Fixing Pulley Innacuracy

[aviphysics 9]

What is difficult to judge about the quality? You can pretty easily measure the eccentricity of the stock pulleys, as well as the backlash.

 

[zumfab 3]

What is difficult to judge about the quality? You can pretty easily measure the eccentricity of the stock pulleys, as well as the backlash.

 

That is kind of hard to do before you buy and recieve them :mrgreen:

 

[aviphysics 9]

That is kind of hard to do before you buy and recieve them :mrgreen:

 

Well, for that you just have to make your picks and take your chances.

 

[macua85 52]

C'mon peeps, we need new pulleys!

[izzy 33]

Hi all, this post is probably way to late but reference the set screws.

The main problem with the set screws gripping a shaft is that they bite into it and damage the shaft.

Have a look for no-mar set screws, they come in several forms usually a nylon tip or a small silver tip, both tips deform to grip the shaft surface with out damaging it, although I have used these in aerospace R&D equipment I fear that with the constant direction reversal they would allow some slipage over time so a good quality "cup point set screw" (you get what you pay for stay clear of cheap ones) that would cut into the shaft a little would retain alignment accuracy.

Although adding a little loctite bearing loc to the pulley would lock it in position on the shaft, perhaps added after after calibration alignment.

[jonnybischof 60]

I've used pointy setscrews (no clue how the correct term in english is...). They do bite into the shaft and leave a dot, but they will never ever slip. Provided you don't change your 3D printer gantry all the time, there shouldn't be a problem if the setscrew leaves its mark on the shaft...

 

[zumfab 3]

I am on the fence about what pulleys to get. I like foehnsturm's, but the one thing that bothers me is that the supply is extremely limited. Ordering extras or replacements a little later is not really an option. Quality seems really high though.

So, if I were to go for GT2 pulleys and belts, what do I need to look out for? Material? Attachment screws/method? I do not really have a clue what makes a good pulley and how to differentiate bad from good

 

[jonnybischof 60]

I look for pulleys that are CNC machined instead of the cheap "assembled from several parts" ones. The latter usually have issues with roundness and may often have sharp brows on the edges.

A pulley made on a CNC lathe has the best roundness possible.

I also prefer pulleys that have two setscrews instead of just one.

Materials? You'll probably find only aluminum pulleys. There are of course different kinds of aluminum alloys - but usually you won't find out which type is used.

 

[aviphysics 9]

@JonnyBischof

Please remind me where you got your GT2 pulleys from. I am pretty sure it wasn't Misumi, cause they don't carry the right size.

 

[anon4321 16]

I believe Jonny purchased them from here : http://www.robotdigg.com/

See http://umforum.ultimaker.com/index.php?/topic/1611-fixing-pulley-innacuracy/?p=64995

I purchased mine from here (much more expensive) : http://www.adafruit.com/

 

[aviphysics 9]

That is what I thought. I was pretty sure that both of those are multipart pulleys (though the core might be all one piece.)

 

[anon4321 16]

Mine are working well. I can't see any "out-of-round" and the axis move very, very smoothly. Much better than with the original UMO MXL pulleys

 

[jonnybischof 60]

That is what I thought. I was pretty sure that both of those are multipart pulleys (though the core might be all one piece.)

 

I just took one of the Robotdigg pulleys to the test (milled it into pieces...). It is actually a multipart pulley, BUT not in the traditional way.

It consists of 3 parts: the core and two rings. The core is a tube with the shaft bore on the inside and the pulley teeth on the outside. This is one single, CNC machined unit. The two rings are then just what the other multipart pulleys are.

But, afaik, the usual multipart pulleys have four parts: A tube, a folded piece that wraps around the tube and forms the teeth, and two rings that hold everything together. These are the "bad" pulleys which are likely to have roundness issues.

The robotdigg ones don't have any roundness issues and are very well machined even when inspecting them under a stereo-microscope.

...

Actually, I just took apart one of the UMO pulleys - they are made in the very same way as the Robotdigg pulleys (no folded piece for teeth, it's actually solid). Machining quality is worlds apart though. Very ugly under the microscope, and they only have one setscrew.

Maybe that second setscrew adds more to the overall "roundness" than I thought.

 

[zumfab 3]

Actually, I just took apart one of the UMO pulleys - they are made in the very same way as the Robotdigg pulleys (no folded piece for teeth, it's actually solid). Machining quality is worlds apart though. Very ugly under the microscope, and they only have one setscrew.

Maybe that second setscrew adds more to the overall "roundness" than I thought.

 

Just so I understand correctly - I imagine that with two set screws, you still push the axle to one side. Am I wrong and how so? I would imagine you need four screws to eliminate that, but I might be missing something.

 

[anon4321 16]

I did notice that with the single set screws original pulleys, the out-of-roundness problem was always in the direction opposite the force the set screw applies.

So when people say tighten the heck out of the set screws, be careful.

I think with two, the force required is much less so is less likely to distort the pulley. One, the force is split between two pulleys and two because two screws biting into the shaft "grabs" more.

 

[zumfab 3]

I did notice that with the single set screws original pulleys, the out-of-roundness problem was always in the direction opposite the force the set screw applies.

 

Do I understand correctly that the out-of-roundness is not strictly a manufacturing problem, but something that happens when you tighten the pulley and force it out of whack?

 

[jonnybischof 60]

Do I understand correctly that the out-of-roundness is not strictly a manufacturing problem, but something that happens when you tighten the pulley and force it out of whack?

 

It seems like that's the case. My UMO pulley (that has never been used) seems perfectly round. But it's made of "weak", soft aluminum and not a hard alloy like aircraft grade aluminum or something. So if you put a strong force on it through the single setscrew, it may very well warp.

 

[anon4321 16]

I agree with Jonny, It's the force from the set screw that distorts the pulley.

 

[zumfab 3]

I have finally slapped a belt tensioner on the printer. Yes, singular, as I grossly overestimated how far the screws actually go into the stepper motor. The design had been slightly adjusted to allow for a greater range, but that begs the question: how tight do the short belts actually need to be? I can pretty much go crazy, but I am a bit scared to stretch the belt and load the stepper motor unnecessarily.

 

I agree with Jonny, It's the force from the set screw that distorts the pulley.

 

Good to know, that will be useful.

 

[jonnybischof 60]

I'm not a professional, so this might be wrong:

I tighten my belts "as little as possible", but "as much as is needed" to reduce backlash to a minimum - if the belt is not tight enough, you can feel backlash movement when you turn the pulleys back and forth.

You don't want to tighten the belts that much as to bend the shafts (which does not take as much as I had thought...).

Also, if you pluck the belts (like a guitar), the sound should be similar on all the belts (of the same length). This works better with the long belts.

Putting too much force on the motor shaft actually can be a problem. But it takes a really tight belt to put that much force on the motor - definitely too tight then.

In that matter, I'm much more concerned when I see the current direct-drive extruders. They put a lot of force on the motor shaft...

 

[Dim3nsioneer 557]

I agree with you Jonny about the force on the motor shaft. However, it doesn't seem to be a problem as I can't recall having read anything about bent motor axes or damaged bearings so far. And the UM2 you are most probably referring to is now out for over a year.

 

[zumfab 3]

I'm not a professional, so this might be wrong:

I tighten my belts "as little as possible", but "as much as is needed" to reduce backlash to a minimum - if the belt is not tight enough, you can feel backlash movement when you turn the pulleys back and forth.

 

I think I will go with as much as needed and a little more. That way the strain should not be unreasonable and I have peace of mind that things are tight enough. Maybe it will result in a little extra wear, but it should not be much.

 

I agree with you Jonny about the force on the motor shaft. However, it doesn't seem to be a problem as I can't recall having read anything about bent motor axes or damaged bearings so far. And the UM2 you are most probably referring to is now out for over a year.

 

It's actually an UMO. One of the short belts kept working itself slightly loose, so I finally tried to fix that permanently rather than retightning every so many prints.

I am using a reworked version of this tensioner with a long screw and spring.

http://www.thingiverse.com/thing:40041

 

[Dim3nsioneer 557]

I am using a reworked version of this tensioner with a long screw and spring.

http://www.thingiverse.com/thing:40041

 

They're not bad. I also used them before I switched to direct drive...

 

[zumfab 3]

They're not bad. I also used them before I switched to direct drive...

 

Direct drive is certainly on the list, but I consider it prudent to go one (or two) steps at the time. That way I don't have to go on a wild goose chase after that one variable of the many I changed that is causing trouble. I hope.

I have been thinking about some sort of way to eliminate the couplers, as I don't like them very much. It must be possible to link the shaft to the drivers directly without loading them too much at the first slight misalignment.

 

[Dim3nsioneer 557]

Don't forget you will get rid of the variable 'short belts'...