[Nut-upsuser] blazer_usb MEC0002 problem Fry's Electronics (Turbo-X) [HID PDC?]
Charles Lepple
clepple at gmail.com
Thu Sep 8 12:42:31 UTC 2016
On Sep 8, 2016, at 8:14 AM, Vassilis Virvilis <vasvir at iit.demokritos.gr> wrote:
>
> Hi again,
>
> Thanks for the prompt reply.
>
> Just for the record i am running debian unstable/sid with nut 2.7.4-3
>
Thanks, that is useful.
> 0.447725 refresh_report_buffer: expected 13 bytes, but got 8 instead
This is the first sign of trouble.
> 0.447736 Report[err]: (8 bytes) => 08 03 4d 00 45 00 43 00
Note that for the first HID query, it returned a USB string descriptor for "MEC" (08 bytes, 03 == string, rest is 16-bit Unicode).
> 0.447745 Path: 00860004.Flow.FlowID, Type: Feature, ReportID: 0x01,
> Offset: 0, Size: 4, Value: 3
Also, there is an 0086 Usage Page, but it is often confused with the 0084 page. (If NUT reports a numeric Usage, then it is probably not one of the standard numbers.) If this device can be made to return valid data, some additional coding would be needed to map these to the proper pages.
> 0.452728 Report[get]: (16 bytes) => 10 03 4d 00 45 00 43 00 30 00 30 00 30 00 32 00
Similar to before, this is a string descriptor ("MEC0002")
> 0.623733 Report[get]: (9 bytes) => 16 03 55 00 50 00 53 00 20
Also returned a fragment of a string descriptor ("UPS ").
It is possible that the blazer_usb driver initialization has put the UPS into a bad state. If you can schedule some downtime, you might want to try usbhid-ups again after powering the UPS down completely and unplugging it for a little while. (Often, a few seconds after the output is off and the AC has been removed, you will hear relays disengage.)
The other PDC HID UPSes would usually report some changes every few seconds. If you do not see any changes after "rebooting" the UPS (something other than "Got 0 HID objects..."), try toggling the AC power (recommend using a circuit breaker or power strip to keep the ground and neutral attached) to see what an "on battery" event looks like.
--
- Charles Lepple
https://ghz.cc/charles/
More information about the Nut-upsuser
mailing list