[Nut-upsuser] CyberPower DX600E won't switch up after power

[Franck]

On 2013-01-10 21:52, Franck wrote:
> On 2012-12-05 21:38, Arnaud Quette wrote:
>> 2012/12/1 Franck <franck at secretfatty.net [4]>
>>
>>>> Date: Mon, 26 Nov 2012 19:34:28 +0100
>>>> From: Arnaud Quette <aquette.dev at gmail.com [1]>
>>>> To: Franck <franck at secretfatty.net [2]>
>>>> Cc: nut-upsuser at lists.alioth.debian.org [3]
>>>> Subject: Re: [Nut-upsuser] CyberPower DX600E wont switch up after
>>>> power
>>>>
>>>> right. we need to monitor the UPS while its shutting down...
>>>
>>> Well Id like to try that; but Im 2000km from my UPS and It seems to
>>> be problematic for me to have the test done.
>>>
>>> But anyway I just got this reply to my quite random inquiry to
>>> CyberPower (wrong country) support:
>>>
>>> "I can only make vague guesses because I have never seen the
>>> product you have, and I am not familiar with the software you used
>>> to generate that data.  The following values stand out to me.
>>>
>>> battery.voltage: 4.7
>>>
>>> battery.voltage.nominal: 12
>>>
>>> ups.load: 31
>>>
>>> If I am interpreting them correctly your battery should be at 12
>>> volts, but it is only at 4.7 volts?
>>>
>>> And the UPS load is 31%?
>>>
>>> If the battery is at 4.7 volts it will not pass the power on
>>> self-test. It needs to be somewhere above 10 volts (Perhaps 10.5 or
>>> 11)before it will pass the self-test and let the ups turn on.
>>>
>>> Other possibilities.
>>>
>>> If you have the computers set to auto start when power is restored,
>>> they will turn on simultaneously, if there has been a power loss
>>> that significantly drained the batteries, they will have very 
>>> little
>>> energy when the power is restored.  The power on self test checks
>>> the CyberPower’s ability to run on battery by stopping access to
>>> wall power and forcing the UPS on to battery power.  If the
>>> batteries are very low and the auto startup of the computers hits
>>> while they are being tested then the load of the computers on the
>>> weak battery could cause the voltage to drop and the self-test 
>>> would
>>> fail.
>>>
>>> One or both of your computers has an Active PFC power supply and
>>> your UPS is not a sine wave ups. If you are not familiar with this
>>> problem, just search the internet on the terms “active pfc” and
>>> “sine”
>>>
>>> The battery in the UPS could be defective.
>>>
>>> Again.  I do not know the product you are asking about so I
>>> can’t provide an accurate diagnosis.  I can only suggest
>>> possibilities.
>>>
>>> "
>>>
>>> So if the guy os right and this might be a battery problem.
>>
>> hem, ups.load is the load on the UPS output.
>> what you were looking for is probably battery.charge.
>>
>> battery.voltage seems indeed wrong, but to me, thats another issue.
>>  yours is with the restart function, that is tied to the USB/HID 
>> data
>> I mentioned before.
>> monitoring these counting down should help understanding how these
>> actually behave.
>>
>> its true that being 2000kms away doesnt make things easy.
>>  but there each problem has at least 1 solution:
>> instead of doing a full reboot cycle, we can just monitor for 10
>> seconds, and cancel the procedure
>>
>> - stop NUT after the reboot,
>>  - restart the driver in debug more (/lib/nut/usbhid-ups -DDDDD 
>> myups)
>>  and upsd (simply type "upsd" as root) in another term
>>  - then execute "upscmd -u ... -p ... myups shutdown.return"
>>  CAUTION: wait no more than 10-15 seconds!!! Otherwise, your server
>> will crash!!!
>> possibly monitor your unit with upsc
>> - then execute "upscmd -u ... -p ... myups shutdown.cancel" (mention
>> -u and -p to avoid loosing a few seconds!)
>>  -then Ctrl+C in the driver term.
>>  and send back the driver output, in compressed form.
>>  you can then restart everything as usual...
>>
>
> I coudln't perform the test, so I bought a brand new  USV Aiptek
> Powerwalker VI 650 LCD.
> And guess what...  I have the same exact behavor:
> It doesn't switch up after beeing unloaded; when the power is back.
> Any idea?
>

Arnaud: does it makes sense to perform the exact same test, now ?

---
http://secretfatty.net