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

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

Page 4 / 53
  RSS
TimeWalker75a
(@timewalker75a)
Eminent Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Hi gents, a quick one.

Was getting a lot of artefacts on the Y axis with stock motors and the Y motor was running CONSIDERABLY hotter than the X motor. So after very little hesitation, I had swapped the Y motor with LDO-42STH40-1004ASR (which is supposed to be a Moons clone based on the winding configuration). In stealthChop (silent) mode, literally second layer into the print, I got greeted with TMC Driver Overtemp and the Y motor being scorching hot. ❗

Do I have to play with TMC2130_PWM_GRAD_Y in the firmware? Saw earlier in the thread that value of 5 seems to be adequate or the motor is just a no-go for the axis swap? Thanks for any insights!

Publié : 11/03/2019 10:03 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Assuming the motor you received as actually a LDO-42STH40-1004ASR, it should run without altering the firmware. Verify the motor actually is a TMC2130_PWM_GRAD_Y Cooler unit.

Yes, increasing TMC2130_PWM_GRAD_Y should reduce current draw in Stealthchop, but there is something odd about your printer.

I would DETACH the y-axis belt and verify that y-carriage moves without any resistance.

Also be sure that you are not over tightening the drive belt. That also increases resistance against which the motor must run.

Publié : 11/03/2019 1:39 pm
Jorge
(@jorge-2)
Active Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

This is a great example of development by the community.

I only have to say, thank you, your work is truly helping others.

😀

I have some questions, if you don't mind responding to them:

Regarding this configuration:
X Moons 0.9º MS17HA2P4100
Y Moons 0.9º MS17HA2P4100
E Moons 1.8º MS17HD2P4100

Does it limit in any way the printer? I'm referring to things like velocity, acceleration and deceleration, when compared with the stock configuration?

I can't understand the choice of changing the extruder stepper, since it has the same characteristics then the stock LDO motor. I ask this because I was wondering ok to "upgrade" the machine to an extruder with a gear relation of 1:2 or 1:3 like does from Bondtech with the stock stepper. On the same subject of gear reduction, what about make the reduction on the X Y axis movement with a gearbox? it is pretty simple (direct or with belts).

If you see any benefit I can take a time to create a Desing and provide the STL's files for that.

Best regards,
Jorge Rui

Publié : 11/03/2019 3:46 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Thanks. Happy it is helping others. The VFA elimination quest has been an adventure. I now have a box full of many motors, but we learned quite a bit.

====

Looking back at the LDO Cooler stepper's specs, Its coil resistance is a bit lower than the stock motors - about as low as the 0.9 degree Moons and OMC motors. I didn't spend a lot of time with the LDO Cooler once I found their VFA performance was worse than the Moons 1.8's.

Perhaps they DO need a firmware change to limit their Stealthchop current draw. Might also be an out of spec motor. Coil resistance should be 4.1 ohm. No idea about its inductance spec.On the other hand, the 0.9's only need TMC2130_PWM_GRAD_Y 4 or 5 to not overheat the TMC2130. Something is odd.

Publié : 11/03/2019 4:39 pm
Chicago Keri a aimé
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Does it limit in any way the printer? I'm referring to things like velocity, acceleration and deceleration, when compared with the stock configuration?

I have by no means explored the full range of acceleration and velocity with the 0.9's. I'm having the EINSY step the 0.9 at half the frequency of 1.8's. So, the step frequency is not a limit. It's the same number of addressed motor positions, but the positions are more accurately achieved with the 0.9's than the 1.8's. The limitation to velocity and jerk could conceivably happen with the 0.9's yielding lower max torque, but I have not pushed my machine to explore max velocities with either the stock motors or 0.9's.

The biggest drawback is needing to use custom software to support the 0.9 motors. That can be downloaded from my GitHub branch for 0.9 degree motors support.

https://github.com/guykuo/Prusa-Firmware/tree/0.9-Degree-Stepper-Support

Users wishing to experiment with 0.9 degree motors probably should be at an experience level of being comfortable compiling their own firmware. There is not other way to finely tune the Trinamic parameters. Whereas my firmware fork works for the Moons 0.9 MS17HA2P4100 and OMC 17HM15-0904S, you many need different Trinamic tuning for other motors.

Motors selected should be 1 amp or lower rating to remain within TMC2130/EINSY current limits for Stealthguard.

After seeing the improved print quality using 0.9 motors, I'm simply not going back to 1.8's. All three0.9 motors tested are a leap forward from even the best 1.8 (Moons). I found the Moons MS17HA2P4100 superior to the OMC (stepperonline) 17HM15-0904S. OMC 0.9's are still a considerable improvement over any 1.8's. They don't quite reach the monotonic uniformity of the 0.9 Moons, but are an option for the budget conscious who are will to forego the last 25% of VFA elimination. My choice is the Moons MS17HA2P4100.

I can't understand the choice of changing the extruder stepper, since it has the same characteristics then the stock LDO motor. I ask this because I was wondering ok to "upgrade" the machine to an extruder with a gear relation of 1:2 or 1:3 like does from Bondtech with the stock stepper. On the same subject of gear reduction, what about make the reduction on the X Y axis movement with a gearbox? it is pretty simple (direct
or with belts).

A 0.9 motor on the extruder is NOT the same characteristic as the stock 1.8 degree LDO motor or 1.8 Moons. It is a 0.9 degree motor. As such, it is better able to actually hit the EINSY requested microstep positions. The extrusion is thus more uniform, despite the same number of addressed steps. The extruder actually runs slow enough that the EINSY can step it twice as fast as stock motor - yielding double the addressable microsteps. I have support for that in my firmware branch, but have not been able to elicit much difference going to that many steps.

Yes, going to a geared setup is another way to improve monotonicity. By having the gearing, the motor has an easier time hitting difficult microstep positions and those positions are physically finer. However, we have basically proven that a 1.8 motor which is inherently less uniform simply never becomes that much more uniform even with Trinamic tuning. Physical motor step uniformity is a much larger determinant of step monotonicity than adding more micro steps.

Hence, a 0.9 stepper in the extruder position produces a more uniform extrusion - even in the absence of gearing to reduce torque load.

Publié : 12/03/2019 12:16 am
Jorge
(@jorge-2)
Active Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

in summary, you findings show that a 0,9 Moons stepper is a better solution to all axis except the Z axis.
With this you can enhance greatly print quality just by changing the steppers. that about 125€ of steppers... This is almost the price as the Bontech extruder that fits on the MK3. With the diference that we have to know how to compile the firmware...

That last step (Compiling firmware) is what's worries me about.

Publié : 12/03/2019 9:21 am
Chris
(@chris-32)
Eminent Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

I really want to incorporate your improvements, Guy. I've been waiting for Prusa to release 7x7 mesh bed leveling and fix the extruder temp blips that showed up in 3.5.1, before I update the firmware. My hope is once they published these changes to their firmware, someone can add the Moon's motor driver changes to that firmware, then I'm all in!

Publié : 12/03/2019 2:25 pm
Chris
(@chris-32)
Eminent Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

I really want to incorporate your improvements, Guy. I've been waiting for Prusa to release 7x7 mesh bed leveling and fix the extruder temp blips that showed up in 3.5.1, before I update the firmware. My hope is once they published these changes to their firmware, someone can add the Moon's motor driver changes to that firmware, then I'm all in! Thanks for all the hard work bud!

Publié : 12/03/2019 2:27 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Compiled firmwares

The MK3 head from which I forked, includes 7 x 7 mesh. It even has the new R parameter. For instance....

G80 N7 R1
Tells printer to do 7 x7 mesh with 1 repeat per point.

I got rid of my temperature blips (on Mk3) by turning the filament sensor OFF. With the filament sensor active, false filament issue episodes would make the printer produce those nasty blips.

To understand what I did in the firmware, open in Arduino, look for "kuo" with a search in all tabs. I marked all the changes with //Kuo

Publié : 12/03/2019 4:29 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
What are the odds?

Replacement for the bad MS17HA2P4100 Moons stepper arrived today from Amazon.

What are the odds?!! The replacement unit is also strongly notchy. I was suspicious when I tried turning its shaft and felt coarse, strong detents. When I tried printing with it, VFA's nearly at the severity of 1.8º motors were produced. The motor is worse than the OMC (Stepper Online) motor.

Unbelievably, I received another bum unit. Perhaps it was the way it was packaged. Amazon put it in just a bubble wrap envelope plus the Moon's product box. The Moons box is foam lined. I don't see any dings on the Moons box, but something is definitely awry with the motor.

Trying again, but this is pretty surprising. My first two units are fantastic, but I need one more for the y-axis. Meanwhile, the OMC unit goes back on the y-axis.

Publié : 13/03/2019 2:47 am
Jorge
(@jorge-2)
Active Member
Re: What are the odds?


Replacement for the bad MS17HA2P4100 Moons stepper arrived today from Amazon.

What are the odds?!! The replacement unit is also strongly notchy. I was suspicious when I tried turning its shaft and felt coarse, strong detents. When I tried printing with it, VFA's nearly at the severity of 1.8º motors were produced. The motor is worse than the OMC (Stepper Online) motor.

Unbelievably, I received another bum unit. Perhaps it was the way it was packaged. Amazon put it in just a bubble wrap envelope plus the Moon's product box. The Moons box is foam lined. I don't see any dings on the Moons box, but something is definitely awry with the motor.

Trying again, but this is pretty surprising. My first two units are fantastic, but I need one more for the y-axis. Meanwhile, the OMC unit goes back on the y-axis.

Can this explain some diferences in VFA's on stock machines? Could it be that some of the stock motors are better then others?

Publié : 13/03/2019 11:29 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Motor variation vs motor type

Can this explain some diferences in VFA's on stock machines? Could it be that some of the stock motors are better then others?

Probably a little, but the amount of variation in a normally functioning unit of any particular 1.8 I have used has been pretty small compared to the leap that occurs when switching to a (properly functioning) 0.9 degree stepper.

From what I have observed through multiple stages of physical refinements in alignment, bearings and rods, VFA appearance can be totally obscured by large scale noise and inaccuracies with sloppier mechanicals. The printer acts as though there are loci of happiness where a balance of motor and linear mechanicals match each other. A non-uniform motor matched with sloppy mechanicals can look ok - not so sharp and detailed but noise seemingly hidden via "blurring." If one improves the mechanicals to be more revealing, shortcomings of the motors become more and more apparent. If you get your machine mechanically refined enough that all the bigger noises are removed, you have the VFA's as the next biggest and visible (of the lesser remaining issues). Going to 0.9 degree motors can remove the VFA's. Then you begin dealing with the next layer of issue - 0.2 mm vertical pulley/belt artifacts.

Yes, I'm now working on those and yes. Changing the pulleys to a better match to the belt does reduce the 0.2 mm, vertical, tooth artifacts. Not all pulleys have the same tooth profile or machine finish. That will become another thread once I have more tests completed with 6 to 7 different brands of drive and idler pulleys.

Meanwhile, I just want another good Moons 0.9 stepper for my y-axis..

Publié : 13/03/2019 4:56 pm
bhawkeye
(@bhawkeye)
Eminent Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Maybe I'm slow, but what do belts and/or pulleys have to do with vertical artifacts? Seems to me like only Z & E could be involved, and neither has belts or pulleys. What am I missing?

Publié : 13/03/2019 10:07 pm
vintagepc
(@vintagepc)
Membre
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)


Maybe I'm slow, but what do belts and/or pulleys have to do with vertical artifacts? Seems to me like only Z & E could be involved, and neither has belts or pulleys. What am I missing?

Think about what's going to happen if you loosen your X/Y belt tension and your printer makes sudden x/y direction changes...

Publié : 13/03/2019 10:12 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

The fitment of belt and pulley teeth creates tiny changes in effective diameter of the pulley during rotation. If the fit isn't exactly right, those changes in diameter translate into periodic acceleration/deceleration as pulley rotates. That makes the axis vary velocity ever 2mm of travel. The changing travel velocity alters how much filament is laid down/distance. The print effect is a vertical surface wave that is spaced every 2 mm. Depth, shape, and surface finish of the pulley teeth vary between pulleys. They are not all machined the same nor are all "GT2" belts the same shape/size teeth. Prusa currently ships with Gates 2GT belts. I'm only going to work on matching pulleys to that belt. Non-Gates, GT2 belts, may need a different pulley than I find.

I have about 1/2 dozen different brands of "GT2" drive pulleys in hand now. They definitely differ in tooth profile, depth, and finish. Finding a best match to the 2GT Gates will entail testing each pulley to determine best tension and comparing the best for each. Then there are the idler pulleys. Those I don't have as many variants to test.

I have found one pulley thus far that minimized the effect. Surprisingly it is not the Gates licensed, ED3 pulley, but rather a cheap, BALITENSEN GT2 Timing Pulley 16 Teeth. https://www.amazon.com/gp/product/B07BH26P2D/ref=ppx_yo_dt_b_asin_title_o04_s00?ie=UTF8&psc=1

It's NOT a 6 mm wide pulley, but our x and y belts self center in it just fine and I am seeing the vertical, 2 mm "tooth wave" decreased by about 50% compared to the Prusa drive pulley. The extra width also avoids belt/flange effects.

Surprisingly, the very well machined, nicely flanged, ED3 drive pulley made the 2mm toot wave worse. That was unexpected because it is made with actual Gates licensing for the 2GT profile. These are preliminary results and much more testing is needed.

Tooth issues have been relegated as inconsequential, but once you have the machine tuned past eliminating VFA's, the tooth wave is left as the most prominent issue. Keep in mind that at each level, we're looking at smaller and smaller magnitude effects. If any larger issue is still at play, the lesser ones are obscured.

Publié : 13/03/2019 10:22 pm
bhawkeye
(@bhawkeye)
Eminent Member
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

Thanks for that detailed explanation. Now it makes sense to me 😀

Publié : 13/03/2019 11:26 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Grinding a flat on motor shaft

Many of the motors tested in this thread lack a ground flat or have only a partial length flat on their shaft. Consequently, drive pulleys may not engage a flat portion of the shaft when positioned to correct belt alignment. Here is a nice video showing a simple technique for grind a flat so our grub screws can lock securely into position.

I particularly like his use of a bag to protect the motor from grindings. A nice tight fit over shaft would be ideal. Dremel quickly cut the flat.

Do not take stepper motors apart like he does with the helicopter motor. Taking a stepper apart can permanently drop motor torque. Mere removal and reinsertion of the rotor can partially demagnetize components. Rest of video is applicable to our situation.

Publié : 14/03/2019 10:09 am
vintagepc
(@vintagepc)
Membre
Re: Grinding a flat on motor shaft


Taking a stepper apart can permanently drop motor torque. Mere removal and reinsertion of the rotor can partially demagnetize components.

... That's not how magnetism works. It does not just vanish into the ether like gas when you open a motor.

More likely is either improper alignment during reassembly, physical damage to the core as part of the process, etc.

Publié : 14/03/2019 2:21 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)

More likely is either improper alignment during reassembly, physical damage to the core as part of the process, etc.

You are quite certain of that?

Publié : 14/03/2019 3:47 pm
Chicago Keri a aimé
vintagepc
(@vintagepc)
Membre
Re: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)


More likely is either improper alignment during reassembly, physical damage to the core as part of the process, etc.

You are quite certain of that?

More certain than I am that the magnetism magically vanishes, unless you can show me a credible source for that claim.

Publié : 14/03/2019 3:49 pm
Page 4 / 53
Partager :