Page 2 of 2

Re: Unexpected y-axis shift mid print

Posted: Sun Sep 20, 2015 9:00 pm
by Vandal968
This is absolutely a weight/acceleration related issue, and if you haven't added a heavy aluminum bed, the default values of 3k for M201, M202 and M204 work perfectly fine for a part weight up to about 45grams. This is consistent across many different parts. Dropping those values to 2k (from 3k) allows part weights up to about 150gr without shifting, and with a value of 1.5k, I just printed an almost 16hr print (241gr, more than 1/2 a pound) with no shifting. I have never altered the x/y movement speed of 18000 since weight has almost no effect on top speed, but has a large impact on acceleration.

For those that think weight is no issue, then why it's the Y axis that always shifts rather then X? The reason is that the Y-axis becomes heavier and more difficult to rapidly accelerate as the parts grow. The X-axis only has to accelerate the extruder and its weight is fixed. The maximum part weight before shifting matches up perfectly when plotted against the acceleration values.

Yes, I know about the Marlin bug, but this firmware is loaded with the earlier IDE and this is a separate issue. And yes, simply reducing the values to 1k makes it go away at the expense of printing speed, and this is the reason why that works and reducing the acceleration values below what is necessary to ensure a successful print just wastes time.

cheers,
c

Re: Unexpected y-axis shift mid print

Posted: Sun Jun 03, 2018 1:43 am
by Vandal968
Updating a very old thread.

I printed for a few years with no issues by simply reducing the acceleration values as described in the post above. I just had to rebuild my computer last week, and performed a fresh install of S3D, using the default profile. Surprise, surprise, got the same Y-axis shift on a large part that I hadn't seen once in almost 3yrs. Looked up this thread to find out what values to use (listed one thread up). It took me a moment to remember where the values get plugged-in (into the starting script, under the scripts tab). While I was digging, I found another description of a related acceleration issue, so adding that info here as well to make it easier to find in the future.

FYI, the part I was printing just now shifted at about 55gr which is consistent with all of my my earlier observations.

https://forum.simplify3d.com/viewtopic.php?t=6049

These are the lines to add to start script:

M201 X1500 Y1500; // max accel print (default 3k)
M202 X1500 Y1500; // max accel travel (default 3k)
M204 S1500; // default accel for normal moves (default 3k)

cheers,
c

Re: Unexpected y-axis shift mid print

Posted: Mon Jun 04, 2018 2:40 pm
by insta
Vandal968 wrote:
Sun Sep 20, 2015 9:00 pm
This is absolutely a weight/acceleration related issue, and if you haven't added a heavy aluminum bed, the default values of 3k for M201, M202 and M204 work perfectly fine for a part weight up to about 45grams. This is consistent across many different parts. Dropping those values to 2k (from 3k) allows part weights up to about 150gr without shifting, and with a value of 1.5k, I just printed an almost 16hr print (241gr, more than 1/2 a pound) with no shifting. I have never altered the x/y movement speed of 18000 since weight has almost no effect on top speed, but has a large impact on acceleration.

For those that think weight is no issue, then why it's the Y axis that always shifts rather then X? The reason is that the Y-axis becomes heavier and more difficult to rapidly accelerate as the parts grow. The X-axis only has to accelerate the extruder and its weight is fixed. The maximum part weight before shifting matches up perfectly when plotted against the acceleration values.

Yes, I know about the Marlin bug, but this firmware is loaded with the earlier IDE and this is a separate issue. And yes, simply reducing the values to 1k makes it go away at the expense of printing speed, and this is the reason why that works and reducing the acceleration values below what is necessary to ensure a successful print just wastes time.

cheers,
c
This sounds like a fantastic use of a post-processing script to dynamically lower the acceleration as the print gets heavier / taller.

Re: Unexpected y-axis shift mid print

Posted: Thu Jun 07, 2018 6:10 pm
by Vandal968
Ok, this pains me but I have to circle back on this. Changing ONLY the acceleration values(down to 1500) solved the shifting problem for me, and kept it solved for years. I just rebuilt my S3D computer and the issue returned when printing a large battery organizer. I'm making a bunch of these and they're all the same overall size (7.5"x 6.5") but with varying cutouts for different types of batteries (AA, AAA, C, D, 18650, etc.)

When I printed the first one for D cells, I got a Y layer shift, so I came on here to jog my memory, found this previous thread, remembered the settings and so updated my accels from 3000 to 1500. I then successfully printed AA, AAA and C organizer trays with no shifts. Then, I tried to print the 18650 tray. Got a shift. Cleaned, lubed the rails, increased Z-hop, tried again. Got a shift. Checked belt tension on Y, a bit loose and it was a big shift (4mm) so tightened belt. Tried again, big shift. #**((#$(#$!!!!! Went back in, further reduced accel to 1000 for all three axes AND reduced the XY speed to 9k per Jimc. NO shift.

So, it looks like while only changing the accels to 1500 mostly solves it (and in-fact I went YEARS with no shifting), some nearly-identical parts are apparently more prone to shifting possibly as a function of the paths that the slicer chooses and the recommended values that Jimc posted are a more robust cure than what I was doing previously. I'll keep an eye on it and report any new developments.

cheers,
c

Re: Unexpected y-axis shift mid print

Posted: Sat Nov 23, 2019 2:09 am
by ensign
Thank you so much for posting the solution to this years-old issue I have had. Without question, this is a weight and acceleration issue - in this case. Don't be fooled by the ability of the steppers to throw a heavy metal bed around. At some point, the added material on the bed is going to become the proverbial straw that broke the camel's back.

My Y-axis layer shift problem had been occurring infrequently for years (I own one of the first M2's), but didn't happen often enough to keep me from printing. I print mainly smaller pieces. I'll never forget my first large print when the famous T-Rex skull became available. I hit my first y-axis shift. I tried passive cooling, active cooling, slicing repairs, layer observations with a camera, and I could never figure out the issue. I would reduce the part count on the bed, since the lower pieces had printed just fine, the problem would disappear on the reprint, and I would forget about it for a year.

At some point, to prevent ringing, I had slowed my prints down in S3D. I printed for a very long time like this, and I only recall a couple shifts during that time. Earlier this year, I updated S3D and started seeing shifting on larger parts/part counts. Following this thread, I reduced my speeds from the S3D defaults to:

default printing speed: 1000 mm/m
outline underspeed: 60%
solid infill underspeed: 80%
support structure underspeed: 80%
x/y axis movement speed: 9000 mm/m
z axis movement speed: 1200 mm/m

Printing at these values eliminated the y shift. Restoring the default values (18000 mm/m x/y axis movement!) brings the layer shift back. I can reproduce this in back-to-back prints using the same gcode.

Thanks again!