[Nut-upsdev] Driver for Cyberpower PR2200

[Arjen de Korte]

Russ Romano wrote:

> I'm a bit curious about how much of the OEM functionality I can expect
> the final driver to support.

Most likely everything you already have in your OEM driver and probably
something extra goodies too. :-)

> Having not gotten it working, I didn't go
> through the exercise of fully configuring the clients, etc.

Wait with that, until we have confirmed that everything is working.

> Using the OEM linux software (and I'm sure the windows software is the
> same) it is possible to set many aspects of the UPS's behavior.  Here is
> a capture of the "information" dump from the OEM linux software:
>
> UPS model : PR2200
> UPS version : 1.100
> output format : 120 V
> battery format :
> buzzer : [on],off
> cold start : on,[off]
> enable status record : [on],off
> shutdown system after power fails : [on],off
> enable auto save function :[on],off
> status record interval : 10 minute 0 seconds.
> power failure causes system shutdown after : 30 seconds.

We don't do this by default (we rather wait until the battery is empty),
but you can script this quite easily through 'upssched.conf' and its
assortment of helper functions.

> time to wait system complete shutdown : 2 minute 0 seconds
> communication port : None,Auto,ttyS0,[ttyS1]

We don't do auto detection on serial ports. Much too risky.

> high input voltage configuration :
> [138],139,140,141,142,143,144,145,146,147
> low input voltage configuration : 88,89,90,91,92,93,94,95,96,[97]
> percentage of battery critical capacity configuration
> :[25],30,35,40,45,50,55,60
> battery mode output voltage configuration : 110,[120],130

Assuming that you can provide us with enough information what your
existing software is sending to the UPS, I'm quite confident that we
will be able to include the same functionality in the 'cyberpower'
driver and (consequently) into NUT.

> Is it safe to assume that the UPS is simply reporting what it monitors
> and the software handles the rest?

No. Usually, you configure the levels when to switch over to battery and
when to report a critical (low battery) condition. Based on that
information, the UPS will decide when to act upon this. The serial link
is much too slow for that. Note that your UPS communicates via a 1200
baud link, so each character will take at least 8 ms to send. So when
the input power goes away, it will take something in the order of 100ms
or more for NUT (or any other piece of software on your computer) to
realize that it needs to switch over to battery. So the decision when to
switch to battery or raise an alarm is made by the UPS, not the OEM
software, although the latter tells the UPS the levels when to do this.

> Or is the "shutdown system after power fails" a setting in the UPS itself?

Most likely, yes.

> In order for my machines to come back on when power recovers, it's
> necessary for the UPS to have completely powered off after all the
> computers shutdown, and then power back up after the line power
> returns (and batteries are sufficiently recharged).

Yes.

> With the over and under voltage conditions, it is possible for this UPS
> to generate 120V output, while keeping it's batteries charged with the
> under or over voltage line power.  Will I loose such functionality with
> NUT, or is that a hardware configuration that I can set with the OEM
> software and then switch to NUT?

You will keep that functionality. It's a fair assumption that the
protocol is similar to the text based cyberpower protocol as we
implemented in the 'powerpanel' driver. We could do it there, so I have
no reason to believe we can't do it here too. It's a matter of decoding
the protocol. Since I doubt we have the protocol somewhere in writing
(which would obviously be much more convenient), we rely on you sniffing
the communication between host and UPS.

Best regards, Arjen