No matter what I do I can not get rid of the third hump

I have never had this issue before Everything I have read suggests this is a belt path issue. The belts are all runing more or less in the centre of the bearing stacks The belts are running in the centre of the motor pulleys The belts are not rubbing on anything The gantry is square and moving very smoothly Both sides of the gantry hit the front stops at the same time. Altering the belt tensions seems to never have an effect on the third hump...it is always there and only in the upper belt and never in the lower one. I have swapped out nearly all the bearing stacks (only found 2 that sounded like they might have had a bit of wear on them.) The belt lengths are the same to the exact tooth count. This might make me take up drinking. @8Complex
No description
55 Replies
8Complex
8Complex3mo ago
how did it get that bad? i thought it was relatively normal other than a minor second hump]\
MFBS
MFBSOP3mo ago
Nope at all stages I have had that third hump
8Complex
8Complex3mo ago
I'd almost consider removing one toolhead item at a time and retrying until the curve changes... like CPAP hose, fan, etc
MFBS
MFBSOP3mo ago
I have removed the CPAP hose and the fan. No change. Will remove the Orbiter now.
8Complex
8Complex3mo ago
Maybe an odd suggestion, but put a small (3mm) spacer between your endstops and toolhead to try a different position?
MFBS
MFBSOP3mo ago
Worth a shot Running that test now OH that changed things a lot and I have not had that before Now I have three humps in both belts!
MFBS
MFBSOP3mo ago
No description
8Complex
8Complex3mo ago
Ok, now that is strange
MFBS
MFBSOP3mo ago
Belts are exact same length to the tooth count...but extruder which is heaviest item on the print head is offset... Just doing another test with the extruder off.
MFBS
MFBSOP3mo ago
With Orbiter extruder off the tool head At least the correlation between the two curves is good (best I have seen yet on this printer) Frequency of humps is higher than before but that is to be expected given the lighter weight of the tool head with the Obriter extruder removed.
No description
8Complex
8Complex3mo ago
The amplitude is higher than it had been... which makes sense, but I'd guess that whatever it is primarily is still there
MFBS
MFBSOP3mo ago
Yep exactly. I think next step is put the Orbiter back on and take the hot end off.
8Complex
8Complex3mo ago
Stupid question, but have you nut & bolted everything?
MFBS
MFBSOP3mo ago
Yep I should mention I have the Microlight print head But same one I have been running for over a year now
8Complex
8Complex3mo ago
Any chance something cracked during dis/reassembly? It's likely been through a lot of heat cycles
MFBS
MFBSOP3mo ago
and crashes
MFBS
MFBSOP3mo ago
Oh I just did something silly while I was trying to think of the next test. I mounted the orbiter on backwards....with the weight of its motor hanging over the front now.
No description
8Complex
8Complex3mo ago
Hmmm... preloaded the carriage roll to the front maybe/
MFBS
MFBSOP3mo ago
Yes
8Complex
8Complex3mo ago
Maybe needs a grease repack in there or something... some bearings 'loose' bumping around in there?
MFBS
MFBSOP3mo ago
Its about a six month old Hiwin Z1 contact linear rail and carriage. Not sure if is genuine Hiwin but definitely has less play than the old one I swapped out. That test was also in the offset postion you suggested so running it again with print head in normal centred position. Result only changed very slightly in the non offset positon.
8Complex
8Complex3mo ago
So probably not the motion system then Did you have issues before the Toro tube? Maybe there is a way to change it's resonance on purpose... like putting a clamp on it, or another fastener in the midspan Hell, maybe even just some tape
MFBS
MFBSOP3mo ago
Yeah I like your line of thinking. There have been a lot of changes to the printer recently New hybrid joiners New long shaft core XY motors with bearings in the top motor plates Toro tube added....Need to play with the mounting of that.
8Complex
8Complex3mo ago
Well, as long as you take the methodical approach, you'll narrow in on it eventually Anyway, time for me to turn in for the night... I'll check back tomorrow and see if anything has turned up... good luck!
MFBS
MFBSOP3mo ago
Hey many thanks you have me headed down a fresh process, I was running out of ideas. Much appreciate and good night.
MFBS
MFBSOP3mo ago
Linear rail screws on Toro tube tightened. Better correlation of two lower peaks but third hump bigger
No description
MFBS
MFBSOP3mo ago
Pre clamped linear rail on to Toro tube with G clamps and then further tighted screws holding linear rail on...
No description
MFBS
MFBSOP3mo ago
And the same as above but with orbiter again reverse mounted. Again gets rid of the third hump
No description
MFBS
MFBSOP3mo ago
Here's my thinking at this point. The more tightly the print head is coupled to the X gantry the greater the accelerations that are measured by the ADXL unit on the print head....there for the vibrations are not coming form the tool head but being transmitted to it by the movement of the gantry? By mounting the Orbiter in reverse on the print head so the motor hang over the front of the print head I am preloading the carriage roll to the front maybe and this is damping the movements measurec by the ADXL unit on the print head?
DocDoyle
DocDoyle3mo ago
do your X/Y graphs contain the third bump? ive seen three bumps on belt/Y graph with a wobbly X rail, where it would not resonate in X/Y but in a rotational component of the X axis, with a small peak of Z movement with the ADXL being very close to the carriage, its Z movement was measured to be very low (at a small distance from the carriage), but the nozzle tip was moving significantly, even causing ringing in the Z axis
8Complex
8Complex3mo ago
Seeing as how different the bottom two were, the loading of either the linear carriage or the pre-twisting of the gantry tube seem to be affecting it quite a bit. That last chart's minimized third hump really makes me think there is some type of twisting action going on there. Have you run the other type of graph? I'm wondering if narrowing it down per the accelerometer axis would help
MFBS
MFBSOP3mo ago
I am using the Beacon Probe which is mounted on the bottom/plate duct of the Microlight...so yep some distance from the rail.
MFBS
MFBSOP3mo ago
Quit a bit oy Y acceleration shows up in the X direction.
No description
MFBS
MFBSOP3mo ago
No description
8Complex
8Complex3mo ago
Yup, that Y direction has me suspicious of your carriage play, although I suppose it could always be the rigidity of the lower half of your toolhead
MFBS
MFBSOP3mo ago
Seems strange that X accelerations would cause movement in the Y direction? I can see it causing acceleratins in the Z moovement if the carriage could rock along its long X axis but movement in the Y axis is harder to imagine....but something very odd is going on.
8Complex
8Complex3mo ago
Not really... you have asymmetry along the Y axis which causes motions going across that direction to sway, which imposes forces in the perpendicular direction Lever arms don't move straight, they follow curves, so their motion is never purely along a single axis
8Complex
8Complex3mo ago
Total crapshoot, but for instance if you move this axis in X, you have the extruder and hotend attached from a cantilevering point behind them... so you'll may get some twist there that could translate into the cantilevered portion bobbing, which would also translate into Y resonance. Of course, if this was the case, you'd likely see a lot more on Z, though...
No description
8Complex
8Complex3mo ago
Just a rough example of what else could be imparting twist/off-axis resonances
MFBS
MFBSOP3mo ago
All good thinking sir. I am going to install another accelerometer on the gantry and one on the top plate of the extruder and do some further testing.
MFBS
MFBSOP3mo ago
@8Complex With my Beacon ADXL probe attached to the XY joiner. So far proving your theory.
No description
8Complex
8Complex3mo ago
Yeah, it looks nice and solid there for the most part
MFBS
MFBSOP3mo ago
Of course not even with any belt tuning
MFBS
MFBSOP3mo ago
Now here is the graph with the Beacon ADXL mounted firmly to the X linear rail next to the print head.
No description
DocDoyle
DocDoyle3mo ago
can you hear what is vibrating on that second peak?
OldKingHamlet
OldKingHamlet3mo ago
This... is almost the exact issue I'm troubleshooting through. Went from perfect/clean two humps to chaos. Granted, I had the frame get strongly hit during a workplace shuffle, replaced the toolhead with the RRTH, and installed a light weight gantry, so there were a lot of changes. I'm actually thinking the rails that were on this printer (I bought this RR used, and it was originally an early V-Core 3) had given up the ghost, so I've got a set of new y and x rails on the way. I had spotted gantry twists being reported in the beacon calibrate process, but only along the back (0.05mm), which incidentally aligns with the umbilical's spring wire pushing the head to rotate forward/over the gantry, and towards the plate. I'm also getting a weird metallic buzz during belt tension testing that comes from the toolhead, but is completely unexplained by extruder gears, loose pieces, or anything else that could rattle. I'm now actually thinking the bearings in the x carriage, and since there's some other evidence of a strong gantry collision at one point. If the cowling for the bearings that keep them in place was damaged, that could explain it some of it, especially now that the much lighter RRTH is installed.
8Complex
8Complex3mo ago
Looks to me like something angularly deflecting in one of the directions (along the axis of travel with the orange belt above)... which would explain the combination X & Y deflections at the same time in the previous graphs, too
mazas
mazas3mo ago
I also had a bit of a similar issue when I tried to run my 300 with one of the lightweight x-tubes. I filmed the head with a high-speed special purpose camera (aka iPhone) in slow motion, and you could clearly see the rotational twist around the x-gantry. Dunno if it would work here, but a slo-mo could help in diagnosis?
8Complex
8Complex3mo ago
I think so, too, but it'd likely have to be quick enough that you won't get rolling shutter... not sure if there are any phones without rolling shutter, but I suppose if the problem was seen in both directions, you'd know it wasn't just a visual artifact
mazas
mazas3mo ago
Yes, that is a good point. Because of the rolling shutter you should not rely on the lines/edges in any one single frame to be straight. However, you can still compare frame-to-frame, and a change in one line should reflect what is happening in real life. Of course the fps needs to be above Nyqvist, but with 240 fps you should be able to check up to 100Hz or so. You can also detect movement above Nyqvist, but you’ll need take the aliasing into account there
OldKingHamlet
OldKingHamlet3mo ago
So, I fixed mine. The rail carriages that were on my vc3, particularly the x rail, were failing. I replaced the rails with some new ones that had some preload on the carriages. I'm assuming the prior owner of mine had a few crashes and/or other z-axis sins. I just replaced the rails and carriages in one go, and now my resonances are clean af *Edit: Mostly. I decided to "fix" some other things and now my lower belt is binding due to some path alignment issues. But at least the gigantic wall of resonance noise is gone.
MFBS
MFBSOP3mo ago
I am getting input shaper recommendations for Y over 11,000 now 0.0% MZV, which I can live with at the moment.
Jamorro
Jamorro3mo ago
Hybrid?
MFBS
MFBSOP3mo ago
@Jamorro (vc4 300) no.
Jamorro
Jamorro3mo ago
11k on Y with no hybrid? thats crazy im at like 6.8k on a 300
Want results from more Discord servers?
Add your server