; at 0.01 kWh it would imply a 0.89 kW heating element (Additional heat generator amount counter is a factor 10 too small since PR #490 BenPru/luxtronik#625).
• Series 1 (V1.x): 1059 reads 0 on every V1 unit in the corpus, so unknown.

So a single scale for 1059 is wrong for one group either way: Energy is off by 10× on V2 units, Energy2 is off by 10× on V3 units (21 of the 30 corpus units are V3). The integration currently handles this with a per-series factor on top of the library, but the library already has what's needed: the since/until fields on definitions are checked against the firmware version in Definitions.build(). I'd propose two entries for 1059:

{"index": 1059, ..., "names": ['ID_Waermemenge_ZWE'], "type": Energy2, "until": '2.99.99', "unit": 'kWh/100', ...},
{"index": 1059, ..., "names": ['ID_Waermemenge_ZWE'], "type": Energy,  "since": '3',       "unit": 'kWh/10',  ...},

with the V1 case (unverified, assumed like V2) covered by the first entry. That would make this the first version-gated parameter definition.

Two more candidates for Energy2, currently Unknown upstream and not in this PR — both display-verified in BenPru/luxtronik#752: 1135 = 410381 vs 4103.8 kWh (cooling heat quantity) and 1138 = 113191 vs 1131.9 kWh (pool energy input).
I'm happy to make PRs for both the 1059 split and the 1135/1138 additions (on top of this one, or separately once it's merged) — or if you'd rather fold them into #264 yourself, that's fine with me too; just say which.

Relates to #198.