I have owfs and owserver running on a BBB under Ubuntu 12.04. It seems to be working, most of the time...
One of the DS18B20 sensors is linked via Node-RED to store a new temperature reading to the emoncms site at five minute intervals. Sometimes it draws nice smooth 12-bit resolution graphs for days. But occasionally the temperature will get stuck for up to an hour, and then make a step change. I'm pretty well convinced that emoncms is storing what they are sent, and Node-RED is sending what it reads from the /sys/devices/w1_bus_master1/28-000000884d88/w1_slave file. --> I'm not sure how to tell if the file is being updated. The date/time metadata does not change when readings change. Is there some way to verify that a fresh reading has been written? Maybe the file is intentionally only written if the value changes? Apparently each of the readings is done separately? temperature 24.6875 temperature10 85 <-- temperature11 24.75 temperature12 24.6875 temperature9 24.5 Watching the owserver page, I sometimes see readings that don't exactly make sense. This sequence was captured by manually copying and pasting and then refreshing the web page, so 15 seconds or more between groups: --- temperature 30.8125 temperature10 30.75 temperature11 30.875 temperature12 30.8125 temperature9 30.5 temperature 30.8125 temperature10 30.75 temperature11 30.75 temperature12 30.8125 temperature9 30.5 temperature 30.8125 <-- looks like this is separate from temp12? temperature10 30.75 temperature11 30.875 temperature12 30.875 temperature9 30.5 temperature 30.8125 temperature10 30.75 temperature11 30.875 temperature12 30.8125 temperature9 30.5 temperature 30.8125 temperature10 30.75 temperature11 30.875 temperature12 30.875 temperature9 30.5 ubuntu@ubuntu-armhf:~$ cat /sys/devices/w1_bus_master1/28-000000884d88/w1_slave e8 01 4b 46 1f ff 08 10 07 : crc=07 YES e8 01 4b 46 1f ff 08 10 07 t=30500 ubuntu@ubuntu-armhf:~$ cat /sys/devices/w1_bus_master1/28-000000884d88/w1_slave ed 01 4b 46 7f ff 03 10 e2 : crc=e2 YES ed 01 4b 46 7f ff 03 10 e2 t=30812 ubuntu@ubuntu-armhf:~$ cat /sys/devices/w1_bus_master1/28-000000884d88/w1_slave e8 01 4b 46 1f ff 08 10 07 : crc=07 YES e8 01 4b 46 1f ff 08 10 07 t=30500 ubuntu@ubuntu-armhf:~$ temperature 30.8125 temperature10 30.75 temperature11 30.875 temperature12 30.8125 temperature9 30.5 --- --> I know the value in the w1_slave file is usually the 12-bit value, but could it sometimes be one of the lower precision values? --> Sometimes it seems like having the owserver web page open makes it much more likely that occasional low-precision values will appear in the file - is that possible? Here is a sequence from today, when the temperature should have been slowly but evenly rising: --- 11 Jun 2014 19:35:24.136[owfs payload] 25.375 11 Jun 2014 19:40:24.210[owfs payload] 25.500 11 Jun 2014 19:45:24.383[owfs payload] 25.500 11 Jun 2014 19:50:24.371[owfs payload] 25.500 11 Jun 2014 19:55:24.413[owfs payload] 26.000 11 Jun 2014 20:00:24.502[owfs payload] 26.062 11 Jun 2014 20:05:26.447[owfs payload] 26.187 11 Jun 2014 20:10:26.461[owfs payload] 26.187 11 Jun 2014 20:15:26.496[owfs payload] 26.187 11 Jun 2014 20:20:26.531[owfs payload] 26.687 11 Jun 2014 20:25:26.554[owfs payload] 26.500 11 Jun 2014 20:30:26.946[owfs payload] 26.500 11 Jun 2014 20:35:26.615[owfs payload] 26.500 11 Jun 2014 20:40:26.648[owfs payload] 27.125 11 Jun 2014 20:45:26.673[owfs payload] 27.125 temperature 27.3125 temperature10 27.25 temperature11 27.375 temperature12 27.3125 temperature9 27 11 Jun 2014 20:50:26.696[owfs payload] 27.000 cat /sys/devices/w1_bus_master1/28-000000884d88/w1_slave b6 01 4b 46 7f ff 0a 10 a2 : crc=a2 YES b6 01 4b 46 7f ff 0a 10 a2 t=27375 ubuntu@ubuntu-armhf:~$ temperature 27.4375 temperature10 27.5 temperature11 27.375 temperature12 27.4375 temperature9 27.5 temperature 27.4375 temperature10 27.5 temperature11 27.5 temperature12 27.5 temperature9 27.5 --- --> I realize that kind of cut and paste logging is hopelessly imprecise when the owfs values are apparently being updated every 15 seconds. Is there some better way to see what is going on? I don't need 15 seconds, the file only gets read once per five minutes. I've tried changing timeouts in the owserver interface and in the /etc/owfs.conf file. The values from the conf file show in the web page, and if I change them there they revert immediately, but they don't affect the actual operation. The hardware seems stuck at 15 seconds. --> Any clues what I might be missing to change the w1 update timing? --> I'm seeing the "15 seconds" with a scope on the hardware bus. Does the timeout value set the time between each individual reading, or the time to update the full sequence of 9/10/11/12/temp or whatever else is being separately read? --> Maybe when you use the w1 interface, the "w1" timeout overrides "volatile"? But that is set to 30 and I still get 15... --> If you think this is just "noise" on the readings, is there some place in the owfs system where I could insert some severe digital filtering? It seems kind of tacky to keep querying your output and filtering that for my purposes. --> While I'm asking questions, does anyone know if trying to read DS18B20 devices with 3.3V hardware causes them to read high? Sensors that match my thermocouple thermometer exactly when read in a 5V system read 2 to 6 degrees C higher on the 3.3V BBB. I've verified the logged temperature matches the reported 1-Wire bits. The readings are stable and (except for the occasional periods of low resolution noted above) change smoothly. I have not gone to the effort of trying to determine if this is a fixed offset or a ratiometric error... For my production system with a dozen sensors along 100+ feet of cable, I'm sure a proper 5V interface with active pullup/downs will be required. But I'm surprised I'm getting offsets when testing with local devices! Thanks for any clues, Loren | Loren Amelang | lo...@pacific.net | ------------------------------------------------------------------------------ HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions Find What Matters Most in Your Big Data with HPCC Systems Open Source. Fast. Scalable. Simple. Ideal for Dirty Data. Leverages Graph Analysis for Fast Processing & Easy Data Exploration http://p.sf.net/sfu/hpccsystems _______________________________________________ Owfs-developers mailing list Owfs-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/owfs-developers
