Cumulus Support

apenwith wrote:Would it be possible to ignore messages for 75ms after a first message in the software.

Well, yes, as I said earlier it would involve a timer and a threshold for when to accept new 'tips'. It introduces complications to what was originally a nice clean simple implementation.

I'm now starting to get a bad feeling about this; this is why I don't intend to add support for any more weather stations. I always end up having to add hacks to work around shortcomings in the hardware/firmware

Hi Steve
You've done enough and we are grateful - and as the RG-11 is yet to be proved as really comparable to a good bucket the tip size workaround is fine. I would make the point that users should check the debug log to see if they need that adjustment.
Regards
Alan

To eliminate possible relay bounce, what would be a good size and type of capacitor?

Duke

duke wrote:To eliminate possible relay bounce, what would be a good size and type of capacitor?

As Alan says, it's not quite that simple: http://www.elexp.com/t_bounc.htm

I think that we need more data on the problem - i.e. results from more people - before we can start to understand exactly what's going on.

steve wrote:

duke wrote:To eliminate possible relay bounce, what would be a good size and type of capacitor?

As Alan says, it's not quite that simple: http://www.elexp.com/t_bounc.htm

I think that we need more data on the problem - i.e. results from more people - before we can start to understand exactly what's going on.

I think more data won't be a problem over the coming weeks .

You've done enough and we are grateful

Ditto.

Duke

I've emailed Hydreon to see if they have any suggestions.

EDIT: Steve beat me to it while posting.

Alan,

As you are the only person with this working at the moment, how about emailing Hydreon and see it they have anything to say about the relay closure time and may be pointing them here?

Duke

@Alan, looking at your results thread, you've obviously eliminated the "double tips". Did you simply half the bucket size in Cumulus?

Duke

I've been exchanging emails with Hydreon. They can't see that there's very much to go wrong with the setup we're using, and suggest that perhaps for some reason the leading and trailing edges of the relay closure are both being treated as 'close' indications. They suggest a test involving using a switch instead of the relay, to see if closing the switch has the same effect.

It would be good to get some results from a real serial port, to see if the USB adapter is causing the problem.

Hi Steve

I tried shorting out the relay wires and obviously I could not replicate a 50ms closure. What was very odd was that he longer I held my 'relay' closed the more ON's and OFF's I got. So somehow a change message is happening even when there is no change.
Nothing venture nothing gain I put a small electrolitic capacitor (I think about 10uF) across the relay leads and against all reasoning it did seem to improve things. I could then reproduce a proper ON/OFF pair. I'll leave it in place for some real rain and see what happens.
Actually I might get the hose out as rain seems unlikely for a while.
I'll let you know shortly.
Regards
Alan

How long is the cable connecting the RG-11 to the serial port?

Do you get the off/on readings if you short out at the serial port end?

Hi Steve
Now with a wet arm (leaky hose) here is the debug log with the caoacitor installed.

In answer to mrossley - yes I shorted at the usb/serial port end -the RG11 is on roughly a 4M cable.
Regards
Alan

Thank you Steve for the integration of the Hydreon RG-11 into Cumulus.

I still have to hook mine up to Cumulus. I currently have a Hydreon RG-11 configured with a Hobby Boards dual counter. I am seeing differences to the Davis Vantage Vue tipping bucket with a little bit of rain yesterday.

The second RG-11 will be configured adjacent to the first one in the next couple of days. Kind of having some drought conditions here lately though.

I rarely get any debounce effects with the other Hobby Boards dual counter I have in place which is covering two water meters.

apenwith wrote:Now with a wet arm (leaky hose) here is the debug log with the caoacitor installed.

That looks good. I guess we now know what the capacitor is for, even if we don't know why.

Hi Steve
I think Michael's capacitor was relay output to ground - mine is across the two relay wires. In theory it should be holding the high for longer and making things worse if the messages continue while the high is on. Actually I didn't check the polarity - I think I had better do that before anyone else tries it. Still - if it works - more time will tell. I've set the tipsize back to 0.2mm.
Regards
Alan