Could you please take a look at the question I've recently posted on SO regarding problems with DHT11: http://stackoverflow.com/questions/43453868/dht11-sensor-failure-on-edison-whereas-working-successfu... python - DHT11 sensor failure on Edison whereas working successful on RPi - Stack Overflow
I read some threads here about one-wire devices workaround but I don't fully understand it and maybe do something wrong. Furthermore, nobody ever experienced those IN/OUT switching port problems on RPi. It is also a genuine computer system with real time OS, isn't it?
Thank you in advance,
The 1-Wire protocol, as you know, requires deterministic timing. It is not possible to do deterministic timing on a computing platform like Edison, or even a PC. You might get somewhere close to it using usleep() but your app may get interrupted at any time, thereby messing up your timings.
What you need to do when using a system without deterministic timing of GPIO is attach the sensor using an interface. That interface should use a small micro which DOES have deterministic timing. (ie a Pic, ATTINY, etc).
First, I want to say thanks to SpiderKenny, we really appreciate your help in the community. And Bart, I recommend you to try SpiderKenny's suggestions, in case you need more information take a look at this thread created by SpiderKenny: https://communities.intel.com/thread/55975 https://communities.intel.com/thread/55975
I hope you consider this useful.