On Fri, Mar 28, 2014 at 8:12 AM, Charles Lepple <clepple at gmail.com> wrote: <<Snip>>> > I'll be honest, I haven't really done much with the CGI scripts lately. For debugging values, it is probably better to be closer to the driver. I'd recommend using "upsc" or the driver logs. > > To that end, the .ac.txt and .bat.txt files both show 24.0 V: > > 0.081935 Entering libusb_get_report > 0.082018 Report[get]: (2 bytes) => 09 f0 > 0.082029 Path: UPS.PowerSummary.ConfigVoltage, Type: Feature, ReportID: 0x09, Offset: 0, Size: 8, Value: 24 > 0.082035 Entering libusb_get_report > 0.082115 Report[get]: (2 bytes) => 0a f0 > 0.082124 Path: UPS.PowerSummary.Voltage, Type: Feature, ReportID: 0x0a, Offset: 0, Size: 8, Value: 24I know there are two 12V batteries in series that provide 24V which is what is being shown, but an actual 24V for new fully charged batteries seems low to me. Between that and the fact that there are no tenths of a volt shown and the battery is shown as 24.0 both on AC and on battery makes me think we are not seeing the actual battery battery voltage but some nominal or default value. I do not think the battery voltage is actually being tracked. ???> > To your other point about refreshing the CGI display, you could try adding a tag to reload the page at a specific interval. I'm not sure if that forces the images to be reloaded, or if the images need a "no-cache" header in their HTTP headers. > > I tend to use collectd to monitor NUT statistics over time: > > https://collectd.org/wiki/index.php/Plugin:NUTI will take a look at that. Thanks> > -- > Charles Lepple > clepple at gmail > > > > _______________________________________________ > Nut-upsuser mailing list > Nut-upsuser at lists.alioth.debian.org > http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser
On Mar 28, 2014, at 9:08 AM, Bill S wrote:> On Fri, Mar 28, 2014 at 8:12 AM, Charles Lepple <clepple at gmail.com> wrote: > > <<Snip>> >> >> I'll be honest, I haven't really done much with the CGI scripts lately. For debugging values, it is probably better to be closer to the driver. I'd recommend using "upsc" or the driver logs. >> >> To that end, the .ac.txt and .bat.txt files both show 24.0 V: >> >> 0.081935 Entering libusb_get_report >> 0.082018 Report[get]: (2 bytes) => 09 f0 >> 0.082029 Path: UPS.PowerSummary.ConfigVoltage, Type: Feature, ReportID: 0x09, Offset: 0, Size: 8, Value: 24 >> 0.082035 Entering libusb_get_report >> 0.082115 Report[get]: (2 bytes) => 0a f0 >> 0.082124 Path: UPS.PowerSummary.Voltage, Type: Feature, ReportID: 0x0a, Offset: 0, Size: 8, Value: 24 > > I know there are two 12V batteries in series that provide 24V which is > what is being shown, but an actual 24V for new fully charged batteries > seems low to me. Between that and the fact that there are no tenths > of a volt shown and the battery is shown as 24.0 both on AC and on > battery makes me think we are not seeing the actual battery battery > voltage but some nominal or default value. I do not think the battery > voltage is actually being tracked. ???"24.0" is slightly optimistic on my part ;-) but the value "f0" is hex for 240, so the UPS is capable of reporting increments of 0.1V. We used to print the floating point values with "%.6f" or similar, which is IMHO slightly more misleading since nobody is measuring with microvolt resolution. That's been on my list to fix for some time now. ConfigVoltage (battery.voltage.nominal) is often an even multiple of 12V, but you're right, the actual voltage should be higher for fully-charged batteries. That voltage may be hiding somewhere else in the report. If there is a battery test command, that (plus the -DDD logs) would be a good way to find where the actual battery voltage is hiding. -- Charles Lepple clepple at gmail
On Fri, Mar 28, 2014 at 9:39 AM, Charles Lepple <clepple at gmail.com> wrote: <<Snip>>>>> >>> I'll be honest, I haven't really done much with the CGI scripts lately. For debugging values, it is probably better to be closer to the driver. I'd recommend using "upsc" or the driver logs. >>> >>> To that end, the .ac.txt and .bat.txt files both show 24.0 V: >>> >>> 0.081935 Entering libusb_get_report >>> 0.082018 Report[get]: (2 bytes) => 09 f0 >>> 0.082029 Path: UPS.PowerSummary.ConfigVoltage, Type: Feature, ReportID: 0x09, Offset: 0, Size: 8, Value: 24 >>> 0.082035 Entering libusb_get_report >>> 0.082115 Report[get]: (2 bytes) => 0a f0 >>> 0.082124 Path: UPS.PowerSummary.Voltage, Type: Feature, ReportID: 0x0a, Offset: 0, Size: 8, Value: 24 >> >> I know there are two 12V batteries in series that provide 24V which is >> what is being shown, but an actual 24V for new fully charged batteries >> seems low to me. Between that and the fact that there are no tenths >> of a volt shown and the battery is shown as 24.0 both on AC and on >> battery makes me think we are not seeing the actual battery battery >> voltage but some nominal or default value. I do not think the battery >> voltage is actually being tracked. ??? > > "24.0" is slightly optimistic on my part ;-) but the value "f0" is hex for 240, so the UPS is capable of reporting increments of 0.1V. We used to print the floating point values with "%.6f" or similar, which is IMHO slightly more misleading since nobody is measuring with microvolt resolution. That's been on my list to fix for some time now.I agree. Three significant figures is the right number. More as you say would be rather meaningless but only two significant figures really does not tell you much.> > ConfigVoltage (battery.voltage.nominal) is often an even multiple of 12V, but you're right, the actual voltage should be higher for fully-charged batteries. That voltage may be hiding somewhere else in the report. > > If there is a battery test command, that (plus the -DDD logs) would be a good way to find where the actual battery voltage is hiding.In effort to find the actual battery voltage I ran /usr/local/ups/bin/upsc from the command line (results attached) and it also reports a battery of voltage of 16.0. I also tried unloading NUT and running CyberPower's latest PowerPanel Linux software but the only thing I could see to do with that software is run "pwrstat -test" and "pwrstat -status" and neither one resulted in any output of battery voltage. If there is something else I should be doing along those lines that I missed, please let me know. I guess if there is any good news it is that it appears that the problem is not with the CGI scripts for both the output AC voltage and battery voltage. I also noted that when running the PowerPanel software that the reported output AC voltage matched the input AC voltage and that the input AC voltage went to zero when on battery. I would take that to mean that the UPS has those values available somewhere and so perhaps it has the right battery voltage as well if they can be accessed.> > -- > Charles Lepple > clepple at gmail > > > > > _______________________________________________ > Nut-upsuser mailing list > Nut-upsuser at lists.alioth.debian.org > http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser-------------- next part -------------- A non-text attachment was scrubbed... Name: upsc_out.txt.gz Type: application/x-gzip Size: 454 bytes Desc: not available URL: <http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20140329/0478a98d/attachment.bin>
I asked CyberPower why their PowerPanel software did not show battery voltage and the response was that battery voltage is not supported in this model. They were surprised that NUT reported 16.0 or any value at all. It makes you wonder where the 16.0 comes from. In any case it appears that the only current problem is the incorrectly reported AC output voltage.
It would appear that there is nothing that can be done to correct the output voltage reading. Is there anyone in this group that would have a suggestion on alternative software for remote monitoring of a UPS installed on a distant server?