2.8.2_1 blazer_usb, nutdrv_qx

dennypage

@LanterIT said in NUT Package (2.8.1 and above):

From my point of view, it is not possible for the UPS to measure just a single battery cell. There are three 12V 9.6Ah batteries in series in it, with two cables, plus/minus and no additional measuring lines.

Fair point. Nominal voltage should be 36 volts.

The blazer_usb driver run shows this:

  1.259834     [D3] send: Q1
  1.494125     [D3] read: (225.7 225.7 230.1 047 50.0 2.18 28.0 00000001
  1.494469     [D2] Status read in 1 tries
  1.494497     Supported UPS detected with megatec protocol
  2.631643     [D3] flush: Timeout
  2.639009     [D3] send: F
  2.749987     [D3] read: #230.0 004 036.0 50.0
  2.750183     [D2] Ratings read in 1 tries

And the nutdrv_qx megatec driver run shows this:

   2.489697     [D3] send: Q1
   2.723945     [D3] read: (226.7 226.7 230.4 049 49.9 2.17 28.0 00000001
   3.798980     [D3] flush: Timeout (-7)
   3.805445     [D3] send: F
   3.915503     [D3] read: #230.0 004 036.0 50.0

Looking at the strings being sent by the UPS, the different runs match well and are within a normal expected variance for the values:

Q1:

input.voltage: 225.7 / 226.7
input.voltage.fault: 225.7 / 226.7
output.voltage: 230.1 / 230.4
ups.load: 47 / 49
input.frequency: 50.0 / 49.9
battery.voltage: 2.18 / 2.17
ups.temperature: 28.0
ups.beeper.status: on

F:

input.voltage.nominal: 230.0
input.current.nominal: 4
battery.voltage.nominal: 36.0
input.frequency.nominal: 50.0

The nominal voltage of 36 matches the physical description of the batteries. The actual reported battery.voltage (2.18/2.17) appears completely fictional. But apparently this is a known/expected behavior with these types of UPSs. See here and here for discussion by one of the NUT team leaders.

FWIW, I am still looking at the code to see what actually ended up in the various releases.

dennypage

@LanterIT Can you confirm which version of nut-devel you used in the later debug tests? If I understand correctly, it was 2024.01.03, but I just want to double check.

The tests I am asking about are:

blazer_usb (post 249)
nutdrv_qx (post 251 & 256)

Thanks

LanterIT

@dennypage

correct, all posts since 241 are with APU2 and fully updated packages

dennypage

@LanterIT said in NUT Package (2.8.1 and above):

all posts since 241 are with APU2 and fully updated packages

Thank you

dennypage

@LanterIT If you have time, would you mind re-running those two tests with a higher debug level (-DDDDDD) please?

LanterIT

@dennypage

Extra Argument:
vendorid=06da
productid=0005
subdriver = "phoenix"
protocol = "megatec"
novendor

nutdrv_qx_202402232122.txt

blazer_usb_202402232127.txt

dennypage

@LanterIT Sorry, can you put the "battery_voltage_reports_one_pack" back in the config please?

LanterIT

@dennypage
only nutdrv_qx output, because blazer_usb does not support the parameter:
nutdrv_qx_202402232234.txt

dennypage

@LanterIT Attached is a nutdrv_qx executable with additional debug information added. Please run let it run for at least 100 seconds before killing it.

Note that I don't have a compatible UPS, so I am not actually able to test the executable. Please do excuse me if I've messed something up.

Thanks

nutdrv_qx_denny_dbg.zip

LanterIT

@dennypage

I am happy when someone takes care of the problem and I can do my part.
nutdrv_qx_denny_dbg_202402252046.txt

dennypage

@LanterIT LOL. I love it when people are willing to help me solve problems.

One more time please...

nutdrv_qx_denny_dbg2.zip

LanterIT

@dennypage

Here you go:
nutdrv_qx_denny_dbg2_202402252237.txt

dennypage

@LanterIT said in NUT Package (2.8.1 and above):

nutdrv_qx_denny_dbg2_202402252237.txt

Did you notice the reported voltage and charge?

LanterIT

@dennypage

It was a bit late last night, so I only dumped the output to the log file.

After a little more than 30 seconds, the battery voltage and battery charge display will change to the correct values.

Does this have an influence on the behavior as long as the status is OL?

EDIT: It also runs on the productive device!

dennypage

@LanterIT said in NUT Package (2.8.1 and above):

Does this have an influence on the behavior as long as the status is OL?

By default, NUT doesn't take action until the UPS declares a low battery, and battery voltage is reported as informational only. This is probably one reason why no-one has noticed the bug. Runtime is the same, although for most of the nutdrv_qx protocols, runtime isn't reported at all.

I'll put up a bug report and PR for the NUT folk later today or tomorrow. I don't know how long it will take for this to flow though to an a release on pfSense. In the interim you can use the last executable--the extra debug statements won't hurt anything.

Thank you for your patience and perseverance.

dennypage

@LanterIT NUT issue #2325.

dennypage

@dennypage said in NUT Package (2.8.1 and above):

@LanterIT NUT issue #2325.

The fix for accessing battery_voltage_reports_one_pack has been merged, and I expect it to be included in nut-devel snapshot.

I was hoping that the flag would simply be deleted, but the NUT maintainer seemed resistant to that. You will need to maintain the flag in your Extra Augments to driver section.