Am 01.07.2016 um 13:29 schrieb Guy COLIN: > > My conclusions are: > -1 as suggested by others I'm going to write an "error tolerant" script. In > case of error I'll use the previous value. Because my goal is just to have no > "holes" or wrong values in my graph. And I log every 5 minutes only. > You could also do an immediate retry. It's very likely a retry will produce a valid value.
I'm not going to show you my dispenser application. It's very cleanly programmed but it has so many nested retry loops in different levels of the object hierarchy it's still hard to understand when it retries and why. It recovers even from errors like full disconnect and later reconnect of devices. > -2 I still hesitate to totaly give up with "owfs /mnt/1wire fuse file system" > and use owread instead. I think I'll do it for the reading values > (temperatures, humidities) but I'll keep fuse file system for the writings > (PIO). Just because my current scripts use it. > Uh, "owwrite" is as easy to use as "echo". And you can get rid of FUSE which may break with kernel updates, too. > -3 I'll stick with 2.8p15. Except if there are obvious good reasons to update > 3.1 As said previously my system is full read only and I want it to be a > black box running 24/7 and never update anything. It is doing the job. > If it does it for you it's fine. The most obvious reason not to stick to 2.8p15 is when you want to use the w1 subsystem of a linux-3.16 or later kernel. Then you *have to* update to 3.1p1. Other reasons include various fixes to the other busmasters and devices. And of course, if you have errors again. Because, nobody wants to look for errors which may be already fixed for years. As said, if it does it for you, it's fine. > Side note about the "out of range" errors temperature = 85°C > I did some tests on an arduino connected to a DS18B20 and the official > arduino 1wire software available here: > http://playground.arduino.cc/Learning/OneWire > This program requests a temp conversion to the 1wire sensor, waits 750ms, > reads the result in the scratchpad. > I noticed that if I reduced the 750ms to a lower value at some point I read > 85°C. > Exactly on testing on the breadboard at 591ms I get a good value at 590ms I > get 85°C. Of course anything lower than 590ms will give 85°C. > So what about if sometimes a temp sensor on the 1wire bus takes a little more > time than usual? More than owserver is waiting? Is this possible? In this > case we'll read 85°C... No? Is it correct? > In that case you should get the previous conversion value. Or the power-up default of 85°C. Kind regards Jan ------------------------------------------------------------------------------ Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San Francisco, CA to explore cutting-edge tech and listen to tech luminaries present their vision of the future. This family event has something for everyone, including kids. Get more information and register today. http://sdm.link/attshape _______________________________________________ Owfs-developers mailing list Owfs-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/owfs-developers
