Making your own timing pulleys - a cautionary tale

[David Morrow]

The following is a copy and paste from the Summer 2013 edition of Digital Machinist, Page 61. I don't make my own timing pulleys but do use them on my CNC router. I thought it was an interesting story that falls under that Rumsfeldian category of "We don't know what we don't know". What appears to be simple can have unintended consequences that may be hard to resolve.


" SPEAKING FROM EXPERIENCE
The article, "How I Used CNC to Make a Cutter and Timing Pulleys," in the Spring 2013 issue brought back memories of working for a small company that made small industrial robots that were driven by timing belts. As I was new, I studied the design of the robots carefully. One day I asked the owner if the robots lost or gained a couple of encoder counts occasionally. He said yes, that had been a problem and they had not resolved it. I pointed to the timing pulleys driving the encoder and told him to look closely at the tooth profile. You could clearly see that the belts didn't match properly. It turns out that he had sent a timing pulley to a manufacturing operation they had set up in Russia and they reverse engineered the
pulleys from that sample. What they didn't account for was that as the belt wraps around the pulleys it changes shape, and the diameter of the pulleys makes a difference. So, a small pulley requires a very different profile than a large pulley!

Take a timing belt and wrap it around a small diameter, smooth cylinder. You'll see that the teeth on the small pulley must be much smaller and almost rectangular in shape, while on a larger diameter they have the more typical trapezoid profile. I was aware of this because I once tried to buy a hob for making timing pulleys and found out that I would have needed five different hobs to cover the whole range of pulley sizes I was interested in. I suspect that the large timing pulley made in the article works quite well. However, the small pulley is going to run quite poorly and wear the belt rapidly. "

[Roger]

That quote is not entirely correct.

I would not try to use a gear cutting hob to make a toothed belt pulley UNLESS the hob was designed for that sized pulley. Wrong technology.
But you CAN make any size toothed belt pulley you want if you use a single-tooth cutter. (This is a bit like using a single point thread mill to cut all sorts of threads.)

To explain: yes, the gap between the teeth will seem to change depending on the number of teeth on the pulley. But this is deceptive, because you really need to start at the 'pitch line', not at the surface of the pulley. The pitch of the belt stays constant. Furthermore, the shape of the 'tooth' on the belt does NOT change, and it is that shape you must focus on. (It's a bit hard to change the shape of a solid lump of hard rubber after all!) You do NOT focus on the shape of the tooth on the pulley: that is irrelevant.

Yes, I do make my own GT2 toothed belt pulleys of a range of sizes. To do this properly you need a cutter of the shape of the tooth on the toothed belt, which is a bit tricky as such cutters are not normally available. The profile is patented or copyright or something by Gates, who are a shade close about it. So first buy some belts from Gates, examine the profile and read the patents (they are deliberately utterly obscure), then make your own cutter to that profile. Don't just try to guess the dimensions: either look them up or calculate them. It works.

Caution: the Chinese sell lots of 'GT2' pulleys and belts in a very small range of sizes - for the RepRap market. Well, they CALL them GT2, but they are NOT the Gates profile. Dunno what they are - just a Chinese copy and simplification I think.

Cheers
Roger

[Nate]

...
To explain: yes, the gap between the teeth will seem to change depending on the number of teeth on the pulley. But this is deceptive, because you really need to start at the 'pitch line', not at the surface of the pulley. The pitch of the belt stays constant. Furthermore, the shape of the 'tooth' on the belt does NOT change, and it is that shape you must focus on. (It's a bit hard to change the shape of a solid lump of hard rubber after all!) You do NOT focus on the shape of the tooth on the pulley: that is irrelevant.
...

Do you model the pitch line a polygon, a rounded polygon, a circle, or some other shape?

[ArtF]

    The pulley DXF from Gearotic does show the pulley tooth profile, its spit out with the DXF if you output a dxf.
What youl notice is the tooth shape usually agrees with the belt, but not always, some Gates for example are purposely
misshaped and are deeper to allow the rubber to flex into the tooth. This means there is a limit to the minimum size of
a pulley when they do that..and they make some adjustment to the form for that.. ( or just to keep the patent.. )

  When it comes to specs, I have seen them purposely wrong, arc radii that dont match to other arcs, or wrong
coordinates.. so yeah, you do have to check and be carefull , but if you have a 3d printer, its quick to print a test,
and if your cutting a blank, I really advise a cheap test on a wooden or plastic blank just to be sure and to see where
one may want to adjust his/her numbers a touch..

  Ive had very good luck 3d printing gt-2's, my Laser tables Z drive is driven by two printer gt2-3 pulleys and so far
seem to be working very well..

Art

[Roger]

Do you model the pitch line a polygon, a rounded polygon, a circle, or some other shape?

It won't be a polygon because the rubber and the cords are not infinitely flexible.
You could use a rounded polygon, but the difference between that and a circle would be extremely small - maybe sub-micron. I use a circle, and so do Gates as far as I know.

there is a limit to the minimum size of a pulley

Definitely what Art says. Bend the belt too tightly and you would weaken it as well.

When it comes to specs, I have seen them purposely wrong, arc radii that dont match to other arcs,

Funny about that. In fact, even the authorised suppliers do this to some extent: they publish engineering drawings of the profile but those drawings are only approximations.
However, remember that we are dealing with a 2 mm pitch here, so the deviations are very small, and the face of the belt is covered with (very special) cloth which does have some compliance. You can get close enough without too much effort.

One other thing: you have to put a little tension on the belt to get it to settle properly into the teeth on the pulley. If you don't have enough tension you may mash the teeth. And you do need to get the diameters correct for the same reason.

Cheers
Roger