Pasted from my Thingiverse entry: Description Over the last couple of months I've read in the forums about people having difficulty with the belts' teeth rubbing on the pulley flanges or fender washers and causing print quality issues. The other night I remembered that flat belts on a crowned pulley will automatically find the center of the pulley. Flanges or fender washers are completely unnecessary! I consulted my old copy of Machinery's Handbook, 24th edition, and even though it's from the early 1990's, it has a chapter on flat leather belts, how to splice them, how to design pulleys, etc. The quoted rule of thumb is for 1/8" of crown for every foot of pulley width. That would be imperceptible and probably ineffective on a pulley this size. I suspect that recommendation has to do with the geometry and strength of the leather belts. I took a guess at a 10mm radius crown for the 6mm wide GT2 belt, and it worked great on the first try. This will concentrate the belt tension in a narrow band instead of spreading it across the whole width of the belt, but I think there is more than enough strength in the belt for that. The belt doesn't track exactly on center due to the various inevitable misalignments of my printer, but it does return to to its happy spot pretty quickly when disturbed. Video demo on my MakerGear Prusa with GT2 belts:YouTube Files in my GitHub fork of the Prusa Mendel: Github Instructions Just print and press the bearing in, then mount as usual. Versions for 608 and 688 bearings are included. My MakerGear Prusa uses 688 bearings, so I haven't tested this with a 608. It should work just fine, but please post in the comments if you find another crown radius works better. I also haven't tried it on my Y axis because I don't want to disassemble the printer just for this. https://reprap.org/forum/read.php?14,108277,108277#msg-108277 Sun, 16 Aug 2026 00:41:05 -0400 Phorum 5.2.23 https://reprap.org/forum/read.php?14,108277,142005#msg-142005 https://reprap.org/forum/read.php?14,108277,142005#msg-142005
Because if anything the diameter itself of the bump would throw the belt away, and would never draw it in. To draw it in, it would have to be a groove and not a bump. But the bump is small enough so the this "inclined plane" vectorized force can be overcome by the bigger friction force. As long as the materials friction coef is good enough. Make the bump too big after a certain point depending on the friction coef of the materials and the belt would be just thrown away and would never stick there.

So diameter is measured in cm, and cm is not a measure of force, hence regardless of how many cm u get of "something" it cant make things move by itself. Belt tension is like a spring, but also doesnt mean anything by itself except that at a point the belt mai fail.

The only force that is actually a force and as such, it is capable of moving things here is the friction force. It increases in the bump area because the belt increases its vectorized tension effect there. It presses more on the bump, so the friction force is directly proportionally biger. Not much, just a % hint, but enough as it only needs to be bigger with a minimal increment. Friction by definition oposes the movement, hence the belt gets stuck at that point because the biggest friction force 1) first wins out as a vector against all other forces, and 2) afterwards succeds in opposing to movements in all other directions.

The thing with "belt finds the bigger diameter" is a sort of a big overlook/shortcut. Exactly what kind of explanation you receive from some1 that doesnt want to give explanations. Its counterintuitive and doesnt help understanding. Sorry to resurect the old thread but would of been a shame to leave it at that.]]> NoobMan Mechanics Tue, 10 Jul 2012 07:54:52 -0400 https://reprap.org/forum/read.php?14,108277,141986#msg-141986 https://reprap.org/forum/read.php?14,108277,141986#msg-141986
So its a bit the other way around: it finds the point with higher friction, which happends to be on the center. Not the other way around.

Where the guy in that tutorial says "tension" you should read friction. The belt tension by itself also like the diameter, has no meaning. What it does do, it increases the force to which the belt presses against the pulley. Now friction force is [friction coef * perpendicular force].

By increasing the tension in the belt, the belt pushes harder against the bump, and therefore its "weight" increases, and the friction in that point increases even if the friction coef stays the same. Thats the "magic" of attraction. If you think of centimeters diameter, ofc its "magic". Once you think of friction, there is no "magic" anymore.

When we walk friction is the force that vectors to our behind and depends on our weigt * friction coef of our shoes with asphalt. The area of contact has no deal in it. We can have 1 ton of mass on ashpalt and either if it contacts on a 100sqm area or on the top of a needle, the friction force is the exact same regardless.]]> NoobMan Mechanics Tue, 10 Jul 2012 06:25:36 -0400 https://reprap.org/forum/read.php?14,108277,110824#msg-110824 https://reprap.org/forum/read.php?14,108277,110824#msg-110824 Buback Mechanics Wed, 21 Dec 2011 13:31:40 -0500 https://reprap.org/forum/read.php?14,108277,110821#msg-110821 https://reprap.org/forum/read.php?14,108277,110821#msg-110821 Dale Dunn Mechanics Wed, 21 Dec 2011 13:21:23 -0500 https://reprap.org/forum/read.php?14,108277,110820#msg-110820 https://reprap.org/forum/read.php?14,108277,110820#msg-110820 Rarely is that in the middle.
[woodgears.ca] explains it pretty well and has pictures.

V drive pulleys are the ones that track to the middle and are move common, these days, in belt drives.]]> criswilson10 Mechanics Wed, 21 Dec 2011 13:05:35 -0500 https://reprap.org/forum/read.php?14,108277,110817#msg-110817 https://reprap.org/forum/read.php?14,108277,110817#msg-110817 nophead Mechanics Wed, 21 Dec 2011 11:52:46 -0500 https://reprap.org/forum/read.php?14,108277,110768#msg-110768 https://reprap.org/forum/read.php?14,108277,110768#msg-110768 -------------------------------------------------------
> The other night I remembered that flat belts on a
> crowned pulley will automatically find the center
> of the pulley.

To put it right, they dont automatically find the geometric middle. What they do find is the point where the friction is maximal. Or better yet they automatically get "stuck" or get "attracted" in that place with most friction. It could be the middle, but its not necesarily so.]]> NoobMan Mechanics Wed, 21 Dec 2011 04:53:17 -0500 https://reprap.org/forum/read.php?14,108277,110690#msg-110690 https://reprap.org/forum/read.php?14,108277,110690#msg-110690 Buback Mechanics Tue, 20 Dec 2011 10:59:14 -0500 https://reprap.org/forum/read.php?14,108277,110682#msg-110682 https://reprap.org/forum/read.php?14,108277,110682#msg-110682 Dale Dunn Mechanics Tue, 20 Dec 2011 09:16:40 -0500 https://reprap.org/forum/read.php?14,108277,110638#msg-110638 https://reprap.org/forum/read.php?14,108277,110638#msg-110638 In a pinch, I've use a heated, bent paperclip to burn gear teeth into a crowned plastic pulley.]]> criswilson10 Mechanics Tue, 20 Dec 2011 00:50:12 -0500 https://reprap.org/forum/read.php?14,108277,110598#msg-110598 https://reprap.org/forum/read.php?14,108277,110598#msg-110598 Buback Mechanics Mon, 19 Dec 2011 16:12:50 -0500 https://reprap.org/forum/read.php?14,108277,108662#msg-108662 https://reprap.org/forum/read.php?14,108277,108662#msg-108662 belt align the pulley. If the prototype moves on to production, then I use "regular" pulleys to save a penny or two.

I haven't tried printing one yet, I usually just screw a regular wide pulley to a drill bit, spin the drill up, and drop
a 120 degree file across the belt area to cut a crown.]]> criswilson10 Mechanics Fri, 02 Dec 2011 14:52:22 -0500 https://reprap.org/forum/read.php?14,108277,108495#msg-108495 https://reprap.org/forum/read.php?14,108277,108495#msg-108495 -------------------------------------------------------
...
> Crown pulleys will work, but they cost more to buy
> and take more material to make than a standard
> pulley

Commercial ones, yeah. Printed ones are dirt cheap and won't catch on the belts' teeth.

Do you use crowned pulleys on a regular basis?]]> Dale Dunn Mechanics Thu, 01 Dec 2011 16:26:12 -0500 https://reprap.org/forum/read.php?14,108277,108488#msg-108488 https://reprap.org/forum/read.php?14,108277,108488#msg-108488
Crown pulleys will work, but they cost more to buy and take more material to make than a standard pulley]]> criswilson10 Mechanics Thu, 01 Dec 2011 14:28:54 -0500 https://reprap.org/forum/read.php?14,108277,108277#msg-108277 https://reprap.org/forum/read.php?14,108277,108277#msg-108277 entry:

Description

Over the last couple of months I've read in the forums about people having difficulty with the belts' teeth rubbing on the pulley flanges or fender washers and causing print quality issues. The other night I remembered that flat belts on a crowned pulley will automatically find the center of the pulley. Flanges or fender washers are completely unnecessary!

I consulted my old copy of Machinery's Handbook, 24th edition, and even though it's from the early 1990's, it has a chapter on flat leather belts, how to splice them, how to design pulleys, etc. The quoted rule of thumb is for 1/8" of crown for every foot of pulley width. That would be imperceptible and probably ineffective on a pulley this size. I suspect that recommendation has to do with the geometry and strength of the leather belts.

I took a guess at a 10mm radius crown for the 6mm wide GT2 belt, and it worked great on the first try. This will concentrate the belt tension in a narrow band instead of spreading it across the whole width of the belt, but I think there is more than enough strength in the belt for that.

The belt doesn't track exactly on center due to the various inevitable misalignments of my printer, but it does return to to its happy spot pretty quickly when disturbed.

Video demo on my MakerGear Prusa with GT2 belts:YouTube

Files in my GitHub fork of the Prusa Mendel: Github

Instructions

Just print and press the bearing in, then mount as usual.

Versions for 608 and 688 bearings are included. My MakerGear Prusa uses 688 bearings, so I haven't tested this with a 608. It should work just fine, but please post in the comments if you find another crown radius works better.

I also haven't tried it on my Y axis because I don't want to disassemble the printer just for this.]]> Dale Dunn Mechanics Tue, 29 Nov 2011 22:52:51 -0500