Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
 
Notifications
Clear all

Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)  

Stránka 15 / 53
  RSS
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

So far so good, 1500 seems to be working after a few selftest.  I'm about to start a quick test print with a high bed temp (which is where things fell apart previously).

The first part of the selftest still hits pretty hard on the left side X-axis.

... and just as I was typing this reply that last selftest finished up and didn't detect it's last X home move to the left side and reset the printer.  Going to increase X to 1750.  

This post was modified před 5 years by AnatomicFlack
Napsal : 20/06/2019 4:28 pm
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

https://imgur.com/a/kwDykNT (video has sound)

I'm now battling a new (homing related) behavior.  Just trying to run an autohome the X hits once, then on second hit it skips steps until the machine resets.  Does this mean it's correctly triggering the threshold on the first hit, but never seeing it on the second hit?

I've tried adjusting homing feedrate up and down (from 1000-3000), belt tension up and down, and I tried to adjust the homing threshold (2, 4) but no change.  

A selftest has similar results, it will complete the first steps on X, Y and Z hitting the limits, but then on homing of X it will hit the left stop fine multiple times, then what might be the last hit it hangs and the stepper skips.  

I'm not sure what to try next.  

This post was modified před 5 years by AnatomicFlack
Napsal : 20/06/2019 5:38 pm
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

My previous post with a video showing what was happening with the X failing to autohome is "awaiting moderation" 

In the meantime I seem to have found the sweet spot and at least selftest and autohome is working without failure at the moment.  

X: 263
Y: 267

I was able to run the selftest completely and I've run autohome several times and it's completing without issue.

I ended up with a firmware Y homing feedrate of 1800 and X at the default of 2000.

Time to try a print... *fingers crossed*

Napsal : 20/06/2019 6:21 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

I wish I had a reproducible gauge for the tensions set on my belts. I think it is about  7 or 8 lbs tension. Yes, there is a tension level that seems to be needed to get good homing. 

I hold the allen key at angled end so it acts as a stick to amplify my feel for changes in tension level. Hold it parallel to bed plane. Push down on mid belt and feel how its tension changes as you go down. Initial tension felt is light, but it stiffens noticeably when you reach begin to pass from taking up slack to stretching the belt. It's easiest to learn that feel when belt is intentionally loose. Adjust tension so the change is when 2-3mm deflection. Then go another 1/3 turn tighter on the bear x-idler tension screws. The ability to fine tune tension on the Bear x-axis is wonderful.

Napsal : 20/06/2019 6:27 pm
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

I've always been pretty decent at tuning my tension by ear and feel.  And I usually hit about (within +-5) these same values on first re-tensioning.  The tension range to hit the threshold reliably seems really narrow. 

Second selftest run results were X: 264, Y: 265 (no change by me to tension)

This also makes me wonder if the reason yesterday I was initially able to pass the selftest, but was getting the reset on Y homing was because the 80 degree bed heat may have pushed the Y belt tension just a bit out of range.  

Now that the printer is cooperating for the moment I can see the movement that the X was failing on in the video.  On the last endstop hit (17 on selftest, 2 on autohome) the carriage does a double endstop bump sort of movement.  I don't think I'd ever noticed it before on the printer, but was I really looking this closely at the movement, probably not.  It would seem that small second bump maybe wasn't enough to trigger the threshold at less than ideal tension.  I thought you maybe mentioned something similar about the hit distance in the early part of this thread, but perhaps that was the reset distance it uses between hits, not the distance on the final little hit?  

Napsal : 20/06/2019 6:38 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Oh, I totally forgot because it was so long ago.

The trinamics autotune themselves upon power up and 1st motor motion. That can introduce a variable that conf0unds testing. The TMC2130's remember that tuning data until the next power down. If you full power down AND get good homing behavior after power up and NO further adjustments, they will repeat that self cal in and end up in the same state. That makes system stable.

On the other hand, if the TMC2130's did their self-calibration prior to a big belt change, the next time they self cal, they can end with different internal settings because they are seeing different motion resistance. That throws things off. Depending on the whether it learns its self cal in a good state or bad, it's a behavior that can make the machine weirdly perplexing or oddly able to "adapt" keep functioning correctly once things are set up correctly.

PSA - For those unfamiliar with stepper motors an drivers. Please do NOT plug/unplug motors with printer turned on. Turn printer OFF before plugging or unplugging motors.

This post was modified před 5 years 2 times by Bunny Science
Napsal : 20/06/2019 6:49 pm
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

How interesting and infuriating all at the same time.  😉

So maybe step one to troubleshoot odd stepper behavior when belt tuning (or just general setup)... make sure you power down and back on before making any radical changes? Especially if the machine was working fine previously? 

Oddly enough, just before things started cooperating (more or less) again I had powered down because I had to manually lower the Z axis downward since the autohome kept failing which keeps raising Z on each attempt.  

When does the TMC perform the self-calibration?  Only on first motion after power up, or during any other times triggered by firmware?

This post was modified před 5 years by AnatomicFlack
Napsal : 20/06/2019 7:11 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Best as I can tell from the Trinamics documentation is it is only upon 1st motor motion after power up. I don't think our firmware can trigger it - just a power off/on. 

Napsal : 20/06/2019 7:18 pm
nikolaistolstoy@gmail.com
(@nikolaistolstoygmail-com)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Posted by: AnatomicFlack

https://imgur.com/a/kwDykNT (video has sound)

I'm now battling a new (homing related) behavior.  Just trying to run an autohome the X hits once, then on second hit it skips steps until the machine resets.  Does this mean it's correctly triggering the threshold on the first hit, but never seeing it on the second hit?

I've tried adjusting homing feedrate up and down (from 1000-3000), belt tension up and down, and I tried to adjust the homing threshold (2, 4) but no change.  

A selftest has similar results, it will complete the first steps on X, Y and Z hitting the limits, but then on homing of X it will hit the left stop fine multiple times, then what might be the last hit it hangs and the stepper skips.  

I'm not sure what to try next.  

I am still having homing issues but on X. I did reseat the cable when the carriage was at max left and it helped some but not totally consistent. In terms of temporary solutions I have found a couple things:

The position of the X carriage in X and Z can affect the homing, so I found a good reproducible position to start from (X232 Z25.5) and edited the end gcode to send the carriage there after a print. This was able to reduce my frustration as it works 70% of the time now.

Giving the cable a slight push to the right to reduce the spring force pushing the carriage away from the end stop can also help, but it’s a bit finicky and you can add too much pressure.

My next stop will be to turn on the homing tweaking gcodes in the firmware so hopefully I can get 99% reliable performance.

The behavior for me is that it hits the end stop once and then goes back and hits it again and starts making noise and stalling with restart. I guess this makes sense as homing works better when the carriage is far from the end stop, but the second hit is from very close. Based on my description is there a setting I should target or just try the speed first?

Napsal : 20/06/2019 8:55 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

position of the X carriage in X and Z can affect the homing

along with a push helps is consistent with how my x-axis misbehaved until I got the belt tension correct. 

Also absolutely verify that the only contact at the left end is the plastic vs plastic. No bit of cable going around the corner on the back of the extruder should be making contact. Sometimes a cable pokes out a little from the guide fin and blunts the homing impact.

Power the machine off. Set belt tension like I described above. Turn on and see how it behaves. If it misbehaves, turn off again, slightly increase belt tension and try again. It really should be 100% of the time homing successfully once set .

Napsal : 20/06/2019 11:10 pm
CybrSage se líbí
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

I'm running through printing some XYZ cubes.  I've reached the conclusion that the issue may be my RAMBo and my TMCs.  When running in Stealth mode I'm getting layer shifts on Y really easily.  This is an issue I was previously having with X before the swap.  Printing in Normal mode would mostly eliminate those issues.  However, my printer has always been SUPER sensitive to layer shifts from nozzle hits.  

I've been printing a lot with CF filled polycarbonate and PETG and I kept getting built up on the nozzle, which was enough to cause the slightest of impacts.  They do not register to the machine, and I have had crash detection disabled forever since it was so unreliable for me. 

I tried my first XYZ cubes in the CF-PETG (which was already loaded), and the first 3 had layers shifts in almost the exact same height.  I then printed the same gcode in Normal and it finished fine, but with some obvious nozzle build up and redeposit on other parts of the print. 

Right now I'm running PLA using the standard PRUSA profile.  The normal print finished perfectly, the stealth is running now.   

I can also avoid the shifts in stealth if I'm using a full pro silicon sock with an E3D nozzle that it fits properly.  I've been using my Slice Engineering and Midwest Tungsten nozzles with my filled filaments (neither of which can use the full sock and the big hole sock is useless to help keep the nozzle clean) so it becomes a boogery mess during prints.  

I'll post pics of the cubes when done.  

Next step will be swapping out the pulleys to the Baletensens.  

All of my issues may have just been related to trying to do everything in stealth mode and my board being hyper sensitive.  I've already ordered a new RAMBo from PRUSA.  Worst case is this finally gives me the last remaining piece to put together a whole new printer from my old spares (even if it can only print in normal mode).  

Napsal : 21/06/2019 6:32 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Layer shifts happening more often with hotter materials and less likely with sock also suggests another problem.

Voltage sags causing the EINSY and TMC2130s to get upset.

Have you checked your power supply and its leads/connectors? What voltage is your power supply adjusted to output? What voltages does the EINSY see?

I hope you are not running on a Prusa supply. 300 watt Mean Wells are dirt cheap and give more head room. 

Napsal : 21/06/2019 6:43 pm
nikolaistolstoy@gmail.com
(@nikolaistolstoygmail-com)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Posted by: AnatomicFlack

I'm running through printing some XYZ cubes.  I've reached the conclusion that the issue may be my RAMBo and my TMCs.  When running in Stealth mode I'm getting layer shifts on Y really easily.  This is an issue I was previously having with X before the swap.  Printing in Normal mode would mostly eliminate those issues.  However, my printer has always been SUPER sensitive to layer shifts from nozzle hits.  

I've been printing a lot with CF filled polycarbonate and PETG and I kept getting built up on the nozzle, which was enough to cause the slightest of impacts.  They do not register to the machine, and I have had crash detection disabled forever since it was so unreliable for me. 

I tried my first XYZ cubes in the CF-PETG (which was already loaded), and the first 3 had layers shifts in almost the exact same height.  I then printed the same gcode in Normal and it finished fine, but with some obvious nozzle build up and redeposit on other parts of the print. 

Right now I'm running PLA using the standard PRUSA profile.  The normal print finished perfectly, the stealth is running now.   

I can also avoid the shifts in stealth if I'm using a full pro silicon sock with an E3D nozzle that it fits properly.  I've been using my Slice Engineering and Midwest Tungsten nozzles with my filled filaments (neither of which can use the full sock and the big hole sock is useless to help keep the nozzle clean) so it becomes a boogery mess during prints.  

I'll post pics of the cubes when done.  

Next step will be swapping out the pulleys to the Baletensens.  

All of my issues may have just been related to trying to do everything in stealth mode and my board being hyper sensitive.  I've already ordered a new RAMBo from PRUSA.  Worst case is this finally gives me the last remaining piece to put together a whole new printer from my old spares (even if it can only print in normal mode).  

I was just going to bring up the topic of layer shifts you kind of read my mind. I don’t have enough of a statistical sample yet but it seems to me like I am having more layer shifts in X now (the one acting up in homing for me). 

Im just trying to think how the .9 motors could be causing this. I guess it’s related to the homing issue, but in that case it seems to be it is not sensitive enough whereas for layer shifts one would think that it is too sensitive and thus “giving up”  when it slightly touches something. None of the layer shifts I have had showed big buildups that the motor could not overcome, so I am thinking it makes a slight hit and decides that was a stall. But that seems at odds with the homing results.

Napsal : 21/06/2019 11:11 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

nikolaistolstoy are you printing in normal or stealth mode? Also if in normal is crash detection on?

Trying to understand whether your layer shifts are simply missed steps or a detected crash followed by a failed rehoming/repositioning.

My machine is always in stealth and (obviously) crash detection is off.

BTW, the system switches temporarily out of stealth during homing. Otherwise homing would never work. 

 

Napsal : 22/06/2019 12:52 am
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Posted by: guy.k2

Layer shifts happening more often with hotter materials and less likely with sock also suggests another problem.

Voltage sags causing the EINSY and TMC2130s to get upset.

Have you checked your power supply and its leads/connectors? What voltage is your power supply adjusted to output? What voltages does the EINSY see?

I hope you are not running on a Prusa supply. 300 watt Mean Wells are dirt cheap and give more head room. 

Man... it's been one heck of an day.  

I switched from the Prusa to the Balitensens pulleys which turned into a parade of crap.  

Once I got them on, I had to crank my belt tensions down MUCH more than with the prusa pulleys.  This was in running the machine in normal, but I was able to get a selftest to complete with belt tensions about ~250 on each belt.  Then I decided to try in stealth... I kept getting the issue I made a video of the other day on X.  The motor would fail on it's last double-bounce and reset the machine.  I kept tightening the X until I was well beyond where I am normally on the tensioner by at least 2mm deeper in the x-end.  I never passed a selftest before the X belt slipped a tooth in the BNB Universal Carriage.  I'm guessing that version you made, after the one I printed, with the 0.2 belt path tightening was for a very good reason. 😉 So I think part of my issue was the belt shifting in the carriage just enough on X hits to upset the machine.  It still needs more tension than the prusas or printing in normal mode, but it's within reason now. 

Thankfully, because you built such an awesome extruder and carriage, it only took me about 10 minutes to pull apart the bearing holders reseat the belt and insert some card stock on the flat side to ensure it doesn't slip until I can print the new carriage with the tighter belt path. 

So after doing all that, I was able to pass a stealth mode selftest with the Balitensens installed. The belt tension still needed to be a touch higher, but the hear/feel test is more in line with what I was expecting from my past experience.  I think I wound up about X:255, Y:265.  

Running an XYZ in stealth mode now to see how it compares to the two I ran on the prusa pulleys.  I could see the VFA from the pulleys on the Y side, but the X was actually pretty clean.  I think I just got lucky with the pulley depth so that it was never touching the sides of the pulley. But the blocks were already near flawless and it took a good hard look to spot the very tiny VFAs.  

I'm going to try to attach the photo I took today of the 3 layer shifter CF-PETG blocks, the one success with some blobbing, and the two PLAs I ran, one in normal, one in stealth that I could see the VFAs on. 

I'm getting closer on this... I think... I hope! 🙂

 

The "Add Media" shunk the photo so much it's useless:  

This post was modified před 5 years by AnatomicFlack
Napsal : 22/06/2019 2:08 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

I am a bit concerned by how tight you describe you tension up the belts. I completely ignore the support menu numbers because they are very distantly related to belt tension. I don't understand how changing the drive pulley would require cranking up belt tension. Just get the tension 6-7 pounds. Super tight and you're into bearing wear and increased friction territory (which would also create more layer shifts)

Napsal : 22/06/2019 2:13 am
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Posted by: guy.k2

Layer shifts happening more often with hotter materials and less likely with sock also suggests another problem.

Voltage sags causing the EINSY and TMC2130s to get upset.

Have you checked your power supply and its leads/connectors? What voltage is your power supply adjusted to output? What voltages does the EINSY see?

I hope you are not running on a Prusa supply. 300 watt Mean Wells are dirt cheap and give more head room. 

Yes, I do think something else is up with my machine, we'll know in a few days when the new RAMBo shows up.  

Yes.  I swapped out my PSU for a Mean Well LRS-350 a while ago.  The clicking on the stock one from Prusa drove me mad.  I also added a Noctua to replace the Mean Well fan on the outside of the PSU in a pull config.  It lives nice and cool now and the fan spins up and down during even the hottest prints.    

Napsal : 22/06/2019 2:15 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

The stock and Meanwell supplies have a voltage adjust pot inside them. Did you tweak it to be 24 volts when installing it? My Meanwell was about 1/2 volt under when it arrived. BTW, don't over volt by much or you'll will feed more than 5 volts into the EINSY CPU.

Napsal : 22/06/2019 2:19 am
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Posted by: guy.k2

I am a bit concerned by how tight you describe you tension up the belts. I completely ignore the support menu numbers because they are very distantly related to belt tension. I don't understand how changing the drive pulley would require cranking up belt tension. Just get the tension 6-7 pounds. Super tight and you're into bearing wear and increased friction territory (which would also create more layer shifts)

As I had mentioned in the Thingiverse Messages... I don't think these Balitensens are as nice as the ones you have.  

I was trying to get a measurement from both to get an idea, but I don't know how to accurately measure a gear profile.  

Balitensen: 
Shortest measurement of across gear profile: .940-.945
Spinning gear to catch the tallest teeth measurement: .962

Prusa: 
Shortest measurement of across gear profile: .951-.953
Spinning gear to catch the tallest teeth measurement: .969-.970

(measured two of each gear with digital calipers)

I don't think I have the tools needed to measure anything else of substance.  

Napsal : 22/06/2019 2:27 am
AnatomicFlack
(@anatomicflack)
Eminent Member
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Posted by: guy.k2

The stock and Meanwell supplies have a voltage adjust pot inside them. Did you tweak it to be 24 volts when installing it? My Meanwell was about 1/2 volt under when it arrived. BTW, don't over volt by much or you'll will feed more than 5 volts into the EINSY CPU.

I didn't have to adjust it, it's was dead on 24V from the factory.  I always see 24V on the Power and 24V (occasionally 23.9V) on the Bed.  I also replaced all my power wires a while back for 16ga silicone wire from PSU to EINSY, and 18ga silicone wire from the EINSY to the bed.  

This post was modified před 5 years by AnatomicFlack
Napsal : 22/06/2019 2:30 am
Stránka 15 / 53
Share: