M593 Dynamic Acceleration Adjustment no effect

AlexLin

Hi
I loaded fw version 2.02(2018-12-24b1) and tried the M593 Dynamic Acceleration Adjustment but I cannot get any effect see image below I tried the some frequencies and also adjusted my belts (in case that had some effect) but all my cubes turned out the same. I made sure the value is not reset ,(I set it in my start.g).What am I missing or doing wrong?
I measured the distance between 6 waves.Divided it by 6 , resulting in 0.96mm. And divided the print speed of 30mm/s by it resulting in 31.25Hz....
The ringing doesn't look as bad as it does in the picture, I just upped contrast and clarity to make it more visible.

toskium

Same here, I measured my ringing frequency at 52.2Hz and set M593 accordingly in config.g, but it doesn't make a difference in the output.

dc42

What XY jerk setting are you using? Dynamic acceleration adjustment has less effect when using high jerk. I'm looking at changing the jerk algorithm to reduce the use of jerk at sharp corners, so that dynamic acceleration adjustment can work better.

AlexLin

these are my setting, don't think jerk is high, or is it?
M92 X640 Y640 Z3200 U3200 E800 ; Set steps per mm
M203 X12000 Y12000 Z500 U500 E1200 ; Set maximum speeds (mm/min)
M201 X800 Y800 Z250 U250 E250 ; Set accelerations (mm/s^2)
M566 X600 Y600 Z20 U20 E600 ;jerk

gnydick

Only effect I experienced was slower printing.

AlexLin

Extra info..it is a core xy with direct exruder(titan and e3d compact motor, i.e. some weight)

briskspirit

Tried M593 too with different jerk settings - saw nothing. Maybe we calculating frequency wrong? I've tried from 47 to 63 Hz (with 2Hz steps).
But in my case I think I hear high frequency resonance and can't find where is it going from

toskium

@dc42 this is my current jerk and acceleration config:

M566 X720 Y720 Z12 E120 ; Set maximum instantaneous speed changes (mm/min)
M201 X5000 Y5000 Z250 E450 ; Set accelerations (mm/s^2)

Don't think it's too aggressive, do you?

dc42

One reason why DAA may not work is that the acceleration limits are set too low. The optimum acceleration may be higher than the maximum you have specified in M201. Try increasing acceleration to 2000 or 3000. For the purposes of testing DAA I suggest you reduce jerk to 100 or 200; but don't try to print curves with this setting.

AlexLin

Thanks for the hint dc
It had an effect. The acceleration was very conservative based on my old cartsesian....
What do you mean (and why) not use such a low jerk on curves? Do you mean G2/G3 movements? Do slicers generate them?
Now I'm seeing an effect I can try to tune the frequency

briskspirit

@alexlin Can you show your accel/jerk/max speed settings and M593? Because I still can't understand what am I doing wrong... still can't get rid of rippling at any frequency...

AlexLin

@briskspirit
M350 X64 Y64 Z64 E16 I1 ; Configure microstepping with interpolation
M92 X640 Y640 Z3200 U3200 E800 ; Set steps per mm
M203 X12000 Y12000 Z500 U500 E1200 ; Set maximum speeds (mm/min)
M201 X3000 Y3000 Z250 U250 E250 ; Set accelerations (mm/s^2)
M566 X300 Y300 Z20 U20 E600 ;jerk
M906 X1200 Y1200 Z1200 U1200 E1400 I30 ; Set motor currents (mA) and motor idle factor in per cent

and
M593 F25

dc42

@alexlin said in M593 Dynamic Acceleration Adjustment no effect:

What do you mean (and why) not use such a low jerk on curves? Do you mean G2/G3 movements? Do slicers generate them?

I don't mean G2/G3 moves, I mean sequences of short straight segments that approximate a curve. If the XY jerk is set too low then the printer will slow down at the junctions between segments.

I will look at adjusting the DAA algorithm to work even when the acceleration limits are low.

AlexLin

@dc42 Thanks , reading your responses in the pressure advance thread explained this

Wyvern

Very cool stuff, I gave this a shot last night and after 2 attempts it killed all the ghosts, along with pressure advance (direct feed) I finally have a perfect single perimeter cube.

My question is so long as I don't set my Acc/jerk too low or too high, the dynamic Acc should stay about the same if I adjust them?

dc42

@wyvern said in M593 Dynamic Acceleration Adjustment no effect:

My question is so long as I don't set my Acc/jerk too low or too high, the dynamic Acc should stay about the same if I adjust them?

I haven't done extensive testing of the effect of changing acceleration and jerk, but this is what I expect:

If you set the acceleration too low in M201 then at higher print speeds, the required acceleration computed by DAA will be above the M201 limit, so DAA won't be applied. The maximum print speed for DAA to be applied at a right angle corner will be about:

V = a/f + j

where a is the acceleration, f is the frequency you are cancelling, and j is the jerk expressed in mm/sec (so 1/60 of the value in M566).

Increasing jerk will reduce the effectiveness of DAA, because it changes the overall acceleration profile from the constant acceleration assumed by the DAA algorithm.

Patrick K

So I am having issues making M593 effective. No matter what F I use, I don't see anything changing.
I bumbed up my vel, my acc and lowered jerk.

From --> to
Vel = 45 to120
acc = 500 to 2000
jerk = 900 to 300

from there my result got instand better even without appling M593. So my ghosting is defenatly depending of frequency. I calculated M593 and depending on my acc (even used 8000 on acc) i got results from M593 F64 till F154.
I ran test where I changed F every mm from 30 to 180. Nothing changed. I did a lot of testing, runing different acc to jerk to frequence matrixes. NOTHING

I figured something only very late and I will have to do some test, but I guess I realized what my issue is.
I am using a high pressure advance value of 1.2. Which gives me charming results.
Therefore I also raised my jerk on E to compensate my high S1.2. So I went from 250 up to 500.

but that seems to little. I think the PA is reducing the Jerk of E so much, that it will influent my motion totally and even impact M593.

My last result I did I ran different setups in changing my motion parameters. After running Vel 120 and acc from 2000-3000 I decided only on the last mm on top to turn of PA and see, no ghosting. not a little. Of course crapy edges but no ghosting.

So I tried with higher jerk for E in PA = 1000. that got me close to the edge of my extruder unit not tu junk or eat the filamet.

So I wonder if there is a way, for me using PA and DAA. righ now I can't see it and I think that might be the issue why most user won't see an effect on DAA. At least with bowden setups as I have.

I assume, to calculattion of the motion profile to avoid certain frequency, is using the jerk and acc values set for x and y. But the actual movement is dictated by the Extruder depending in the PA setting.
Only my assumption.

And just for real lovers, I am running an Omerod2 modified to a monster. (Duet2Wifi and latest firmware to this date)

dc42

@patrick-k said in M593 Dynamic Acceleration Adjustment no effect:

I calculated M593 and depending on my acc (even used 8000 on acc) i got results from M593 F64 till F154.

Ringing frequency does not depend on acceleration. Ringing pitch does depend on the speed at which you print the perimeter (frequency = speed/pitch). At low accelerations, the pitch of ringing will be compressed close to the corners because the head isn't up to speed yet, so ignore the pitch in that area.

Dav

Hi
I calculate my hz but where I've to put the m593 command inside the conf.g?

bot

@Dav You can do that. But, for testing, you can just send it via the web interface so you can enable it only when you want to test it.