Duet 3 G31 'Heater Number'

dc42

@Danal said in Duet 3 G31 'Heater Number':

@dc42 said in Duet 3 G31 'Heater Number':

I was intending to remove the temperature compensation facility because I didn't think anyone was using it; but when I became aware of Pinda probes with built in thermistors, I changed my mind.

Should we remove that statement from the dozuki?

I already have.

dc42

@Garfield, first check that you are getting a sensible temperature reading on sensor 2. Name the sensor in the M308 command using an A parameter, then look in the Extras tab of DWC to see the reading.

If you do get a good reading but you still get that error message, try adding a short delay between the M308 command that creates sensor 2 and the G31 command that uses it, for example by inserting G4 P500 command between those lines in config.g.

[Edited]

Garfield

I'm on it - I shall let you know.

Garfield

Same error with the delay, the sensor is reading - showing 23.9 perhaps a tad high but I hadn't got as far as verifying calibrations but it is definitely reading (called it PINDA) .

Garfield

This seems to be working now BUT ....

It didn't work following a board reset, power cycling the Duet seems to have resolved, the error has disappeared.

dc42

Thanks. I've rewritten that code in RRF 3.1 so that the delay should not be needed.

Garfield

Well unfortunately no code change but the error is now back ... so I still get the error.

Definition code as confirmation

; PINDA - Z-Probe
M308 S2 P"temp2" Y"thermistor" T100000 A"PINDA" B4725 C7.06e-8 ; configure sensor 2 as thermistor on pin temp2
G4 P500
M558 P5 C"!io5.in" H2 F120 T6000                      ; set Z probe type to switch
G31 P1 X-57 Y-7 Z2.0 C0.2 S40 H2                      ; set Z probe trigger, offset, trigger height, temperature compensation

error has returned ...

G31 Error.png

antlestxp

@Garfield how far did you get with setting up your pinda. I have the v2 installed currently along with the temperature readout. I am seeing huge changes in probe height depending on temperature. Not sure how to get it to actually compensate for the temperature it reads.

Garfield

It is working pretty well now, took a while to dial in. The compensation is done in the G31 line, the C0.2 S40, which is the compensation I'm applying and the temperature I base it on, it seems mostly to work although I do occasionally see glitches in the matrix when it comes to the initial layer. If I use an Aluminium compensation (expansion coeefficient) it doesn't seem to work as well.

The trigger height, not the Z2.0, but the actual height that triggers the probe does seem to vary regardless of temperature change, as yet I've found no solution. Basically if I run concecutive probes in the same spot the trigger height will change by a very tiny amount until you get to the 10th iteration or so and then it seems to stabilise.

I know my bed movement is consistent and accurate as I have measured it with a dial test indicator.

I now have it mostly working so I haven't pursued this much - suppose I should - but the pursuit of perfection is a never ending pursuit

antlestxp

@Garfield I am going to see if your c0.2 works for me. I can't find any details on how it is calculated as well as the s40. Right now I left out the C and S command and only home or probe the bed at room temperature. I am also using the the temperature of the pinda as reference. I wonder if it would be best to use the bed. I was hoping to be able to get it to home with at least some heat in the bed. If I try and run another print with the bed still warm I risk the nozzle hitting the bed.

Garfield

The S40 is basically the temperature at which the Z height is correct with no 'adjustment' for temperature - as in the calibration point with no adjustment needed.

I thought the co-efficient would be the co-efficient of expansion for the bulk of the materials - in my case mostly aluminium but that as a thermal expansion co-efficient of 0.000023 (m/m per Deg C), which I tried but it doesn't function as expected, I keep telling myself to inspect the code and see how the formula is applied but never got around to it.

Basically what I tried was to set S to my room temp (21), calibrate Z, then heat my bed to 80 Deg C, I expected the trigger height to be reduced by 0.008 (ish) but the trigger height simply isn't repeatable enough to claim any accuracy - and the number so small that I stopped caring to be honest.

I guess at some point I will circle around to it again.