M593 DAA
-
With reference to the documentation for DAA (Dynamic Acceleration Adjustment). Can someone please give some context to the following:
DAA is most useful to avoid exciting low-frequency ringing, for which S-curve acceleration is ineffective and may make the ringing worse.
What is low frequency? What range of Hz?
'High jerk' has also been reported to effectively disable DAA. What is high jerk? Is it a fixed value relating to DAA or relative to your print speed and/or acceleration?
-
@DocTrucker So source code allows values between 4Hz and 10kHz. But I think to remember that it was said it's most effective with up to double-digit values. But I can't find the source of this anymore.
In regards to jerk I have no idea.
-
Thanks. I'm searching through for an introduction of DAA or some kind of feature discussion that expands on the gcode reference. Is there such a thing?
-
@DocTrucker The DAA feature came from the S curve discussions. David decided against a general s-curve feature because he found out that it's more important to control specific vibrations (ringing), which are printer specific. The result is the DAA setting.
My little repository says you can find detailed information here:
Klipper: https://github.com/KevinOConnor/klipper/issues/57
Duet:
○ https://forum.duet3d.com/topic/5951/periodicity-of-ringing/39
○ https://forum.duet3d.com/topic/2629/s-curve-sinusoidal-jerk-acceleration
○ https://forum.duet3d.com/topic/5513/s-curve-in-marlin-2-seems-like-a-huge-quality-improvement
○ https://forum.duet3d.com/topic/4802/6th-order-jerk-controlled-motion-planning
○ https://forum.duet3d.com/topic/2151/acceleration-as-a-function-of-speed/7
reddit: https://www.reddit.com/r/3Dprinting/comments/9w7psw/dynamic_acceleration_adjustment_is_pretty_amazing/
Duet document M593: https://duet3d.dozuki.com/Wiki/Gcode#Section_M593_Configure_Dynamic_Acceleration_AdjustmentThere were a lot of discussion what jerk is, whether to implement s-curve etc.
The reddit user created a thingiverse and found 75 and 80 Hz for his fast printer,
closed S-Curve acceleration? #57
-
Thank you!
I will knock acceleration up to 3000, jerk to 600 (mm/min) and get the before square vase underway...
-
Intermediate report!
I tried a few cracks at DAA but it had no effect.
I tried upping the acceleration to 6000 and to my surprise the machine didn't loose steps, but the ghosting/ringing predictably worsened. It also showed up that my instant speed and max speed settings on the extruder were way to high. I've needed to drop instant speed to 1000mm/sec^2, and max speed to 2100mm/min.
I'll try later today but I'm a little unsure on how your are supposed to balance all of the instant speed, acceleration, and DAA settings. Trial and error shall continue!
-
@DocTrucker In one of the threads mentioned was a method to find out the frequency by measuring the ringing distance and taking into account the specific belt/pulley/... properties. The thread was about testing and reporting ringing and printer properties. I think it was this here: https://forum.duet3d.com/topic/5951/periodicity-of-ringing/62
Another possibility is, that your ringing has maybe another cause than vibration. Some unbalanced pulley (out of center) or similar. I don't think one can resolve this with DAA settings if this is the cause.
-
@JoergS5 thanks for your reply. Sorry I wasn't clear on what I meant. Put another way you would normally set instant speed and acceleration at a level that is low enough to cause minimal ghosting. In essence DAA is a tool to produce better parts quicker. So how much ringing/ghosting can DAA be expected to cancel out?
I do also appear to have a belt issue that is causing another constant ripple on the surfaces parrallel to y (bed slinger), and I am not expecting DAA to cure that.
-
@DocTrucker - OT: You have a lot of different printers, so you can compare the print results very good. I never thought about this aspect and it's a reason to build multiple printers. Example: printing a vase, a polar printer may be best etc.
-
@JoergS5 the multiple machines are kind of my problem too! My original Ormerod (personal rather than the work one in 2014ish) produces nice parts in 0.3mm layers although doesn't like overhangs. Since then with my other printers I've tried to better that machines results. Extrusion quality and general part finish have been improved but I struggle with both the belt (2mm pitch) ripple and ghosting.
I'm currently looking at a test sequence to try and figure out best speed, instant speed, and then acceleration. I think vase mode prints are probably the way to go for these. So planning:
Speed:
One vase mode with- Speed: 65 70 75 80 85
Instant Speed:
5 octagonal Vases...- X=400, Y=400 500 600 700 800
- X=500, Y=400 500 600 700 800
- X=600, Y=400 500 600 700 800
- X=700, Y=400 500 600 700 800
- X=800, Y=400 500 600 700 800
Acceleration:
5 octagonal Vases...- X=500, Y=500 750 1000 1250 1500
- X=750, Y=500 750 1000 1250 1500
- X=1000, Y=500 750 1000 1250 1500
- X=1250, Y=500 750 1000 1250 1500
- X=1500, Y=500 750 1000 1250 1500
...then go back to DAA!
-
@DocTrucker The Ormerod design is interesting, because all mechanics is on one side, accessible from the other sides. There was an interesting thread on reprap https://reprap.org/forum/read.php?177,711747 which discussed some construction topics.
-
My first effort was in reworking the Ormerod into a more stable design. I hit up against belt ripple issues, and needed to replace the belt reduction on the z-axis with a worm gear reduction. I shelved it to get the P3's running as I had thought that would be a quicker route to a machine capable of running finer nozzles or higher speeds - in other rigid and stable.
https://reprap.org/forum/read.php?340,839175
Edit: I'm aware of many area for improvement on that Custom Cantilever. I'm not opening it up for critique just yet.
-
If you talk of nozzles, I willl try matchless from Solex. They have multiple input tubes in the nozzle (I bought the 1.0 mm one, it has 3 tubes) and shall be able to print faster, because the tubes heat up the filament faster.
