Trixology
WeatherCat => WeatherCat General Discussion => Topic started by: dmatrix on October 26, 2012, 05:56:51 PM
-
Greetings!
I'm looking for guidance in troubleshooting this rather annoying connectivity issue I'm experiencing.
I have a Davis Vantage Vue station, with the USB DataLogger on my Mac Mini running OSX 10.8.2. I'm running version 3.0 of the USB driver. The console is directly connected to an onboard USB port to the back of my Mini (ie, no external usb hub).
Almost on a daily basis, I will receive alerts that WeatherCat lost communications with the console and is unable to reset it. This results in me either doing a reboot, or physically disconnecting the console from the computer.
Now, I have been running this same hardware setup and USB drivers with other Weather Station software without these issues - for weeks at a time. It seems that as soon as I switched over to WeatherCat the stability and reliability has greatly suffered.
The last incident occurred this morning -and I've attached the WeatherCat Logs showing the event. At the time that this occurred, it showed that I had 44 Comms Errors, and 0 Sensor Errors. Now with my previous weather station software, I was also receiving the occasional comms errors so, the fact that it's showing up here doesn't surprise me. I would like to figure out what I need to do to resolve those errors, but I guess what I'm saying is I don't know if that necessarily is the cause of this issue with WeatherCat.
What are the suggested next steps that I should take to further nail down what's happening here?
Many thanks in advance!
-bill
-
Dear Bill and WeatherCat fans, [cat]
I'm looking for guidance in troubleshooting this rather annoying connectivity issue I'm experiencing.
. . . . .
The last incident occurred this morning -and I've attached the WeatherCat Logs showing the event. At the time that this occurred, it showed that I had 44 Comms Errors, and 0 Sensor Errors. Now with my previous weather station software, I was also receiving the occasional comms errors so, the fact that it's showing up here doesn't surprise me. I would like to figure out what I need to do to resolve those errors, but I guess what I'm saying is I don't know if that necessarily is the cause of this issue with WeatherCat.
You seem to have the fundamentals down, but 44 Comm errors is very high. Most of us using Davis stations have 0 errors for months at a time.
Try going through the troubleshooting tips on the WeatherCat Wiki:
http://wiki.trixology.com/index.php/Davis (http://wiki.trixology.com/index.php/Davis)
One possibility is that you have some EMI interference that is interfering with your USB cable. There is advice in the Wiki on how to deal with that.
Cheers, Edouard [cheers1]
-
Hi Bill,
As Edouard points out, the problem is likely to be electrical interference/noise (a thermostat switching/fluorescent lights/ TV's etc) - this crashes the Silicon Labs driver. WeatherCat will try to recover it (it attempts to unload the USB driver and reload it) but if this fails you have to reboot WeatherCat and/or the machine (the USB driver lives in kernel space so can make it unstable when it crashes).
Davis have an F.A.Q. on this on their site ( http://www.davisnet.com/support/weather/faq/answer/index.asp?ProdFam=2&faqCat=3&faqid=6 (http://www.davisnet.com/support/weather/faq/answer/index.asp?ProdFam=2&faqCat=3&faqid=6) ) and there's also information available on the wiki, specifically: http://wiki.trixology.com/index.php/Electronic_Interference_Effects_on_Station_Hardware/Data_Logger/USB_Cable (http://wiki.trixology.com/index.php/Electronic_Interference_Effects_on_Station_Hardware/Data_Logger/USB_Cable)
Basically the advice is remove the cause of the interference and/or use ferrite clamps on the cables to filter out the noise.
Cheers,
Stu.
-
Hi Bill
I had a Vantage Vue with the USB datalogger until recently. I always had lots of comms errors as well, if run for a few days I could have a couple of hundred errors recorded.
I had to return the Vue back to Amazon due to a firmware problem in the ISS that the UK agents for Davis could not resolve and bought a Pro 2 as a replacement.
I found that the USB datalogger fits very tightly into the Pro2 console, a lot tighter than it did in the Vue console, in fact with the Vue if I removed the datalogger cover for the Vue the datalogger would fall out ! It may be that your datalogger is slightly loose in the console leading to the communications failures.
I very rarely get communications errors now, using the Pro 2 console in the same location with the same USB datalogger.
JC
-
Based on everyones suggestions - I tried the following:
* Used a shorter USB Cable / one with a built in ferrite.
* Physically moved the console near the computer - previously it was on the other side of the wall in the family room.
* Run the console on batteries only. (PS already was the updated one with the built in ferrite)
Upon doing all of this, I still managed to get about 12 comm errors over night.
I just reseated the data logger - so we'll see if this has any effect.
Thanks again for the suggestions!
-bill
-
Just an update - the issue has yet to be resolved. :(
I replaced the USB cable with a new one that included two ferrites, as well as moved the console to a new room as well as connected it to a different computer.
I'm at this point at a loss as to what to try next.
Thanks in advance!
-
Hi Bill,
Might be worth uninstalling the Silicon Labs driver then reinstalling again, just to make sure you have the latest installed (V3.0).
https://www.silabs.com/products/mcu/Pages/USBtoUARTBridgeVCPDrivers.aspx (https://www.silabs.com/products/mcu/Pages/USBtoUARTBridgeVCPDrivers.aspx)
Another thing to try, and it's a bit of a long shot, is plugging the logger into a powered USB hub and connect the computer to that rather than directly to the logger. This goes against all the advice but the thinking is filtering (and possible isolation as it's powered) in the hub might help.
Cheers,
Stu.
-
Hi Stu -
I also tried connecting it to a powered hub - with no change in behavior. The second computer that I tested with, never had the console connected to it... so I had to perform the initial installation of the drivers before launching WeatherCat.
Is there any chance that I'm running into a compatibility issue with the drivers, and Mountain Lion? I also have a PPC Mac Mini that I can also try and use for testing purposes and see if I get any favorable results.
I fired off an email over to Davis late last night to see if they have any other suggestions, will let you know what they say.
Is there any simple test or utility that I can run on a Windows platform to eliminate the chance it's a faulty Datalogger, or console?
Thanks again!
-bill
-
Hi Bill,
I was just digging around the forums looking for other Vue comms issues and came across the following:
http://athena.trixology.com/index.php?topic=123.msg1207#msg1207
Is this similar to what you're seeing? (I'm guessing not as your's is locking up and not recovering?)
One other thing to check - the batteries in the console - if these get low they can drag the station voltage down which will lead to poor comms. If you have a fresh set, put them in to see if it makes any difference.
Cheers,
Stu.
-
Is there any chance that I'm running into a compatibility issue with the drivers, and Mountain Lion?
There are a few of us using the latest Silabs driver, WeatherCat 1.1 build 1570, and Mountain Lion 10.8.2. So I can't see compatibility being the problem.
Steve
-
Stu -
http://athena.trixology.com/index.php?topic=123.msg1207#msg1207
Is this similar to what you're seeing? (I'm guessing not as your's is locking up and not recovering?)
Yes. That sounds pretty spot on. I reset the application earlier this afternoon (as well as replaced the batteries). It's been running for slightly over 4 hours, and has logged 9819 Good Coms, and 12 Errors - 99.89%. In the first hour or so, I had already logged 2 errors.
Where the app is unable to automatically reset the Comms/Driver, the Application will attempt to restart - but fails at discovering the appropriate console (and driver). At this point, it will just sit there, dumbfounded.
I'd just exit out of the application and disconnect/reconnect the console, and restart the App and it will continue on. Or I will just remotely reboot the computer.
I would say that the app has yet to run a full 24 hours without needing to be reset.
i'm running WeatherCat 1.0.2 Build 3.
Thanks...
-bill
-
Thanks Bill. Yes, that's a USB driver crash. Comms errors come in two types, either a failure to reply to the poll or a reply with a bad checksum/crc. When faced with this, WeatherCat will continue to try for a few minutes (just incase it' some spurious issue), then attempt to unload the USB driver. When it does this it'll send out an admin email saying comms have failed. If the driver reloads (which it typically will do if it's just hung) you'll get an email saying comms have restored. If the driver doesn't unload, WeatherCat will reboot (if it's allowed) in an attempt to get the OS to unload the driver, but if the driver has crashed, it's likely this won't work and there's nothing more WeatherCat can do.
Now what's surprising is you're still having this problem after doing everything that's recommended, which leaves me wondering if the noise could be coming through the mains on the computer side (or if there's some other issue)? Do you have a mains suppressor/surge protector you could plug everything into?
Stu.
-
Stu -
Interestingly enough, Weathercat is able to properly recover by automatically resetting the driver when running on Mac 2. Whereas on my Mac 1, it would eventually just crash and die.
So far, on Mac 2, WeatherCat has been running for slightly over 2 days - 122512 Good Data, 90 Coms Errors - 99.93% Much longer than it would have run for on my Mac 2.
In regards to your power suggestion - Mac 1 was protected by a line conditioning UPS. When the console was connected to AC, it was also connected to the same UPS. Also ran the console on battery - which also generated the same errors.
Mac 2 - is connected to standard surge protector. When the console was connected to AC, it was also connected to the same surge protector.
-bill
-
I spoke too soon. I awoke this morning to find that Mac2 was unable to recover/restart the driver.
-
Hi Bill,
I've just completed a 3 day run here with ours - no errors. Now that's a UK unit, so just in-case there's some difference between the UK and US variants, I have one coming in from ATP. Should be here in a few days (at the mercy of the USPS and UK customs), I'll give that a test when it comes in.
Stu.
-
Dear Bill, Stu, and WeatherCat fans,
Sorry, really overloaded so I haven't been able to contribute to this thread. However, I've watched this exchange and I cannot help but wonder if this another case of a defective Vantage Vue like J.C. ran into. Davis makes good products, but the Vue is a new product and obviously there were trying to keep the costs down.
Have you given the Davis tech-support guys a call?
Cheers, Edouard [cheers1]
-
Sorry I've been absent from the thread for the last week or so.
I did email Davis regarding my issue (still occurring), and they want me to give them a call and work with them thru some troubleshooting steps. Unfortunately, it's just a matter of trying to coordinate a time to give them a call when I'm home and the kids are behaving, and they are still in the office. Nothing tricky about that! ;)
-bill
-
Dear Bill and WeatherCat fans,
I did email Davis regarding my issue (still occurring), and they want me to give them a call and work with them thru some troubleshooting steps. Unfortunately, it's just a matter of trying to coordinate a time to give them a call when I'm home and the kids are behaving, and they are still in the office. Nothing tricky about that! ;)
S-s-s-s-s-simple . . . . . . isn't it! :o
Best of luck in finally getting to the bottom of your troubles.
Cheers, Edouard [cheers1]
-
Just an update on this one - that US console still has not arrived! However, I do now expect it to be delivered tomorrow; I'll put it on test over the weekend - if I spot anything I'll let you know.
Check out the incompetence on display here (this is USPS priority international apparently...):
21-11-2012 16:43 Glasgow Central Depot Delivery scheduled
20-11-2012 07:14 Glasgow Central Depot Awaiting payment of charges
20-11-2012 00:42 Glasgow Central Depot Received at delivery depot
19-11-2012 13:20 National Hub Received and processed
19-11-2012 00:27 International Hub Revised charges to be paid
19-11-2012 00:26 International Hub Forwarded for UK processing
15-11-2012 15:37 International Hub Awaiting Customs Charging
15-11-2012 14:02 International Hub Awaiting Customs clearance
15-11-2012 14:01 International Hub Awaiting Customs clearance
15-11-2012 14:01 International Hub Awaiting Customs clearance
15-11-2012 14:00 International Hub Awaiting Customs clearance
15-11-2012 08:34 International Hub Awaiting Customs clearance
15-11-2012 08:33 International Hub Awaiting Customs clearance
15-11-2012 08:28 International Hub Received and processed
13-11-2012 09:16 International Hub Awaiting Customs clearance
13-11-2012 09:15 International Hub Received in destination country
10-11-2012 06:54 Delivery Agent - MIAMI Forwarded for export
10-11-2012 06:41 Delivery Agent - MIAMI Forwarded for export
06-11-2012 14:37 Delivery Agent - UNITED STATES AMERICA Collected from customer
Sigh....
-
Howdy Stu and WeatherCat lamenters of the human condition . . . . . .
Check out the incompetence on display here (this is USPS priority international apparently...):
. . . . . .
Sigh....
Isn't technology wonderful . . . . . . . . ::)
Before the Internet, you could hope that the idiots you ran into where the only ones on the planet. Now, thanks to the Internet, not only are you sure they are everywhere, but their numbers are . . . . increasing!!! (http://www.canebas.org/Weather/LWC_forum/Custom_emoticons/D%27oh.gif)
Oh well, grin and bear it!! :-\
Cheers, Edouard [cheers1]
-
Ha !
I bought my USB data logger from a supplier in New Jersey. It was being sent USPS International priority service. Took 8 days to travel 70 miles from the supplier to Newark, then 48 hours to travel 2 or so miles across town to the departure airport. Then UK customs held it for 5 days before releasing it to Parcelforce to charge me a small fortune to hold it for another 48 hours before I tracked it down and collected it.
JC
-
It's not good is it - how something can take 5 days to get through customs is beyond my comprehension????
Anyway, the new console came in yesterday, put it on test last night and got 30 comms errors in 12 hours. So I have two consoles now - one shows no errors in 3 days and one shows 30 in 12 hours. Something going on here.
Working on it...
-
Whew.. Thanks for the update! I'm glad to hear that it's atleast something that you have been able to reproduce, and it's not just an isolated issue with my setup. I really appreciate you going above and beyond in troubleshooting this for me!
Please let me know if there's anything you'd like me to do from this end.
-bill
-
HI Stu
After reading your post saying that you were now testing a Vue console with similar error rates, I once again connected up my Vue to run in parallel with my VP2 system. And its . . . . deja vu(e) . .
VP2 system: 7 days 11 hours , 276300 data samples with 4 comm errors (known interference and prior to staring Vue test)
vs
Vue system: 16 hours , 29050 data samples with 36 comm errors ! (also one failure/reset after 14 hrs)
I'm looking forward to any explanation you may find. Also happy to do any testing or provide additional data.
Regards,
David
-
Thanks folks. An updated station driver is available in the latest build: http://athena.trixology.com/index.php?topic=514.0 (http://athena.trixology.com/index.php?topic=514.0)
If you could give this a test and let me know how you get on, that'd be great.
-
Stu,
Major improvements!
I let the app run over night. After nearly 11 hours, I had 16,800 good data points, and only (1) Comms Error.. 99.99% for data quality.
This was after moving the console back to my original computer, which exhibited the worst problems. The console is running on batteries, with the ferrite on the usb cable.
I'll continue to run it, and let you know how things look in a 24 hour period.
-bill
-
Hi Stu,
Looks like about a ten-fold improvement in the VUE data with Build 1721. After about 19 hours and nearly 30000 samples I have 4 CRC comms errors, no time out errors and no comm fails.
I'll keep monitoring it and advise.
David
-
23 hours, 44 mins -
37013 Good Data
3 Comms Errors
99.99% Quality
No driver crashes or resets.
-
Thanks for the reports folks - good to know.
-
Stu,
Still running great! At this time, I have 14 coms errors, and 168700 good coms.
Based on your findings, and the updated driver - is there hope that you'll be able to tweak it more to get us at 100%?
-
Hi Bill,
Well, never say 'never' :) but there are a lot of variables, not all of them under my control.
Thanks for the updated report [tup]
-
I've suffered this problem for nearly two years now so it's awesome to finally have it sorted. I have an Australian unit. Since the update I now get 1 comms error every 10000 good data.
-
Good to know, thanks for the report breno.
-
I've just updated my Vantage Vue console firmware to version 3 which apparently beefs up the internal communications from console to logger. Lets see if you can get down to Zero now.
-
When I had this same problem ferrite clamps completely cure it - permanently. Ten of them end to end from computer to console.
-
No change to the 1 in 10000 good data rate with the latest firmware but that is fine with me. Ever since Stu weaved his magic the comms have been rock solid in comparison. I used to get 2-3 coms to station fail emails a day and a whole swag of -73.3 dewpoint readings on wunderground but not any more.
Just out of interest Stu, what was different with the US version of the Vue and the UK/Australia/Metric version that was returning so many coms errors?
-
Just out of interest Stu, what was different with the US version of the Vue and the UK/Australia/Metric version that was returning so many coms errors?
Second this. Also, I'm using the serial weatherlink hooked to a keyspan 19HS, and get about one com error per 10k as well. Was actually curious why I was getting them and found this thread.
I don't have any ferrites on the serial cable, so I might try that, but I figured a 0.01% error is not exactly a problem. Would be cool to eliminate the errors, but hey, it's still much better than my previous LaCrosse.