Heater fault, spurious spikes [SOLVED]

paboman

I have a 4 wires PT100 (E3D 2 wires modified to 4 wires as per DC42 blog) which works fine but some times I get a random spikes in the temp graph that I am unable to reproduce.
Touching and bending the cables does not produce that effect.
All wrings are shielded except the last 30cm that goes into the effector as they are too stiff with the shielding while the pt100 4 wire cable is not shielded at all.

Tonight the printer stopped after many hours with the following message:

06:50:48 Printing paused
06:50:45 Warning: Tool 0 was not driven because its heater temperatures were not high enough or it has a heater fault
06:50:45 Error: Temperature reading fault on heater 1: sensor short to other wiring Resume-after-power-fail state saved

Is there a way to see / save the full temperature logs from the beginning of a print session ?
Can a shielded 4 wires cable solve the problem ?
Where the shielding should be connected ?
It will not be possible to make the shielding goes all thru the pt100 probe as I need to solder the 4 wires of the cable to the 2 wires, so at least the first 4/5 cm after the probe will be unshielded where passing near the heating cartridge wires.
I am worried that this can mak the whole shielding useless, any thoughts ?
One user in the forum solved this problem by changing the PWM frequency of the fans to 10khz, I done this for the layer fans which were off during the last print that gave an error so now I want to do this for the always on fan (the hotend cooling fan)
How to change the PWM of the always on fans ?

thx!

dc42

@paboman said in Heater fault, spurious spikes:

sensor short to other wiring

That suggests it is not a noise issue. Check that there is no way for the PT100 cartridge wires to short to the heater block.

paboman

@dc42
that is quite impossible, maybe is the pt100 damaged somehow (but i dont think so), wires are well insulated.

Can this spikes be caused by an intermittent connection in the connectors?
one is the DuPont on the daughterboard, then I have a second one along one of the towers, which is a 4 pole mini xlr (http://www.switchcraft.com/Category.aspx?Parent=806)

Moving them does not produce any spikes at all

paboman

@dc42 just checked again, none of the sensor wires touches the heater block

deckingman

As a work around, you could "de-sensitise" the fault anomaly detection by increasing the P and/or T parameters in M570 https://duet3d.dozuki.com/Wiki/GCode#Section_M570_Configure_heater_fault_detection

paboman

@deckingman
thank you!

now I have changed the 4 wires cable of the pt100 with a shielded one as this is the only thing I can do.
My plan is to run the printer and later check if random spikes are still there but i cant look at the graph forever so i would like to record the temperature graph somehow.
I know repetier host has a nice interface for the temps but how to connect the duet wifi to repetier ? what is the ip port ?

sonderzug

Hi all,
I've had a somewhat related issue with both PT100 and Type K sensors, leaving me unable to print at the temperatures neccessary for the materials I use. I've documented this in another thread, unfortunately without a working fix at this moment (https://forum.duet3d.com/topic/4966/highly-erratic-temperature-readings-above-certain-temperature). It seems that the insulation inside the sensor cartridges is not stable at higher temperatures (>300°C). While this doesn't really apply to you (you're printing at pretty standard temperatures as it seems), a failing cartridge insulation and short to the heater block might still be the case?

Best regards, Niklas

paboman

@dc42 I have changed the pt100 cable with a shielded one but I am still missing something...

I am sure the PT100 cable is picking up estruder noise as the temperature graph became noisy as soon as E motors is energized

Routing the cable externally to the printer also makes the graph look clean but I cant route the cables in a different way.
Cant understand this behavior as all my cables are shielded, steppers cables are twisted and shielded....what am I doing wrong ?

Look only at the noise, the temperature was raising and falling because of me turning on/off the heater to see if it has any correlation but it hasn't

Shield of the extruder stepper is connected to the extruder motor case and to nothing on the other end
The shielding of the PT100 is connected to nothing on both ends

dc42

@paboman, try connecting the shield of the PT100 cable to ground. There is a ground pin on the daughter board OUTPUT connector, it's the corner pin of that connector closest to the INPUT legend.

paboman

@dc42

Can I use the negative terminal of the main input ?
I mean the GND on the big green connector

thx!

paboman

@dc42

Now its better but still jumping up and down in the 2°C range
I will run the printer to see if it spikes again

Is there any way to filter / flatten the readings even more or this is the best I can achieve ?

thanks again David!

paboman

UPDATE:
SOLVED
I made better connection at the shield line and now my temps are within 1°C