1-Wire device support in Cumulus

[Charlie]

I was just cruising through old posts and I came across this one.
Anyone still interested in getting 1-wire data into Cumulus, please drop me a line. As mentioned elsewhere, I've developed a microcontroller based hardware solution to bridge 1-wire data into Cumulus, and I'm confident enough with the solution at this point to build a few more units for those folks interested.

[spou]

I'm surprised you need a special hardware bridge. The regular 1wire-to-rs232 ( DS9097U ) or 1wire-to-USB ( DS9490 ) should be enough for most needs.

I personally run a good 300 feet, 12 DS1820 drops network over CAT3 cable going into a DS9097U without problems. For me, the real problem is getting the 1wire data into cumulus. This looks like a software problem more than anything else, specially when counters are in the equation (for rain and wind speed)

Or maybe I'm wrong and where is indeed a need for a hardware interface, there are things I dont know

Can you explain more, there might be something I can learn here.
Spou

[Charlie]

I'm surprised you need a special hardware bridge. The regular 1wire-to-rs232 ( DS9097U ) or 1wire-to-USB ( DS9490 ) should be enough for most needs.

I personally run a good 300 feet, 12 DS1820 drops network over CAT3 cable going into a DS9097U without problems. For me, the real problem is getting the 1wire data into cumulus. This looks like a software problem more than anything else, specially when counters are in the equation (for rain and wind speed)

Or maybe I'm wrong and where is indeed a need for a hardware interface, there are things I dont know

Can you explain more, there might be something I can learn here.
Spou

It has nothing to do with range. 1-wire is purely a transport protocol, and so it dumps data in a variety of formats with no standard or even self consistent meanings. The problem is not one of simple physical interface - that's fairly trivial, if you ignore error detection (sometimes CRC-8, sometimes CRC-16) and addressing issues (every device is unique, so manual intervention is usually required). The problem is to interpret and process the data, and then present it to Cumulus is a fashion similar to the way commercial stations do. The whole 1-wire job is huge, and I suspect is the main reason Steve is not keen to support it directly. There are as many flavors of hardware as there are of weather.

My approach is to offload all of this into a custom processing unit, supporting a small family of common sensors, and emulate a standard station to Cumulus.

A bit more than a "bridge".

Having said that, people like 1-wire because it's simple and robust and sensors can be quite accurate with extremely rare erroneous data. Most of us that like to fiddle with hardware, also like the power of Cumulus for data manipulation and presentation, and there's the usual rub.

[spou]

Most of us that like to fiddle with hardware, also like the power of Cumulus for data manipulation and presentation, and there's the usual rub.

Interesting idea you have, and it proves that for the one that has a hammer, everything is a nail.

I mean: I'm a software guy, and I naturally approach problems with that viewing angle, while you seem like a hardware guy, and look at problems from your own end.

Back to the subject at hand, I think I understand that you more or less would like to replace one of the few supported station by your bridging circuitry. In that case, you indeed need some electronics to do the interpretative work between 1wire and cumulus.

In my personal case, it will have to wait until my 968 goes belly under. The later the better

Spou