Page 1 of 1

Windows 7 reboot issues

Posted: Sun 20 Sep 2015 1:51 pm
by starchild
Hi there,

I've been running Cumulus 24x7 for 7 years and it's been a real gem. I've got quite a few twitter followers for its hourly tweets too! I'm running the cheapo Maplin Watson W8681 weather station which has been surprisingly reliable (especially with an annual squirt of WD40 in the rain gauge to stop spiders nesting!).

However, I do have the problem mentioned in the program readme file regarding the date format issue when rebooting Windows 7. The first obvious sign I get when restarting Cumulus is that the pressure reads zero as soon as the process of fetching stored data from the weather station starts. I then get erratic data collection and the "Processing data" pop-up regularly appears and stays on screen for 10 seconds or so when all stored data has been transferred. At first I was able to get Cumulus to start normally after resetting the region as suggested - although this would take 3 or 4 attempts of complete reboot cycles. Unfortunately after many months of success this work around is no longer able to restore normal operation - and with what seems like almost daily Windows updates requiring a reboot now, I've all but given up using Cumulus.

I've spent many hours scouring through the data files to see where there is a discontinuity but have never found anything wrong. If there is a date format issue I would have expected to see differences in the representation of day/month/year - but every file is consistently presented in UK date format. There's also no zero pressure readings in the files which might be acting as a trigger.

I'm really at a loss now and haven't come across any mention of my particular observations in the forum. Anyone have any ideas?

Re: Windows 7 reboot issues

Posted: Sun 20 Sep 2015 1:55 pm
by steve
Once you've corrected the Windows 7 region/format issue it shouldn't be a problem again, so perhaps you have a different issue. If you zip up the diags folder and attach it to a reply in this thread, I'll take a look.

Re: Windows 7 reboot issues

Posted: Sun 20 Sep 2015 4:01 pm
by starchild
Many thanks Steve.

As I say, the region reset trick worked for a while - even if it took 5 or 6 attempts but I just can't get it to work now. I did find that I previously had more chance of success if I left the region set to US for a few minutes before switching it back to UK.

I have to admit that the machine which Cumulus is running on also runs a variety of applications such as home automation, media server, file server, web server, CCTV server, cloud sync, local backup, etc so there's a chance it may be something to do with these - although I always ensure they're not running when I reboot.

The zip file contains two sets of diags:

"June 2015 Good" - a saved copy from a successful reboot and where Cumulus has been running for a few weeks (months?)
"Sept 2015 Bad" - Today's failed attempt at running Cumulus after a reboot

Re: Windows 7 reboot issues

Posted: Sun 20 Sep 2015 4:51 pm
by steve
There's a problem with your pressure reading. Back in June, when it was working OK, it was giving a sensible absolute reading:

17/06/2015 23:03:26.621 : EWUSB: Read pressure values: rel=1021.4 abs=1006.9 offset =14.5

But now the absolute value is impossibly high, and you've had to use a large negative offset to bring the relative reading into range:

20/09/2015 14:40:03.799 : EWUSB: Read pressure values: rel=1025.6 abs=1215.9 offset =-190.3

But when Cumulus reads the logger entries, it's getting a more reasonable absolute value, but subtracting the large offset makes the relative pressure far too low:

20/09/2015 14:38:04.920 : 14:38:04 EWUSB: Processing history data for = 20/09/2015 14:30:04
20/09/2015 14:38:04.921 : EWUSB: Ignoring bad data: pressure = 820.799926757813
20/09/2015 14:38:04.921 : EWUSB: offset = -190.300048828125

The above corresponds to an absolute pressure of 1011.1 hPa, which is in line with what it was reading in June.

You could try doing a full reset (remove batteries and USB cable for a few minutes), that might get it behaving again. What absolute and relative pressure values are being displayed on the console?

Re: Windows 7 reboot issues

Posted: Sun 20 Sep 2015 10:26 pm
by starchild
That's strange Steve - I see what you mean though.

The display shows 1007.7 for absolute but I'm not sure this device it's capable of displaying a relative pressure? I can't get it by tapping the touch screen anyway.

How would the large negative pressure offset have been entered? I can't see it in the configuration anywhere.

The problem started in February when I upgraded from XP (yes I know, I was a very late adopter but I needed to take a deep breath to migrate my home automation!). Whenever I managed to get Cumulus to display a non zero pressure after reboots and resetting the region format I would always copy the entire Cumulus directory next time I would need to reboot for a Windows update in the hope that it would not be corrupt - that's where the June diags come from. I'm beginning to wonder if I have a failing station and the fact that it would sometimes work after a few attempts at region resets is just a coincidence.

By the way, now that Cumulus is displaying zero pressure, although the application shows the correct readings (e.g. outdoor temp, trend, etc), Cumulus is not showing anything in the last updated time field in the bottom right of the window. However, the clock top right is counting up OK. I can also see that Cumulus isn't writing to the data files. It seems odd that a pressure reading issue would cause that to happen?

Re: Windows 7 reboot issues

Posted: Mon 21 Sep 2015 6:53 am
by steve
From the manual, this is the way you switch between abs and rel pressure on the console (on some models, at least):

Touch the PRESSURE section, + button and – button will be flashing. Touch the + button or –
button to Shift the display between Absolute pressure and Relative pressure.

Cumulus doesn't log data until it is reasonably sure your station is working - it looks for a valid pressure reading and a valid outside temperature reading.

The offset is actually set by adjusting the relative pressure until it is correct, rather than entering the offset directly. Cumulus calculates the offset by subtracting the abs pressure from the rel pressure in the 'fixed' block of the station memory at start up. It sounds like the absolute pressure in that memory is not correct. You can override this calculation and supply the offset to Cumulus directly, by editing cumulus.ini (while Cumulus is stopped), and in the [Station] section add an entry:

EWpressureoffset=x.y

where x.y is your pressure offset in hPa (i.e the abs pressure subtracted from the rel pressure). Cumulus will then use this instead of reading the fixed block offset from the station. It may be that the 14.5 which Cumulus was calculating previously is what you need.

Re: Windows 7 reboot issues

Posted: Mon 21 Sep 2015 12:10 pm
by starchild
Many thanks Steve - that makes complete sense. I'll check the relative value on the station when I get back from work.
As I was poking around I did note that the offset value in the ini file was present and set to zero.

So, it looks like there is effectively a corruption in the stored values in the station which upsets Cumulus - or at least there was a large change at some point. That's really where I need to focus my attention otherwise by entering an offset in Cumulus I will at best mask the issue and at worse cause me a problem when the values are overwritten in station memory (or the relative value 'corrects' itself).

Either way, in order to avoid a hole in my data records (from June to now) I'll then need to set about manipulating the data files to ensure consistent pressure readings.

Once again - thanks for your help Steve.

Re: Windows 7 reboot issues

Posted: Mon 21 Sep 2015 9:45 pm
by starchild
OK, so I re-read your message when I got home and have followed your instruction to enter an overriding offset in cumulus.ini (by the way, I have two of these, one in the root directory and another in the data directory - I added the line to the root directory version).

Checking the station as you instructed shows that there is indeed a difference of 14.5hPa so that's what I've set.

The data is now being uploaded from the station and I can see that the pressure is no longer zero as the upload is in progress - it is showing the historic values being read as expected.

I guess the best thing to do now, once the application is back up and running, is to stop Cumulus and pull the batteries/USB cable in order to clear that errant block of memory.

Once again, many thanks Steve - I've learned a bit more about the mechanics of Cumulus which may well prove useful in the future. Maybe I'll get another 7 years out of the system now!

Re: Windows 7 reboot issues

Posted: Tue 22 Sep 2015 7:14 am
by steve
If everything is working OK, you may want to consider leaving it alone rather than resetting it.

Re: Windows 7 reboot issues

Posted: Tue 22 Sep 2015 7:51 pm
by starchild
Checked again and everything is working perfectly.
You're probably right Steve - I'll leave well alone!