[Nut-upsuser] TrippLite SU3000XLCD

[Fairfax, Charles A.]

Thanks for your reply.  I've done a bit more testing and made additional notations below.

-----Original Message-----
From: Charles Lepple [mailto:clepple at gmail.com] 
Sent: Thursday, October 20, 2016 9:06 PM
To: Fairfax, Charles A.
Cc: nut-upsuser at lists.alioth.debian.org
Subject: Re: [Nut-upsuser] TrippLite SU3000XLCD

[please use reply-all to include the list - the headers are not mangled]

On Oct 20, 2016, at 3:22 PM, Fairfax, Charles A. wrote:
>> 
>> I’m having trouble with a TrippLite SU3000XLCD.  I’m using a Raspberry Pi running Raspbian  (wheezy) with NUT 2.6.4. 
>>  
>> I can monitor my older TrippLite UPSs, mostly SU1000RT2Us, with this Pi and a USB to serial adapter.

>Interesting, we don't have that model number listed in our compatibility list[*]. Do they mostly use tripplitesu or tripplite?

Tripplitesu.  
I have a number of older TrippLite UPSs:  SU1000RT2U, SU2200RT2U, and SU3000RT2U without USB ports.  Traditionally I've monitored these with TrippLite's PowerAlert software but I can't get it to work reliably with the Windows 7 systems they support so I'm headed in the direction of using NUT on Raspbian Pi-s.  The tripplitesu driver also works with my newer SU1000XLa and SU1500RTXL2Ua over serial.  Unfortunately I don't get Battery Runtime in my upsstats.cgi output for serial attached UPSs.

>[*] http://networkupstools.org/stable-hcl.html?manufacturer=Tripp%20Lite

>> My first thought with the SU3000XLCD was to use the USB port instead of the serial port as that’s the “modern” thing to do.  When I plugged the USB cable in to the UPS it turned off, dropping the load.  The load is process control instrumentation which, while not damaged, presented some challenges to restart.  So now I’m leery of using the USB port.
>>  
>> With a long history of success monitoring UPSs via serial communication that’s where I am now.  Unfortunately NUT doesn’t see the UPS.
>>  
>> My query is two part:
>> 1)      Is there a reasonable explanation for why plugging my Pi into the UPSs USB port caused it to turn off?  Is there a proper way to connect it that will avoid this?  If so my problem is likely solved if the usbhid-ups driver will talk to my SU3000XLCD.  If I can’t assure my employers that I won’t kill  the process again USB may not be a viable option.

>To be honest, I would have expected the RPi to reboot when inserting the USB cable.

>Was the Pi plugged into a computer or a standalone USB adapter? I wonder if there was a difference in ground potential between the Pi and the UPS. 

The Pi was powered by a standalone AC/USB adapter plugged into the SU3000XLCD so I wouldn't expect a ground potential difference unless one was introduced by the wired network connection or perhaps static from me.  I have submitted a support request via TrippLite's web form asking about the load drop.  I don't have another SU3000XLCD to test.  I have tested plugging and unplugging both USB and serial on a SU1500RTXL2Ua which hasn't yet gone into service with no problems.  

>The entry on the HCL for the SU3000XLCD came from NUT USB testing that Tripp Lite performed in their lab: >http://article.gmane.org/gmane.comp.monitoring.nut.user/8173 . For that reason, I would think it would be reasonable to bother Tripp Lite about this issue, especially if it crops up again.

>> 2)      Connecting to the serial port has not (yet) shut down the UPS.  This is a good start but it would be better if communication happened.  Should I be able to communicate via serial port with this SU3000XLCD with one of these drivers?  Is there another driver that’s known to work with this UPS or can this one be “tweaked” to work?  Can it be as simple as a baud rate mismatch?  I’m not having much luck querying the serial port’s settings.

>Although the old "tripplite_usb" driver seems to be talking to a USB-to-serial chip with a very similar protocol to the "tripplite" driver, I don't think we ever got anywhere on trying to combine the two. Also, since the SU3000XLCD is listed as compatible with the usbhid-ups driver rather than tripplite_usb, I don't know if we can draw any conclusions from the older driver.

>Do you have access to a spare SU3000XLCD? We could try patching the tripplite serial driver to send a protocol query, and work from there. However, I wouldn't want to experiment on the production system.

>Another thing to note: some UPSes don't like switching between the USB and serial ports after they have powered up. I don't remember if we ever confirmed if this affects Tripp Lite hardware, but it is something to be aware of.

The ideal outcome is probably to use the USB interface if for no other reason than that it presumably returns Battery Runtime as with the other UPSs I've tried.  Being currently supported is probably a plus as well.  I guess I'll wait to hear from TrippLite re: my support ticket.  I do have a window of a few weeks while the system (load) is still in development and can with justification do some testing that might cause a shutdown.  My best justification is likely that we won't want any surprises from the UPS when the system goes into production.

Thanks for taking a look at this
-Charles.