Jump to content


Photo
- - - - -

An odd polling failure problem with the anemometer board


  • Please log in to reply
4 replies to this topic

#1 ksdehoff

ksdehoff

    Cocooner

  • Registered
  • PipPip
  • 36 posts

Posted 14 August 2011 - 09:09 AM

Eric:

I've had the anemometer board in place over a year. It is in a weathertight box on my roof and I connect to it via a 50 ft rj45 cable. I also have the insolation/temp/humidity sensor chained from this via a short 6 ft rj45 cable and finally I've also connected the rain sensor to the screw terminals on the anemometer board via a short 3 ft cable. This all runs back to the active hub.

The problem I'm having is that the wind direction routinely stops updating after a lengthy period (typically a week to week and a half). Wind Speed, rain sensor and insolation/temp/humidity continue to report back just fine. I've tried polling rates from 28 sec up to 60 sec and the time to fail is very consistent. I'm wondering if you can suggest anything that I could consider in figuring out how to troubleshoot this.

I'm using Michael Sharry's xapmcs1wire package that I suspect you aren't all that familiar with but with over 100 devices I'm polling in my house, the entire system is impressively stable other than this one input.

Thanks for any suggestions!
Ken

#2 ericvic

ericvic

    Dedicated Cocooner

  • Registered
  • PipPipPip
  • 754 posts
  • Location:Birmingham, AL

Posted 15 August 2011 - 11:17 AM

Ken,

It seems really strange that only one of the 1-Wire chips would go missing. Can you try using the OneWireViewer and let it run to see if you start getting errors so we can try to narrow it down to either hardware or software.

Eric

#3 ksdehoff

ksdehoff

    Cocooner

  • Registered
  • PipPip
  • 36 posts

Posted 23 August 2011 - 11:36 AM

Ken,

It seems really strange that only one of the 1-Wire chips would go missing. Can you try using the OneWireViewer and let it run to see if you start getting errors so we can try to narrow it down to either hardware or software.

Eric


Eric:

After the failure occurred I switched to the 1wire viewer. I was able to successfully see the direction sensor but I noted some interesting behavior I want to ask about. Sorry this is a 1wire basics question. When I was polling in 1wire viewer at 30 seconds with the one channel active I would see the devices disappear and reapper. When I increased the polling to 10 seconds, the devices remained - didn't disappear. My understanding of appropriate polling rates is that slower is better, but is there instability from slower polling and does it make sense to you that the 10 sec polling rate would appear to be more reliable?

Also regarding the comment about the rc filter - would you recommend this to all us hub owners with odd problems? If so could you post a little more detail so I'm not dangerous with my soldering iron?

Thanks
Ken

Edited by ksdehoff, 23 August 2011 - 11:37 AM.


#4 ericvic

ericvic

    Dedicated Cocooner

  • Registered
  • PipPipPip
  • 754 posts
  • Location:Birmingham, AL

Posted 24 August 2011 - 10:54 AM

Eric:

After the failure occurred I switched to the 1wire viewer. I was able to successfully see the direction sensor but I noted some interesting behavior I want to ask about. Sorry this is a 1wire basics question. When I was polling in 1wire viewer at 30 seconds with the one channel active I would see the devices disappear and reapper. When I increased the polling to 10 seconds, the devices remained - didn't disappear. My understanding of appropriate polling rates is that slower is better, but is there instability from slower polling and does it make sense to you that the 10 sec polling rate would appear to be more reliable?

Also regarding the comment about the rc filter - would you recommend this to all us hub owners with odd problems? If so could you post a little more detail so I'm not dangerous with my soldering iron?

Thanks
Ken


There is no reason I can think of that polling quicker or slower would make the devices disappear and reappear. Devices usually disappear when there is a short on the 1-Wire network or when they lose external power.

As for the rc filter issue. If your 1-Wire network is not very long then you can just remove the capacitor C11. If you have a longer (300' or longer would be my guess) then you would need to cut a trace and move C11 to the other side of R5 with a piece of wire.

Eric

#5 ksdehoff

ksdehoff

    Cocooner

  • Registered
  • PipPip
  • 36 posts

Posted 09 October 2011 - 11:49 AM

So it appears that removing that capacitor has really helped. I'm at 2 weeks now. Given the MTBR was long before I can't be certain, but this is the first time I've ever gone 2 weeks - so I'm cautiously optimistic!

Ken




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users