thermistor noise

T3P3Tony

@jens55 yes possibly. what steps/mm are you running your extruder at?

jens55

@T3P3Tony
M92 E25.3:51.875:51.875:51.875
Note that prints were done on the first 3 boards. The first tool uses a different extruder from the other two but blips showed up just as they did on the other tools.

The above was from config.g, below is what I get back from sending M92 to the printer:
m92
Steps/mm: X: 100.000, Y: 100.000, Z: 3200.000, U: 30.578, E: 404.800:830.000:830.000:830.000

First extruder is an LGX, others are BMG clones

T3P3Tony

@jens55 can you try using 16 microstepping with interpolation (it looks like you are using 256 microstepping?) that could well be the issue with the hiccups.

jens55

@T3P3Tony , from config.g: M350 E16:16:16:16 I1 which I am interpreting as 16 microstepping.

oliof

are you running .9 degree steppers on the extruders?

jens55

@oliof, I actually do not know the degrees/step for the two type of motors but the e-steps of the two kinds of motors are different while the temperature blips happen with both.
I don't quite get the connection between degrees/step or microstepping setting as it pertains to either the temperature blips or the hiccups.

oliof

it's just a mind goblin that's vexing me looking at possible "overload" sources on toolboards that may cause these blips. I have nothing to back this, so it's safe to ignore (-;

jens55

@oliof .... 'mind goblin' ..... must remember that .... priceless

T3P3Tony

@jens55 yeah lets address the source of the hiccups (likely to be too high microstepping) and then see if the blips are still there.

What does M350 report when run from the console?

See:
https://docs.duet3d.com/en/User_manual/Reference/Gcodes#examples-104

Depending on which comes first, M350 and M92, they influence each other

jens55

@T3P3Tony

Microstepping - X:16(on), Y:16(on), Z:16(on), U:4(on), E:16(on):16(on):16(on):16(on)

T3P3Tony

@jens55 hmm ok, not sure how that can be the same and M92 be so different.

Sorry to be a pain but please post the full config.g and confirm that the output on the console is:
M350
Microstepping - X:16(on), Y:16(on), Z:16(on), U:4(on), E:16(on):16(on):16(on):16(on)

And
m92
Steps/mm: X: 100.000, Y: 100.000, Z: 3200.000, U: 30.578, E: 404.800:830.000:830.000:830.000

jens55

@T3P3Tony
For ease of calculation, the microstepping was applied after the M92 command. I believe this was arranged like that in the Jubilee config files and I kept a good portion of the example files including that section.

I can confirm the M350 and M92 outputs.

Here is my rather verbose config.g file:
config(2).g

I see that I didn't update the comments on my M570 commands, please ignore

oliof

@jens55 why does U only have 4x microsteps?

T3P3Tony

Ok so now I see. This config is made to calculate everything from first principles so hence the weird numbers you reported before:

M350 X1 Y1 Z1 U1 E1    ; Disable microstepping to simplify calculations
.....
M92 E25.3:51.875:51.875:51.875
.....
M350 E16:16:16:16 I1   ; 16x microstepping for Extruder axes. Use interpolation.

That together will give you:
25.316 = 404.8
51.87516=830

Which are what is reported by M92 when its running.

So scratch the idea that the microstepping is too high and that directly is causing hiccups.

Can you do the following:

First run a test print to see if you get hiccups, every time you send M122 Bnn the hiccup counter is reset. so if you send it multiple times, keep a record of each one.

Then generate a simple set of test print jobs that moves back and forward (e.g. between X10 and X100) for enough times that you would definitely see the blips if it was a print job. (I would use a text editor to cut and past the line in blocks)

Run with movement, the hot end at temperature, but no part cooling fan and no extrusion. Look for blips and also grab M122 and M122Bnn
Run with the movement, no part cooling, but extrusion (you can have filament not fitted to avoid having a bit of spaghetti to clear up and wasting filament). Look for blips and hiccups and also grab M122 and M122Bnn
Run test 2, but manually setting the part cooling fan to the levels used by your slicer, and looking to see if there are blips as its on and off, or from (say 80% to 100% or other combinations). Look for blips, hiccups and also grab M122 and M122Bnn

Thanks.

jens55

@oliof
U is the actuator that locks the tool to the carriage. It does not need to be accurate. In fact, the actuator connects to a clutch mechanism that puts a fixed pressure on things so the position of U only needs to be sort of in the ballpark to do it's job.

jens55

@T3P3Tony
Thanks, I will do some test runs tomorrow (well later today as it is 2:30 am here)

jens55

@T3P3Tony
Unfortunately I don't think this is going to be too helpful:

I did a test print and got just shy of 7000 hiccups

I then did the tests you asked for - none showed hiccups (well one had 1 hiccup)

test#1, movement, nozzle at operating temperature, no extrusion .... absolutely flat temperature curve
test#2, movement, nozzle at operating temperature, extrusion at 4.4 mm3/sec .... only two blips over about a 6 minute test run but still no hiccups, Frequency of blips are much less than during an actual print run.
test#3 - same as test #2, very rare blips and no hiccups

I then printed the same job I started with but with no filament loaded. I got 6586 hiccups and a bunch of blips on the temperature graph.

I ran test#2 again, I saw a number of blips on the temperature graph but hiccups was zero again

lastly, I ran test #1 again - flat temperature line, a single tiny blip which was so small that I missed it in my initial look and no hiccups

T3P3Tony

@jens55 thanks for doing those tests. I think it's right to focus on the cause of the hiccups and see if removing that also solves the temperature blips.

Can you share the shortest print file you have that shows high hiccups and temperature blips please.

jens55

@T3P3Tony
This is a file for a knob used in woodworking fixtures. Not necessarily the shortest but it is what I used in my tests.
knob.gcode

jens55

This is a 'bump' for this thread.
It is no longer a 'thermistor noise' issue but one of 'tons of hiccups'.
If I should open a new thread I can do that but I thought I could continue this here as a lot of tests were done to narrow down the 'hiccup' issue.
To recap, normal/ordinary prints have thousands of hiccups and although the prints are completing correctly, I feel that it is important that the source/reason for the hiccups is tracked down.
I have previously reported that on occasion the printer skips a lot of steps at the end of a print as it is moving to the parking position. This is a rare issue that I can't reproduce but I would not be surprised if this somehow relates to the hiccups as well.