Trying to tune Vcore 400mm
I attached the ADXL345 accelerometer to Vcore 400mm and ran the macros in RatOS for belt tension and input shaping.
I gather working on belt tension should be done first, but my graph looks nothing like others, so what does it mean? Too tight, not tight enough?
25 Replies
I incrementally tightened the belts and retested but it made little difference to the shape of the graphs above. So decided to take the tensioning blocks (from the EVA toolhead) apart and sand them down a bit so they slide easier in the toolhead. Reconnected everything and retensioned the belts and have new graphs as follows.
Still not sure if this is any good or not. Secondly I don't understand why the resonances_X graph shows no X or X+Y+Z lines?
Looking at other peoples graphs the belt tension from 0 to 25 Hz is always pretty flat, but on mine its high straight away. What does that mean?
I dont have much experience with this either, but what I read is, when the graphs look very different from the goal, usually something mechanical is not correct. Loose screw etc. And did you watch this video from the pins? Explains a lot. https://youtu.be/M-yc_XM8sP4
3D Printers & a Whiteboard
YouTube
How to Read and Analyze Input Shaper Graphs Generated by Klipper
This video is supposed to be a brief overview on what the input shaper graphs can tell us. It is the first video of a series of videos which will give a more in depth view on the kinematic system and input shaping in Klipper.
As always: feel free to jump to those parts, that help you with your problems:
00:00 Intro
00:57 What we can learn from...
Yes, I have watched that video, but it doesn't provide me with any ideas of what to do.
Is the accelerometer attached properly? Screws decently tight?
The response is all over the place
Yes I have checked the accelerometer is tight and all other screws on the EVA head (and the frame too).
I think the tension on your belts is low, do you happen to have an Android phone?
If so, install "Spectroid" from the play store
Then push your gantry all the way to the back, touching the Y-endstop
Then, pluck your belts like a guitar string and measure the frequency with the Spectroid app
What size is your printer? Belt frequency should be at least 50 Hz for a 500 size, 60 Hz for a 400 size, 70 Hz for a 300 size
The printer is a 400mm. Here is a screenshot of the Spectroid graph, red line represents the plucked belt. Are you able to send me a screenshot of what the spectroid graph should look like as a good example?
Hmmmm, unfortunately that seems OK
is it the same for both belts?
when you run the shaper graphs, do you hear things buzzing or rattling around?
Yes, both belts produce the same shaped spectroid graph. I have made a video running the belt tension macro, I do hear buzzing/vibrating but I kind of expect it does that too a certain degree versus rattling of a loose bolt for example. https://youtu.be/csMBvypBn_Q
That is the graph produced in the video.
I wonder if the cable attached to the accelerometer has too much freedom and it's transferring motion of the cable to the accelerometer. Can you tiewrap the blue cable to the black cable and test again?
I'll make a video of mine to compare, a moment
You have things vibrating nearly constantly, I assume you have checked all the screws etc 😄
@swdee start by executing
ACCELEROMETER_QUERY
and
MEASURE_AXES_NOISE
to verify if the sensor is faulty, your graphs shows basically nothing but signal noise.Tie that adxl-cable down to make sure that the cable is moving with the toolhead.
I have tied mine down into the green cable-gland and (I think) also tied it down on the right side of the toolhead just by the extruder or something
*striped it down
Not certain that this is it, but it's a theory which could explain it.
I cable tied the cables and bowen together but it didn't make any difference to the graph above.
I was wondering if the sensor could be bad also so brought another one, hopefully it will arrive in the next few days. Running those commands gives me the following output;
accelerometer values (x, y, z): -151594.176922, -3034.844362, 156819.748441
accelerometer values (x, y, z): -151594.176922, -2886.803174, 156819.748441
Axes noise for xy-axis accelerometer: 0.000000 (x), 87.583522 (y), 0.000000 (z)
Axes noise for xy-axis accelerometer: 0.000000 (x), 87.162055 (y), 0.000000 (z)
Ran them twice for comparison. I am not sure what the ranges should be, I did see online the noise axes should be in rang of 1-100.... hence those zero values maybe an issue?
I tried tieing down the cables but it did not make any difference unfortunately.
accelerometer values (x, y, z): -151594.176922, -3034.844362, 156819.748441 accelerometer values (x, y, z): -151594.176922, -2886.803174, 156819.748441it shows that it accelerates almost 2G in X and Z direction, it should be close to 0 on 2 axes and about 1G on on the 3rd one the accelerometer is broken @swdee ^
ok, thanks for that... i'll wait for the new one to arrive and re-test
New accelerator sensor finally turned up and this one actually works and I am now getting data which looks similar to what it should be doing.
Axes noise for xy-axis accelerometer: 20.746252 (x), 18.938507 (y), 47.780332 (z)
Axes noise for xy-axis accelerometer: 21.558181 (x), 19.457533 (y), 53.079426 (z)
accelerometer values (x, y, z): -296.082377, -9178.553681, 1685.410096
accelerometer values (x, y, z): 0.000000, -9104.533087, 1302.362347
New graphs generated by a working accelerometer above, these look more like the examples I see from others. Any comments on them? @Prav @oskar7122 @Steef @Technofrikus
The last one - belt tension is uneven. But i'm not very good at these.
Max accel < 5300 for Y with MZV at 42hz, I think you can safely tension belts a bit more.
I think @Helge Keck would say that both your vibrations on the last graph / belt tension are a bit "crazy" and that you should try to tighten stuff (screws, make sure cables etc can't shake) to decrease those vibrations > 100hz.
Try tightening the upper belt 0.5-1 turns and rerun the belt tension graph.
Also run an input_shaper calibration every full turn or something and make sure you don't go too high in MZV hz for y. I can't remember if ~40hz is "tight enough" for Y for a 400mm³ (meaning, you're already a little bit too tight) or if it's closer 45-50hz that's the "max"
Y on my 400 is tuned to 44hz mzv and gives max accel at ~5900mm/s², but I have a little bit lighter weight gantry (and probably a heavier fan - sanyo denki 9gax something)
in the last graph frequency is the most important thing, not the amplitude. the amplitude difference can be caused by one belt being a bit more streached than the other or unequal length, like a single teeth of difference. good to fix not the end of the world.