Hi,
I would like to build the DC42 RepRapFirmware from Dr. DC's source code.
I am using two MAX31855 thermocouple chips for bed and hot end temp measurement.
Every now and then (say one or twice in 5 minutes) I see a temp single reading of 2000'C, and of course the firmware sees this and shuts down the relevant heater.
I have tried various power supplies, everything else disconnected, a 1 inch thermocouple, the MAX31855 are well decoupled and have 10nF across T+ and T-
I cannot make the problem happen by flexing the PCB, bashing it, flicking wires etc.
I'm wondering if it's an SPI timing issue, I have yet to check the clock, data and cs signals. Need to dig out my USBee analyser.
A reading of 2000'C is significant, I need to check the MAX31855 datasheet further.
The reason is that I wish to tweak the code in Module Heat.cpp
found in the GitHub repos here
[
github.com]
line 215 onwards......
else
{
// Error condition is not expected to correct itself (e.g., digital device such
// as a MAX31855 reporting that the temp sensor is shorted -- even a temporary
// short in such a situation warrants manual attention and correction).
badTemperatureCount = MAX_BAD_TEMPERATURE_COUNT + 1;
}
I want to be able to tolerate the odd bad reading, not crap out on the first, so change line
badTemperatureCount = MAX_BAD_TEMPERATURE_COUNT + 1;
to
badTemperatureCount++;
This may not help in long print runs.
Maybe implement a rolling average of n readings, discarding highest and lowest, although this might upset PID sampling.
Any other suggestions would be appreciated.
Thanks.
J.
Edited 1 time(s). Last edit at 03/14/2016 10:57AM by JuJuDelta.
