V-Core 4 400: 3 peaks in belt tension + Y axis weirdness?

[pulsar harness]

I've recently finished building my V-Core 4 400mm and I've reached the calibration stage of the commissioning guide.

I'm trying to understand what I've got going here, the X axis input shaper graph seems fine but the Y looks suspicious (low acceleration and everything between 75-150hz) but I don't know what to look for.

The printer is on the floor.

What I've tried so far:

Tightened X linear rail.
Took belts out, checked for X/Y binding, seems fine.
Cut the belts to the exact length.
Rerouted the belts so that both belts are tightened to the same tooth.
Made sure both belt tensioners were the same length.
Frequency matched them to 87hz.
Checked gantry squareness on the rear XY joiners.
More zip ties on printhead.
Tightened idlers on X gantry.
Made sure belt is centered on drive gears.

I've added the belt tension, input shaper and real time analysis graphs.

The real time analysis graph macro is the one from the ratos guide (https://os.ratrig.com/docs/guides/real-time-analysis/#custom-power-spectral-density-graph-macros).

I'm hoping someone can point me in the right direction 🙂

[ornate acorn]

Your idlers can be too tight, you should go around and roll all of them without the belt to check for friction. Also be sure that the belt is centered on the drive gears and not rubbing.

[pulsar harness]

I will remove the belt again and confirm all idlers are running smoothly, forgot to mention but I shall edit my main post that when I rerouted the belts I made sure they were centered on the drive gears.

[pulsar harness]

So I removed the belts and checked al idlers for smooth operation, 1 idler on the lower belt X gantry did not feel very smooth so I loosened the nut and bolt holding the idler slightly and now it feels smooth again.

Will have to do some new belt testing and input shaper to see if anything improved.

[pulsar harness]

I don't think much changed, these are the new graphs. Y looks worse but X improved?

[ornate acorn]

How'd you do the tensioning on the belts?

[pulsar harness]

Same method as before, made sure the belt is attached to the same tooth on the back of the print head and then used the app they recommend to 87hz and measured the tensioners with calipers.

Both tensioners measure 45mm and the X gantry also hits the XY joiners in the back at the same time.

One thing I haven't mentioned is that I still have roughly 60mm of belt sticking out of each side on the back of the print head, I don't want to cut that off yet until I've figured out how to get the belts to match on the belt tension graph.

[ornate acorn]

I'd suggest cutting it, if your belts are as you described, you're fine. You can leave a few teeth tho, but 60mm will be enough to throw off your graphs

Also, I just want to be clear... IS charts are not the end all be all

Mine look like trash, and it prints quite honestly beautifully.

I seriously recommend, if you are confident in your tensions which I would be in your case, enter in the values and kick out some prints.

[pulsar harness]

I did a re-test where I stuck the excess belt under some other wires and did not notice any difference in the Y input shaper graph.

I did notice that the right X gantry carrier was slightly rattling, so I could disassemble and reassemble that…

Another thing that I possibly could check is to remove the Orbiter Smart Sensor, I remember that there is a metal ball in there and it might be that. I need to inspect that anyway because it falsely triggers filament runout…

[pulsar harness]

If I decide to continue with these results, which shaper do I select? The documentation just tells you to pick MZV.

[ornate acorn]

MZV always, idk why but MZV. The others have drawbacks that MZV doesn't really.

[pulsar harness]

So I loosened the bolts on the right XY joiner and pushed the whole assembly more towards the center of the frame and X gantry that were holding idlers, so for the lower belt, and tightened them only a little bit, it seems to give me slightly better results now...

[ornate acorn]

It does look better! Like I said, you will almost certainly start printing and have little to no issues that are solved by "fixing" these IS graphs. Skew is the biggest one, once that is taken care of, let it rip

[knotty hawk]

Got same problem with 75-135 hz did you fixed this ?

Almost exactly same graph

[knotty hawk]

Measured with toolboard T0 (ebb42) better result.