eaton-ats16-nm2-mib.c: fix mul coef. for ambient.temperature and humidity

[FrancoisRegisDegott-eaton]

No description provided.

[aquette]

Lgtm. Would be worth testing on more than 1 unit to be sure

[jimklimov]

Does this apply to all devices? IIRC at least in some other drivers, we had a problem of multipliers adjusted between firmware versions, so we have to either require that hardware is flashed to newest FW, or to adapt in the driver to devices in the field, by some mapping functions etc. like "if (version < X) then ... else..."

I do hope this one is not a case like that? :)

[FrancoisRegisDegott-eaton]

I agree. But its not easy to check regression related to fw/MIB version...
Here is the upsc fingerprint for the 2.1.5 fw... (useless info removed)
ambient.humidity: 4.50
ambient.humidity.high: 90
ambient.humidity.low: 10
ambient.temperature: 2.6
ambient.temperature.high: 80
ambient.temperature.low: 0
device.mfr: EATON
device.model: Eaton ATS 208V 20A
device.part: EATS220
device.serial: GA07H14138
device.type: ats
driver.name: snmp-ups
driver.version: 2.7.4.1
driver.version.data: eaton_ats16_nm2 MIB 0.18
driver.version.internal: 1.16
ups.firmware: 00.00.0010
ups.firmware.aux: 2.1.5
ups.status: OL

[jimklimov]

For the data dump above, is the humidity really "4.5%" as in a desert (and not a moderate 45% in the range okay for electronic devices -- too dry means lots of static!), and temperature a freezing "2.6" not "26" degrees (Celsius, I suppose)?

Combining the two... many people died in Sahara unexpectedly freezing at nights, not hiding in shadows during the day, but those nights also allowed them to collect condensed water so it was not too dry at that time ;)

[jimklimov]

Also, whatever coefficient is used, shouldn't "high" and "low" use the same?

I wonder if those high watermarks can tell us whether to divide by 10 (something over 100% certainly should be, at least...)

[jimklimov]

And regarding FW/MIB versions, can you get in touch with HW team or their changelogs, I suppose there would be a couple of values (or ranges) in the end to check for?..

[jimklimov]

Per comments above, seems to me that additional work is needed here. Or please prove me wrong :)

[aquette]

hereby confirming that Jeff's above dump is on ATS16 with NM2 and EMP002 backward compat (exposed like the legacy EMP001).
With Jeff's patch, we do well have the right values.
Also confirming that ATS16 with EMP001 are not impacted:
ambient.humidity: 31.10
ambient.humidity.high: 90
ambient.humidity.low: 5
ambient.temperature: 22.8
ambient.temperature.high: 40
ambient.temperature.low: 5
device.mfr: EATON
device.model: Eaton ATS 230V 16A
device.part: EATS16N
...
driver.version.data: eaton_ats16_nmc MIB 0.18

Finally confirming that the new Eaton Sensor MIB exposes temperature and humidity in "tenthOfDegXXX" (Kelvin, Celsius or Farhenheit), as per TemperatureUnitType OID