Am I doing something stupid?

DjDemonD

No it doesn't seem to make a difference, I wouldn't expect it to really.

number40fan

Did you send M591 D0 C3 before the M591 R-50?

DjDemonD

No, I have M591 D0 P3 S23.5 C3 R50 E3.0 in config.g, although I have also tried with R-50 instead, which didn't make a difference.

then I sent via console M591 R-50 to set calibration mode as per https://duet3d.com/wiki/Duet3d_Filament_Monitor

If I remove the M591 from config.g and send

M591 D0 C3

I get

M591 D0 C3
Extruder drive 0 has no filament sensor

number40fan

Just reading the directions, it says to send the M591 D* C* before the R-50. I would assume you still need the full M591 in the config.g.

DjDemonD

Just tried that, now getting further.

So I put the M591 D0 P3 S23.5 C3 R50 E3.0 back in config.g

Then sent M591 D0 C3
Then sent M591 R-50

Then printed an object.
Now when I send M591 D0 I get back just the parameters I entered in config.g but if I send M122 then at the bottom I now get

Extruder 0 sensor: angle 293.9, ok, 28.05mm/rev, tolerance +8% -10%

Which is progress. I assume it has derived the 28.05mm/rev value, and determined the tolerance, not sure what +8% to -10% means, but I suppose I enter a tolerance into M591 in config.g which allows for this measured tolerance and some overhead to prevent false positives, i.e. 20%.

Now in config.g it looks like M591 D0 P3 S28.05 C3 R20 E3.0

I'll try it now but stop the filament moving and see if it works to pause the print.

dc42

It shouldn't whether you configure the filament sensor in one command or more than one, but you need at least the D, P and C parameters to configure it.

When the firmware is in calibration mode, a certain amount of filament has to be extruded during the print before calibration data starts to become available.

DjDemonD

I think I have it calibrated now. As for the tolerance, what value would you recommend I enter for R parameter based on "tolerance +8% -10%"?

DjDemonD

Might adjust the wiki slightly as its not very clear how to do this.

dc42

@DjDemonD:

I think I have it calibrated now. As for the tolerance, what value would you recommend I enter for R parameter based on "tolerance +8% -10%"?

I suggest at least 20%. You may find that you need more, depending on whether your test print covered worst-case conditions.

DjDemonD

Thanks thats what I figured. Why does the wiki mention " Use a positive or negative sign as reported by M591" surely its +/- x%?

DjDemonD

Okay so i gave it test, reached in and undid the idler. After around 3-4 seconds it went to pause and displayed:

I re-tensioned the idler and hit resume and maybe missed around 1/4 of a layer of infil on a 20x20 cube, so quite recoverable.

It does work.