Notifications
Clear all

Bolt-On Gearbox Extruder Experiment  

Page 2 / 3
  RSS
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

If you use a 1.8 degree motor, setting e-steps should be sufficient.

Going to a 0.9 degree motor, a firmware change is needed to bring down msteps rate so EINSY can handle step rate.

Publié : 07/04/2019 4:10 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

The Bunny Science gearbox improves extruder uniformity compared to the stock motor.
It even slightly beats my best 0.9 degree motor.
This experiment has worked out so well that it is my current extruder motor setup.

Now published on Thingiverse https://www.thingiverse.com/thing:3550304

You got to follow its development here on the forum.

Publié : 08/04/2019 8:44 am
K-Lab
(@k-lab)
Active Member
Re: Bolt-On Gearbox Extruder Experiment

I wonder why the 17HS08-1004S isn't working for you (too low torque) but the 17HS10-0704S works?
Both have 13 Ncm and the 17HS08-1004S is the one which is recommended for the Skelestruder (listet here: https://www.ifixit.com/Guide/Skelestruder+for+Prusa+MK3+Intro/117967 )
The Skelestruder recommended 0.9° motor has 18 Ncm which makes me wonder if a 0.9° pancake motor with reduced microsteps is the better choice.
Have you tried a 0.9° pancake motor with your gearbox yet?

Publié : 10/04/2019 4:49 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

Yes, I tried two 0.9 pancakes (with reduced micro steps) and they didn't quite work. Extrusion was amazingingly uniform, but layer restarts were flawed with 2-3 mm of non-extrusion. Even tried with higher current settings without effect. I guess one could adjust layer start/stop parameters in slicer, but you should not need to do that if the motor is adequate.

Full size 0.9's work fine. Full size 1.8's work fine.

The 0.9 small bodies tested and failed with layer restart issue....

- Stepperonline 17HM08-1204S (0.9 degree, 1.2 amp, 11 N cm)
- LDO-42STH25-1404MAC (0.9 degree, 1.4 amp, 18 N cm)

To further test if the layer start problem was a 1.8 vs 0.9 degree issue, I also tried an underpowered 1.8. It too exhibited the layer start problem. It really looks like a torque issue rather than 1.8 vs 0.9.
- Stepperonline 17HS08-1004S (1.8 degree, 1 amp, 13 N cm)

One reduce size unit has been completely perfect. It is the same used on the Bondtech extruder upgrade.
Don't get this one confused with the 1 amp unit above. This has same N cm at lower amps.
- Stepperonline 17HS10-0704S (1.8 degree, 0.7 amp, 13 N cm) <--- very uniform and perfect layer restarts with Bunny Science gearbox.

Publié : 10/04/2019 6:33 pm
Spacemarine
(@spacemarine)
Estimable Member
Re: Bolt-On Gearbox Extruder Experiment


Extrusion was amazingingly uniform, but layer restarts were flawed with 2-3 mm of non-extrusion.

Do you really only see this only on the layer restart?

I'm running the LDO-pancake 0.9° on the bondtech extruder (as you remember from the other thread) and I'm having a lot of issues with some prints, especially with PETG being mechanically very weak on detailed prints. PLA prints fine most of the time.

After your post I looked at my prints again, and it looks like I also have spots of non-extrusion. But I have them several times on one plane, so I guess this phenomenon occurs on a lot of the unretractions, not only layer restarts (layer changes).

I might have to try going back from the 0,9° LDO pancake to the 1.8° 17HS-10-0704S that was supplied with my bondtech.

Publié : 10/04/2019 9:12 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

Do you really only see this only on the layer restart?
I only tested with my ecore tower and didn't do other prints once I saw the problem. The tower has only one retraction/layer.

I'm running the LDO-pancake 0.9° on the bondtech extruder (as you remember from the other thread) and I'm having a lot of issues with some prints, especially with PETG being mechanically very weak on detailed prints. PLA prints fine most of the time.

I was wondering HOW on earth that motor was working on your Bondtech.

BTW, don't forget to change my firmware to turn OFF 0.9 degree on E axis if you swap back. Should be commented lie below to turn of 0.9 degree support


//#define E_AXIS_MOTOR_09 //kuo exper

Publié : 11/04/2019 12:21 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

The Bunny Science Gearbox can dramatically improved extruder motor uniformity.

Final gearbox is now moved to https://www.thingiverse.com/thing:3553286

Publié : 11/04/2019 2:41 am
JLTX
 JLTX
(@jltx)
Reputable Member
Re: Bolt-On Gearbox Extruder Experiment


The Bunny Science Gearbox can dramatically improved extruder motor uniformity.

Final gearbox is now moved to https://www.thingiverse.com/thing:3553286

Also check out skelestruder for this and more. 😀

Regarding belt vs gearing, I did build a version with printed double helical gears and even though it had some backlash, the extrusion was still much better than stock. Overall I find the belt solution better and more robust. But I’d say the gear ratio is the biggest win if just looking at print quality.

Publié : 11/04/2019 12:40 pm
Spacemarine
(@spacemarine)
Estimable Member
Re: Bolt-On Gearbox Extruder Experiment


Do you really only see this only on the layer restart?
I only tested with my ecore tower and didn't do other prints once I saw the problem. The tower has only one retraction/layer.

Ok, this strengthens my belief that is a retraction/unretraction problem.


I was wondering HOW on earth that motor was working on your Bondtech.

Very funny indeed. Did you try to increase motor current to compensate for the lack of torque?


BTW, don't forget to change my firmware to turn OFF 0.9 degree on E axis if you swap back. Should be commented lie below to turn of 0.9 degree support


//#define E_AXIS_MOTOR_09 //kuo exper

Is this the same as going back from 16 µsteps to 32µsteps? I just flashed the 1.8° firmware from here: https://github.com/bgiot/Prusa-Firmware-BMG

I aready swapped the 0.9° for the 1.8° that was delivered from bondtech this morning. My prints are SOOOO much better now! I had a lot of other factors under suspicion which I resolved now one by one, that's why I didn't think that the motor was the problem. But it turns out it was ONLY the motor that was responsible for the problems.

Here are some quick comparison shots of your tower, made with my smartphone against the diffused sun through a frosted windows:
https://drive.google.com/drive/folders/1-aG05mUYI70VUQ-aKP3QY1xkS8EEhSUN?usp=sharing

The seam on layer change has been much improved and now VFAs have shown up. Otherwise the print quality is about the same. So I conclude that the 0.9° does not bring any benefit even if it had enough torque, so I'll stay with the 1.8° for now.

One thing I noticed: If I place some objects in between the diffuse light source and the print, the VFAs become much more visible. This might help when trying to take pictures of the VFAs with a smartphone and natural light.

Publié : 11/04/2019 12:41 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

Regarding belt vs gearing, I did build a version with printed double helical gears and even though it had some backlash, the extrusion was still much better than stock. Overall I find the belt solution better and more robust. But I’d say the gear ratio is the biggest win if just looking at print quality.

Thanks. You saved me the trouble of doing that experiment. Definitely agree that gearing ratio creates the marked improvement. Gears are more compact, but it is a lot easier to print a smooth running pulley system like your Skelestruder.

I universalized your pulley and belt to NEMA17 mounts. Now that advantage can be had on multiple extruder designs - not just specific to you Skelestruder. If people like what they achieve with my gearbox, they can be more confident about taking the big leap.

I played with integrating the pulley into the MK3S motor plate and extruder body, but that shifts the motor only 4 mm closer to the x-axis rods. For the amount of effort, it doesn't seem worth doing.

Publié : 11/04/2019 5:42 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

Is this the same as going back from 16 µsteps to 32µsteps? I just flashed the 1.8° firmware from here: https://github.com/bgiot/Prusa-Firmware-BMG

In my firmware, that also does a few things with trinamic tuning for the 0.9 degree motors.

I did add BMG support in my fork so it is easier to mix the BMG with 0.9 degree motors on the x and y axes. With a geeked extruder and 0.9 degree motors on the X & Y axes my MK3 prints like a completely different machine. The largest surface noises now are pulley/belt tooth interactions. Those I've already more than halved with Balitensin 10 mm long drive pulleys. Somehow, they managed to shape their teeth to more cleanly involute with the GT2 belts than other pulleys I've had.

Publié : 11/04/2019 5:46 pm
Spacemarine
(@spacemarine)
Estimable Member
Re: Bolt-On Gearbox Extruder Experiment


Gears are more compact, but it is a lot easier to print a smooth running pulley system like your Skelestruder.

In addition, a belt drive has usually less noise (only noticeable at high speeds), less play, is much easier to adjust and (when using softer plastics) has less wear because the force gets distibuted over much more teeth at the same time.

So if space isn't the most important aspect, I'd choose a belt-system any time, especially if I would print the gears myself.

Publié : 11/04/2019 6:26 pm
Chocki
(@chocki)
Prominent Member
Re: Bolt-On Gearbox Extruder Experiment


I wonder how much of this improvement is from the belt as opposed to the gear reduction. This is one of the reasons turntables and tape decks used belt drives - it helped reduce the motor non-uniformity that would manifest as wow and flutter in the audio.


Belt drive turntables have a heavy rotational mass turntable driven by a rather thin belt drive, this allows the belt to stretch and contract quickly whilst damping any speed changes with the mass of the turntable. In this application you have a relatively thick non stretching belt, driving a low rotational mass extruder.
I don't think the belt is doing much to reduce the visible artefacts, but rather the ratio reduction. This is just my thoughts so don't take it as gospel.

Normal people believe that if it ain’t broke, don’t fix it. Engineers believe that if it ain’t broke, it doesn’t have enough features yet.

Publié : 11/04/2019 9:20 pm
vintagepc
(@vintagepc)
Membre
Re: Bolt-On Gearbox Extruder Experiment



I wonder how much of this improvement is from the belt as opposed to the gear reduction. This is one of the reasons turntables and tape decks used belt drives - it helped reduce the motor non-uniformity that would manifest as wow and flutter in the audio.


Belt drive turntables have a heavy rotational mass turntable driven by a rather thin belt drive, this allows the belt to stretch and contract quickly whilst damping any speed changes with the mass of the turntable. In this application you have a relatively thick non stretching belt, driving a low rotational mass extruder.
I don't think the belt is doing much to reduce the visible artefacts, but rather the ratio reduction. This is just my thoughts so don't take it as gospel.

Oh, I definitely recognize there is a "moving" inertia involved in those cases, but in applications like this where there is back pressure on the other side of the belt (from the liquid plastic) that provides a limiting factor if it is more than the belt's natural springiness/stretch. Thus, the belt's natural flex will offer some damping to irregular motion from the extruder motor (exactly the cause of the artifacts) so long as the bondtech gears have some sort of preload on them from hot-end back pressure.

Whereas for direct drive on the bondtech, the back-force is almost certainly less than the motor torque so the stepper will just plow its way through.

Publié : 11/04/2019 9:28 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

I'm with Choki in thinking the gearing ratio is making much more difference than vibration reduction by the belt.

Time period of extruder irregularities, seen as diagonal lines in test towers, is far longer than mechanical vibration frequencies that could be dampened by such a short, stiff belt.

Publié : 12/04/2019 1:30 am
teookie
(@teookie)
Eminent Member
Re: Bolt-On Gearbox Extruder Experiment

Wow this is so cool. I was considering going skelly but I think I'll give this a shot first.

Thanks for sharing your idea/design!

Publié : 12/04/2019 1:51 am
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

The Skelestruder is a nifty design, but even I feel a bit intimidated by it. This is a smaller project to test the waters.

I have an entire bag of printed Skelestruder parts and a Bondtech extruder upgrade kit in hand. Neither are going to happen soon - this is working so well.

Meanwhile....

What happens if you take a motor service modified Bear extruder,
Slice it down the filament channel like an MK3S to create front motor plate and easy extruder install.
Add internal shaft hubs for gearing like a Bondtech or Skelestruder
Fuse the new motor plate to the front 3/4 of a Bunny Science gearbox to make a unified motor plate/gear assembly

Abomination or wonder?.... Moves motor 4 mm further towards x-rods. Is it worth that much work? Bunny scientists are printing one now..

Publié : 12/04/2019 3:38 am
teookie
(@teookie)
Eminent Member
Re: Bolt-On Gearbox Extruder Experiment

Are you printing that with supports then? I was looking at doing that but didn't see a way to print w/o them.

Publié : 12/04/2019 7:53 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

Yes, the combo motor plate and gearbox can't be printed without supports. Still refining...

If I can get it to work, should be applicable to MK3S extruder. It's still a bit nuts compared to just adding the Bunny Science gearbox. This only gains about 10% moment arm improvement, but it is a fun exercise.

Publié : 12/04/2019 10:25 pm
Bunny Science
(@bunny-science)
Noble Member
Topic starter answered:
Re: Bolt-On Gearbox Extruder Experiment

If you are using a Bear extruder and want to play along, here is mfs file with the experimental parts already plated.

[EDIT ]
Fixed an interference on idler door against motor plate. Changed plating to place motor plate front down.

Also STL for optional easy motor service x-carriage & carriage clamp

Publié : 13/04/2019 12:02 am
Page 2 / 3
Partager :