Thoughts on Calibration Cube

After trying to tune everything in pid, e Steps, linear Advance, retraction, temp tower, etc… This is what I have. Dimensions are .04-.08 mm off yet depending on the side with the z within .01 mm.

My issues: sight bulge on 1 side corner that lines up with top and bottom of Y… No idea where to start with that.

There appears to be some visual Z stripes though they look worse then they feel. During running a “tick sound” started (or I noticed) related to Z movement as it appears the machine compensated for bed level as it traveled. Thinking that could be the stripe look as I can feel the tick on the rods. Everything runs up and down smooth on the Z and before tuning I spent a lot of time leveling everything including using a 123 block to verify. So not sure what would have went wrong in a couple hours of test prints.

Open to all feedback on the cube to get it even better. Thanks!!



Forgot bottom

First thing I would try is a bed heat PID tune.

The bottom layer looks very good, your Z offset is set well. The little bump out on the sides at the corner occurring below and above the lettering happens because when it doesn’t have to move in and out to form the letter it has a longer straight path down the side of the cube, and that allows it to accelerate to a higher speed, which changes how it moves around the corner a bit. Not sure how to fix that.

So the main thing that needs work is the banding in the layers. I’ve seen similar happen when the bed heat PID tune is off, so instead of maintaining a stable temperature on the bed, it cycles up and down a few degrees as it tries to get to the set temperature but keeps over and under compensating. This causes the bed to expand and contract a bit each time it cycles and causes the print to move up and down just a tiny bit, which compresses the layers a little bit when it expands upward. Doing a bed PID tune will make the bed heat more stable and minimize the movement.

EDIT: After I posted I scrolled back up and re-read your message and saw “tune everything in PID.” Forgot that by the time I read the whole message. If you did a bed PID tune, how many cycles did you do? I’ve found that the default of 5 cycles isn’t enough. I usually tell it to do 8 cycles and I get better results.

Thanks for the feedback!

As for the PID, I misstated that. I ran the PID for the hot end several time varying between 5 and 8 cycles then took the numbers that were closest to the average of all the tests… I never even thought about the bed PID cause adhesion has been amazing. So I’m glad you mentioned it as that is now on my try next list.

I really liked your acceleration point and hadn’t thought about that. So I went back and looked at some different test prints and other objects I’ve printed as I’ve went with that in mind. It looks like the issue was slightly more noticeable when I bumped speeds up to 40 mm/s, but got hidden some by the linear advance adjustment. However, what I didn’t realize is when I adjusted those things my seams changed how they looked on these cubes and I didnt realize it moved them to IN the letters and I can’t see them so it didnt even cross my mind they were in the letter. BUT above and below the letter they are on a outside corner and the seam blended so well I didnt realize thats what it was. So I think you may be on right track its an acceleration/jerk issue but with starts and ends. So thanks for that, it gives me a different way to look at the issue!

You may also want to look at Linear Advance. I wrote up a post on that and you can find it here: Linear Advance - How to enable this on TAZ Pro or TAZ Workhorse

The idea is that melted filament being extruded has a fractional delay as you start/stop extrusions. But also … as the print head changes speed … it needs to change the extrusion rate and there is a delay there as well. This results in slight under-extrusion or over-extrusion at the start or end of printed segments.

Linear Advance is a feature of Marlin firmware which allows you to tune the firmware to compensate for that delay. They have tool (web page) which allows you to generate G-code to test various values for Linear Advance to find the best settings for your printer/printhead/filament combination.

Little update… did the bed PID tuning with 10 cycles and got very different numbers from what was in there. Adjusted them and after brief battle where using the screen on the taz to change them and save them didnt work (had to connect by computer to get them to save right) I was able to test print again. I think it produced a visible improvement with the banding look, not a huge improvement but enough there is a visual difference on the smooth sides when angled to light.

As for linear advance I have that set at .17 after running the test before the 1st cube. Im thinking though that after PID tuning if my initial LA tests might have been effected?? With this on I have also turned all wiping and coasting functions to 0 as I had read it replaces them, that may be effecting my start/stop to.

It might also be time to dry this polylite PLA roll I’ve been using for testing.

I didn’t think to mention that I always do the PID tune through the console. It’s much easier to keep track of what is going on and gives you more options since you can specify all the parameters.

What has worked best for me is to start with a cold bed and do at least 8 cycles. I usually print PLA with the bed around 60 degrees, so in the console I’ll do the command M303 C8 E-1 S60.

The M303 has the U parameter that says to use the result, not just print it to the console. I’ve tried that and it doesn’t seem to work reliably, so I just let it print the numbers to the console, then I re-enter the results with the M304 set bed PID command and save them with an M500.
(Hopefully I remembered those commands correctly.)