Stuck Pressure Reporting to CWOP

[HurricaneTracker]

I am using the latest build of Cumulus (1.9.1) and am trying it out before moving over from WD. I love the simplicity of Cumulus, and the all-time record keeping is a huge plus over other software. My main issue preventing me from using Cumulus full-time is that my pressure data are not routinely being updated to the CWOP server. See sample output below. The pressure Cumulus is sending is *NOT* the current pressure (i.e., it does not match the value off the datalogger). However, if I close Cumulus and start again, the next transmission (at 5-min intervals) is the current pressure, then it goes back to being stuck again until I restart Cumulus.

Is there a fix for this? I hope so - would love to continue using this great software and will strongly consider a donation once fixed!

Cheers and thanks.

Mike

KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112134z3529.27N/08245.50W_355/000g000t078r000P017h76b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112139z3529.27N/08245.50W_355/000g000t077r000P017h78b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112144z3529.27N/08245.50W_355/000g000t077r000P017h80b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@112149z3529.27N/08245.50W_355/000g000t076r000P017h81b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112154z3529.27N/08245.50W_355/000g000t075r000P017h82b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@112159z3529.27N/08245.50W_355/000g000t075r000P017h82b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112204z3529.27N/08245.50W_355/000g000t075r000P017h82b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112209z3529.27N/08245.50W_355/000g000t075r000P017h84b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@112214z3529.27N/08245.50W_355/000g000t075r000P017h83b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112224z3529.27N/08245.50W_355/000g000t075r000P017h83b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112229z3529.27N/08245.50W_355/000g000t075r000P017h84b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112234z3529.27N/08245.50W_355/000g000t074r000P017h83b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112239z3529.27N/08245.50W_355/000g000t074r000P017h83b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112244z3529.27N/08245.50W_355/000g000t074r000P017h85b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112249z3529.27N/08245.50W_355/000g000t073r000P017h85b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112254z3529.27N/08245.50W_355/000g000t073r000P017h86b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112259z3529.27N/08245.50W_355/000g000t073r000P017h86b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@112304z3529.27N/08245.50W_355/000g000t073r000P017h87b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112310z3529.27N/08245.50W_355/000g000t072r000P017h86b10154eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-2:@112314z3529.27N/08245.50W_355/000g000t072r000P017h87b10154eCumulusDsVP
[CUMULUS WAS RESTARTED HERE]
KC9AZX>APRS,TCPXX*,qAX,CWOP-3:@112336z3529.27N/08245.50W_355/000g000t071r000P017h88b10160eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@112346z3529.27N/08245.50W_355/000g000t071r000P017h89b10160eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@112356z3529.27N/08245.50W_355/000g000t071r000P017h89b10160eCumulusDsVP
KC9AZX>APRS,TCPXX*,qAX,CWOP-1:@120011z3529.27N/08245.50W_355/000g000t070r000P017h89b10160eCumulusDsVP

[steve]

CWOP requires Altimeter Pressure rather than Sea-Level Pressure, so the value sent to CWOP won't match the console - but it shouldn't get stuck like that; I think that's the first time I've seen that reported.

Could you please turn on the debug log, then wait 10 minutes or so, then zip up the diags folder and attach it here, along with the debug.log file.

[HurricaneTracker]

Thanks for the response, Steve. Hope it's a quick fix. Attached files as requested.

[steve]

Once a minute, Cumulus reads the barometer data from the VP2 that it needs to calculate Altimeter Pressure, using a VP2 command called "BARDATA", and if the required data is returned (the normal case), it logs the values to the debug log. Here's an example of mine:

1430.7141 : 12/05/2011 14:36:00 Minute timer 36
1430.7141 : Adding LH entry, ts = 14:36:00 raintot = 230.199996948242
1431.0219 : MSLP=1006.03173828125 SP=29.6906757354736 AP=1005.72320556641 pIn=29.7080001831055 eFt=16 mtF=52 C=1.89999997615814

Here's yours:

0712.799 : 5/12/2011 9:16:00 AM Minute timer 16
0712.799 : Adding LH entry, ts = 9:16:00 AM raintot = 16.5699996948242

The line logging the pressure data is missing. Unfortunately, it doesn't log an error message in this case, so we only have the absence of the message to know that the required data wasn't returned by the station.

How is your VP2 connected to your PC?

[HurricaneTracker]

It is through a USB connection (set to "serial") in the software. I am not using the standard Davis serial-to-usb connector, but am using "Prolific Technologies" PL-2303 serial-to-usb connection on COM PORT 6. It says the device is working properly and the driver is up-to-date. Note that when using Weather Display, all data come across and flow just fine (including the 1-minute pressure update), so I don't think it is the usb connection...

[steve]

Cumulus doesn't handle the serial comms itself; it uses a DLL supplied by Davis for all communication with the VP2. For the case of the special BARDATA command, it tells the DLL to send a 'wakeup' command, followed by the BARDATA string. It then reads all incoming characters until there are no more to be read. In your case it looks like it's not getting enough (or any) characters in response to the request.

I guess something about your USB adapter (timing, buffering, or something) is causing this not to work.

What firmware version does your VP2 have? Is there any way you can try with a real serial port?

[HurricaneTracker]

My VP2 is connected to my laptop, which I purchased early last year, so it does not have any serial port options.

My firmware on the VP2 is fairly old, May 2005.

[steve]

Looking through the firmware release history, there are a couple of changes since 2005 which might have a bearing on this. But your existing firmware requires a Davis Updater, so it's not an easy thing to upgrade.

Other than the firmware and the USB adapter, I'm afraid I can't think of anything else to suggest.

[HurricaneTracker]

OK, well thanks for helping and looking into it.

What is strange is that when I restart Cumulus, it does send the current pressure to CWOP. Moreover, using "View Graphs", I can see the whole trace of the pressure data and it looks great (i.e., not a flat liner).

Would turning off the option to force a VP 1-min pressure update make any difference?

[steve]

What is strange is that when I restart Cumulus, it does send the current pressure to CWOP.

Yes, it's presumably getting a valid response back the first time, but then no more. You could confirm this by stopping Cumulus and editing the cumulus.ini file so that the debug log is turned on at start up. Add a line "Logging=1" to the [Station] section.

Moreover, using "View Graphs", I can see the whole trace of the pressure data and it looks great (i.e., not a flat liner).

Yes, that's because the graphs, and everything else in Cumulus apart from the CWOP update use Sea Level pressure, which doesn't require the special 'bardata' data.

Would turning off the option to force a VP 1-min pressure update make any difference?

It's possible. The 1-minute update is performed in the same way as the 'bardata' read, so anything that goes wrong with one might affect the other.

[HurricaneTracker]

Added the "Logging=1" to the .ini file you mentioned, and restarted Cumulus. Here are a few sample lines from the debug.log file. Notice upon startup the correct bardata alogrithm goes to work...

0014.336 : MSLP=0 SP=27.4713859558105 AP=30.1001625061035 pIn=29.9939994812012 eFt=2522 mtF=69 C=5.40000009536743
0015.990 : 5/12/2011 11:55:00 AM Minute timer 55
0015.990 : Adding LH entry, ts = 11:55:00 AM raintot = 16.5699996948242

I've also turned off the "1-min update". After 15-minutes of running this new setup, the bardata still appears to be "stuck."

[steve]

Yes - so the first 'bardata' works OK, but subsequent ones don't. This suggests to me an issue between the DLL and the firmware, rather than the USB adapter.

[pdthull]

Am also having trouble with pressure readings being sent to CWOP. They are not changing. The pressure readings to wunderground.com are changing and good.
Am using 1.9.1 on Windows7 64 bit with a Vantage Pro 1 with serial port. I was using an earlier version of Cumulus on an XP 32 bit machine and it seems ok. The problems started with the 1.9.1 on the new machine, but don't know if its the Windows7 64 bit or the 1.9.1. Please tell me where I can download 1.8.9 to test on my new machine.

[steve]

Please tell me where I can download 1.8.9 to test on my new machine.

I don't keep old versions of Cumulus, but someone else may have a copy.

Alternatively, you could install 1.9.2 (which has a little more diagnostics for failed 'barread' commands) and turn on the debug logging in cumulus.ini as described earlier in the thread. The debug.log file would then at least show whether it's the same problem, if nothing else. What version of firmware does your console have?

[pdthull]

Am using Rev B from 2006.