Cumulus Support

I have noticed that the low temperature on the dashboard and on the web has been 32deg f at noon (1200) on April 9, 10, and today the 12th but the charts show today the temperature at noon was 66.7 deg f and the overall low temp today was 45.9 deg f at 04:05. Yesterday seems to be correct. I do not know if it has something to do with shutting down and restarting a bunch of times or what. I will keep an eye on it now that I seem to have running the way I like

Thanks Don

If you zip up the MXdiags folder and attach it to a reply, I'll take a look.

Diags files you requested

Thanks for the help

Don

It happens during the 'synchronisation'. It's the same issue as in this thread - https://cumulus.hosiene.co.uk/viewtopic.php?f=27&t=13556

You have a Pi2, which is capable of supplying more power, but I think you have to set a kernel option to enable that? I had to do that for my USB hard drive. Do you have the station plugged directly into the Pi?

Yes it is directly to the pi. FYI it jus happened again

Don

It did it this time a 11:49 and then right after it started with the USB error problem. After reading the other postings I shut down everything and plugged the station into a powered USB hub. Now lets see what happens.

Thanks - that will be a useful test.

Steve

That did not help. At 11:59 it set low temp to 32deg f. This time it missed the data log and did not show a spike on the charts.

I found out that to get full power to the USB ports you need a 2A power supply. I was running a 1A but the hub was running 2A. put a 2A power supply on the PI and moved the station back directly to the PI and moved everything to the hub.

See what happens tomorrow

Don

I don't understand why only two of you are seeing this (although it's not clear how many people have the sync option set). I powered up my original Pi again and connected my FO station to it (via a powered hub) and it performed the sync straight away without any problems. It does look like some kind of hardware/power issue.

I connected my station directly to my Pi last night and left it running. When it came to do the sync this morning, I got the same problem. I'll put it back on the hub and leave it running.

I've my old WH1080 attached direct (without a usb hub) to my Pi 2B and have not seen this (yet) the pi is powererd with 2 A 5V

Steve

I went 24 hours with out the errors and the low temp set to 32 deg f after removing the check on the "Synchronize Fine Offset Reads". Then this evening I used Cumulus.ini file that I had saved which still had the check and I got errors and low temp set to 32deg f right away.

In version 1 when I got "The" error everything locks up with version 3 it just seems to set the low temp and everything else continues to work.

Thanks

Don

I get the problem when synchronising starts if I have the station connected directly to the Pi. I haven't so far seen the problem when using a powered hub (I have no other devices connected). So it seems clear that the problem is with the Pi's USB port - perhaps a power issue. It copes with reading the data every 10 seconds, but not every 0.5 seconds when MX does its synchronising.

I'll continue to investigate, though; there may be something I'm missing.

Hi,

I'm absolutely not an expert, but will try to help with some suggestions on, how to deal with USB related problems.
Assuming the weather station works on a regular laptop/desktop PC, the problem is isolated to interaction with the RPi.
First of all, the current draw from the RPi to the weather station should be below 50 mA, as this should only be used to drive an internal USB to RS232 converter inside the weather station. Unless the the processor in the weather station has a USB interface, the current draw should be close to 0 mA.

With that said, it can still be a power problem.
The RPi is usually powered from a relative small and cheap wall mount power supply (mine is!). Sure it can deliver 5V 2A, but I'll bet the noise levels on the 5V power are high and there is not much on the RPi to deal with this.
This noise will have a negative impact on the USB communication lines (D+/D-), which will lead to bad or lost communication.
This is of course very dependent of the kind of power supply used, USB cable (length/quality) and finally other things drawing current or generating noise of the RPi (keyboard, mouse, ethernet, wifi and so).

Some of you have observed that a powered USB hub can minimize or even get rid of the problem. That might be because the USB hub is "acting" as a "signal sharper" between the RPi and the weather station, so the USB signals are fine to maintain the necessary communication.
What also might help, is using a USB cable with a ferrite core and also a power supply with a ferrite core mounted close to the micro USB connector for the RPi.

madsandrup wrote:First of all, the current draw from the RPi to the weather station should be below 50 mA, as this should only be used to drive an internal USB to RS232 converter inside the weather station. Unless the the processor in the weather station has a USB interface, the current draw should be close to 0 mA.

The station is actually completely powered via the USB interface, so how much current is that likely to require?