Need help for making a timing belt pulley

Hey guys, need some advice here. Long story short, we have an Italian mill with a damaged spindle shaft. We can't get the belt anymore so we need to modify the spindle to match newer style of belt available. According to some info found on the site here I posted a picture below of the numbers taken to get to the diameter I need to machine it to match the new belt. The shaft should be 2.048" OD so the bottom of the teeth are inside the pulley. Original size of the spindle was 2.187" or metric version of that. I made a test on plastic shaft before I cut the real one. It doesn't match. The belt binds when I try to wrap it around the pulley. Where did I go wrong? I also tried with 15 teeth because it was closer to the original size of the spindle. But it still did the same thing. It binds but even worse for the 15 teeth. Hope somebody can help me out. Tks

.500 x 15 teeth = 7.5"
7.5" x 3.1416= 2.387"
2.387" - .180 <- x2 dept of tooth
2.207" compare to the 2.187 original.

I almost don't even know where to begin...

First, you need to educate yourself on the actual physical characteristics of the belt tooth form being used. One cannot simply mill plain grooves into a cylinder and expect any manner of utility ( much less success or accuracy ). There are several tooth forms to discern from. Normally, they are Hobbed like gears. However, these can be milled, if one has the proper profile and suitable milling cutters, to suit.

You should feel glad and lucky that it fit so poorly. If you had tried to actually run it you would undoubtedly have damaged the belt ( if not destroyed it ), as well as possibly the pulley and other surrounding objects.

I do not understand why you simply did not purchase a proper pulley and bore it to suit.

Your second photo says nothing of value because the pulley has one pitch diameter and the belt sitting on top of the teeth effectively has another. So don't judge anything from it.

In the third photo, it looks like your belt has a slighter longer pitch. I would just adjust the pulley by a bit: make it slightly larger in diameter and see how it works.

As for your math, it is the effective PITCH diameter that counts, not the OD. You only need the OD to actually make one. And you can not use the length of the outside of the belt because when the belt is installed, it's outside surface will be above and therefore at a larger diameter than the pitch diameter or even than the OD of the pulley.

I would suggest that the the manufacturer of the belts should be contacted for the actual methods for calculating the size of the pulley blank and the tooth form. I would also suggest that this may qualify as a good part to 3D print.

Black70 said:

Hey guys, need some advice here. Long story short, we have an Italian mill with a damaged spindle shaft. We can't get the belt anymore so we need to modify the spindle to match newer style of belt available. According to some info found on the site here I posted a picture below of the numbers taken to get to the diameter I need to machine it to match the new belt. The shaft should be 2.048" OD so the bottom of the teeth are inside the pulley. Original size of the spindle was 2.187" or metric version of that. I made a test on plastic shaft before I cut the real one. It doesn't match. The belt binds when I try to wrap it around the pulley. Where did I go wrong? I also tried with 15 teeth because it was closer to the original size of the spindle. But it still did the same thing. It binds but even worse for the 15 teeth. Hope somebody can help me out. Tks

.500 x 15 teeth = 7.5"
7.5" x 3.1416= 2.387"
2.387" - .180 <- x2 dept of tooth
2.207" compare to the 2.187 original.

View attachment 252334View attachment 252335View attachment 252336

Click to expand...

I've hacked out a few of these over the years. The OD of the pulley is the pitch circle diameter for timing belts. It is critical to have it correct. The teeth of the pulley provide positive traction, but they cannot fix an out of pitch diameter situation because then the belt wants to advance or retard on the pulley. Then it gets chewed up.

The pitch circle circumference must match the theoretical length of the belt in so and so many teeth. What pitch is your belt?

The belt is a model H so .500 pitch. It's 48 teeth and 24" long. The angle of the tooth is 40 deg. To start I just cut a 1/4 straight endmill cut. After I used a flycutter ground to 40 deg to make the angle on each side. With the math it's very close, but not close enough. Next step will be to play with the OD of the shaft until it just has the spacing right. I thought with the calculation it should of been dead on. Oh well.

Timing belts are in most cases standard. Are you sure the original is not just a metric belt? And I am still not certain what is it that you need to do. If you need to replace the belt and the pulleys find the centre distance and buy two pulleys and a belt that will give you that distance. If there is no combination of belts and pulleys to give the right centre distance, one can get a next size longer belt and add an idler tensioner. But again, it is not clear what is the exact issue here.

Sure, where can you buy the gear. I did educate myself a lot actually on how to make it. Not a lot of info when somebody want to make it themselves. I'm looking for one number. The actual OD of the shaft I need to machine it too. The rest is easy. And according to most websites, when a trapezoidal belt. They are pretty much always 20 degrees. But the OD has to be bang on. Sorry to disturb you Mr. Kopf.

Wlodek said:

Timing belts are in most cases standard. Are you sure the original is not just a metric belt? And I am still not certain what is it that you need to do. If you need to replace the belt and the pulleys find the centre distance and buy two pulleys and a belt that will give you that distance. If there is no combination of belts and pulleys to give the right centre distance, one can get a next size longer belt and add an idler tensioner. But again, it is not clear what is the exact issue here.

Click to expand...

I can't get the belt with the same pitch anymore. They do not exist. It's an old Italian mill with a metric pitch. Somebody tried to repair it but parts are not available anymore. Previous people who worked on it replaced the rear pulley but didn't realize from spindle was not quite the same pitch and ruined the belt. It was kind of Ok when working slow but was jumping a tooth eventually when going faster. So we machined the spindle, now we need to make the splines on it to match the pitch of the new belt. Seemed pretty simple but also not so simple anymore. So, what size does my shaft have to be to machine 14 teeth with a pitch of .500 minus the thickness of the tooth. Which is .090". and that belt will not bind in the splines.

I beleive there is only one pulley.
The spindle shaft is one piece with the belt teeth cut into it to maker the driven "pulley"? If so no reason it could not be cut down to a much smaller diameter and a pulley attached.
Bil lD.

There's a lot of different types of timing belts, but iirc it's common that the pitch diameter is actually in the middle of the belt. That puts it outside the o.d. of the sprocket.

Normally you can find a drawing of the rack that is the belt in a belt supplier's literature.

The sides of the teeth need to be curved, too, btw. If you wrap a tooth with straight sides around a circle then you will see that you generate curved teeth. It's similar to an involute gear. Look for an 'odontograph' to get the idea.

It would possibly be quicker and in the long run cheaper to give the job to a shop that cuts sprockets and gears ... It's pretty easy if you have the correct tools.

Bill D said:

I beleive there is only one pulley.
The spindle shaft is one piece with the belt teeth cut into it to maker the driven "pulley"? If so no reason it could not be cut down to a much smaller diameter and a pulley attached.
Bil lD.

Click to expand...

You are right. That is exactly what I am trying to do. Here is a picture of it. I thought instead of making a whole new spindle, just to weld the spline section, turn to diameter and recut the spline to match new belt I got. There are splines "dog" underneath too for when you are switching from hi to low speeds.

Black70 said:

Sure, where can you buy the gear. I did educate myself a lot actually on how to make it. Not a lot of info when somebody want to make it themselves. I'm looking for one number. The actual OD of the shaft I need to machine it too. The rest is easy. And according to most websites, when a trapezoidal belt. They are pretty much always 20 degrees. But the OD has to be bang on. Sorry to disturb you Mr. Kopf.

Click to expand...

You are not disturbing me. I slept peacefully.
As for where to purchase a pulley, I'm really not sure what to tell you if you have already considered using Google and the approximately one hundred common outlets for such things. That doesn't even begin to point out the obvious, like eBay and Amazon.

In the interest of being helpful, I submit to you a primer that should illustrate it well - HELPFUL ASSISTANCE

Black70 said:

Bill D said:

I beleive there is only one pulley.
The spindle shaft is one piece with the belt teeth cut into it to maker the driven "pulley"? If so no reason it could not be cut down to a much smaller diameter and a pulley attached.
Bil lD.

Click to expand...

You are right. That is exactly what I am trying to do. Here is a picture of it. I thought instead of making a whole new spindle, just to weld the spline section, turn to diameter and recut the spline to match new belt I got. There are splines "dog" underneath too for when you are switching from hi to low speeds.

Click to expand...

Ah. I have to admit that this was not clear when I originally read the opening, either. I think you would be much better off overall if you were to correctly determine the original belt form and simply purchase a suitable pulley for the other one, with belt to match. As others have alluded to - not much has changed with belts over the decades. I'd be pretty surprised to learn that you really could not source a replacement. We've been able to find replacement belts for machines made 60 - 65 years ago.

Still, if you simply must make it, you'd be far and away better off to educate yourself about the proper math for making them, and the various forms extant. It's not rocket science, nor brain surgery. Timing pulleys are fairly simple affairs. There are a number of gear shops in your neighborhood ( Ontario ) capable of knocking that out in less time than it would take you to do the math and set up, but I do understand when a hobbyist just wants to force their will upon something for the rights to say they did so.

Talk with these people about belts and pulleys and stop REINVENTING THE WHEEL

Small Mechanical Components: Precision Gears, Timing Belts, Gear Assemblies, Timing Belt Pulleys, Couplings, Bearings and much more - SDP/SI

Leading Manufacturer of Precision Components | WM Berg

Regarding the pitch line of the belt and the corresponding pitch diameter of the sprocket, what Hu wrote above is incorrect and what Emanuel wrote is correct. On an H section timing belt the pitch line is outside the OD of the sprocket. I'm not at the shop at the moment to tell you the exact offset amount. This is why your belt does not mate up properly with the sprocket you made.
If you must have a custom sprocket, any gear shop should be able to cut this for you without breaking the bank. If you can use an off the shelf sprocket and adapt it, that would be the simplest solution. Good luck.

Dan from Oakland said:

Regarding the pitch line of the belt and the corresponding pitch diameter of the sprocket, what Hu wrote above is incorrect and what Emanuel wrote is correct. On an H section timing belt the pitch line is outside the OD of the sprocket. I'm not at the shop at the moment to tell you the exact offset amount. This is why your belt does not mate up properly with the sprocket you made.
If you must have a custom sprocket, any gear shop should be able to cut this for you without breaking the bank. If you can use an off the shelf sprocket and adapt it, that would be the simplest solution. Good luck.

Click to expand...

Yup, I recalled incorrectly. I would have looked at the engineering in the old Browning catalogue before I went ahead and cut it

This catalogue doesn't show a 15 tooth, but it has a 14 and a 16. Taking the average, the OD of the sprocket should be 2.333"

I politely suggest to the OP that his assumptions are incorrect.

Pulleys are not *hard* to make but are generally very very exact in multiple ways not-obvious to one who is not making industrial-type pulleys or industrial automation routinely.

There is a very important crest on the top of the teeth, and an important shape to the teeth, and a pretty important shape to the root.

Any industrial shop used to making pulleys *can* make any of them, but the cost will likely exceed the OPs budget and expectations by 2-10x.
Because the industrial shop will likely/usually insist on making the part *right* to avoid potential expensive reclamations later.

The easiest and cheapest way is to buy any 2 modern pulleys and one belt for the same transmission ratio, in the same size give or take.
Then bore to fit, or use taperlocks, or bolt to the side (ugh), or whatever, depending on rpm, power, size, costs, torque etc.

The pics + H profile seem to indicate fairly high power in 2-5-10 kW range, of low rpm (sub 3000?), and low positional accuracy.
So does "old", and a milling spindle.

(HTD-5 limited low speed torque and low speed power output) or HTD-8 bored to size or taperlocks are easy, available everywhere, cheapish.
T8 metric or gates GTxx are more expensive and just as good as HTD but likely less noisy in 2000 rpm++ speeds.

To the OP.
You can download OD, PD, pics, data and cad models of pulleys from Misumi EU, and many pulley makers like maybe gates, cross +morse UK, others, mcmaster.

If you download a 3D model of a modern (GT profile, T8, etc.) pulley (misumi, partsexpress) nothing is straight - the teeth are complex curvilinear shapes.

YOU yourself for personal use can grossly approximate the shape of the teeth and the pulleys will work fine, if maybe a bit more noisy and less positional accuracy - unlikely to be needed in this power transmission app.
- Proper shops often don´t want to do this for commercial liability reasons.
- I suggest making a 3D cad model using downloaded IGES models, to check the process.
This would take me less than one hour.

"Inside the belt" is what I would have thought. That means that the teeth will be spaced closer than what the belt pitch would indicate. AND, those teeth on the belt will be just as wide as they were when the belt is straight (or even a touch wider due to distortion) when they are bent around the pulley, SO the spaces between the teeth on the pulleys must be full size and the pulley teeth must be a little smaller. Like an involute gear tooth form, the tooth forms for the pulleys WILL change for each and every number of teeth on a pulley. So you need DIFFERENT cutters for each tooth count. But I am sure that they must come in ranges, just like gear cutters.

You say the belt is 0.5" pitch but then you say it is metric. Perhaps you need a better measurement. Someone probably still make belts and pulleys in that pitch. And they would probably be usable. Or just buy new pulleys and belts in the same pitch and machine them to fit.

English and metric belts:

McMaster-Carr

And pulleys:

McMaster-Carr

There are many other sources. Worst case, you can use a slightly different pitch belt and pulleys AND add an idler to take up the slack. I am assuming that the center to center distance can not be adjusted.

EmanuelGoldstein said:

There's a lot of different types of timing belts, but iirc it's common that the pitch diameter is actually in the middle of the belt. That puts it outside the o.d. of the sprocket.

Normally you can find a drawing of the rack that is the belt in a belt supplier's literature.

The sides of the teeth need to be curved, too, btw. If you wrap a tooth with straight sides around a circle then you will see that you generate curved teeth. It's similar to an involute gear. Look for an 'odontograph' to get the idea.

It would possibly be quicker and in the long run cheaper to give the job to a shop that cuts sprockets and gears ... It's pretty easy if you have the correct tools.

Click to expand...

hanermo said:

Pulleys are not *hard* to make but are generally very very exact in multiple ways not-obvious to one who is not making industrial-type pulleys or industrial automation routinely.

There is a very important crest on the top of the teeth, and an important shape to the teeth, and a pretty important shape to the root.

Any industrial shop used to making pulleys *can* make any of them, but the cost will likely exceed the OPs budget and expectations by 2-10x.
Because the industrial shop will likely/usually insist on making the part *right* to avoid potential expensive reclamations later.

Click to expand...

I agree that buying off the shelf arts is easiest but you are making way too much of the difficulty. Sprockets are simple.

Make sure the OD is correct - not a big deal.

Put the blank on an arbor and indicate it - 15 minutes.

Put on your change gears - 15 minutes.

Touch off, count the number of witness marks to make sure your brain didn't go on a trip. Maybe 15 minutes.

Go check the print again to make sure your memory was not mistaken. That way you can avoid having a 150 pound doorstop that looks real cool but has the wrong number of teeth.

Back up, go into depth minus .015 or .020 and cut teeth.

Measure.

Adjust size, take a finish cut.

Measure.

Bill.

Ship.

It's that simple. All the complicated stuff is done by the interaction of cutter and machine. No one in the gear business should be afraid to cut a sprocket. Or a spur gear. Helicals take a bit more brains, bevels yet more, woims and woimgeahs can be a bitch. But sprockets and spurs, piecea cake.

Real gearshops don't even put teeth on the drawing, it's pointless. We don' gotta show you no steenkin' 3D models !

EmanuelGoldstein said:

I agree that buying off the shelf arts is easiest but you are making way too much of the difficulty. Sprockets are simple.

Make sure the OD is correct - not a big deal.

Put the blank on an arbor and indicate it - 15 minutes.

Put on your change gears - 15 minutes.

Touch off, count the number of witness marks to make sure your brain didn't go on a trip. Maybe 15 minutes.

Go check the print again to make sure your memory was not mistaken. That way you can avoid having a 150 pound doorstop that looks real cool but has the wrong number of teeth.

Back up, go into depth minus .015 or .020 and cut teeth.

Measure.

Adjust size, take a finish cut.

Measure.

Bill.

Ship.

It's that simple. All the complicated stuff is done by the interaction of cutter and machine. No one in the gear business should be afraid to cut a sprocket. Or a spur gear. Helicals take a bit more brains, bevels yet more, woims and woimgeahs can be a bitch. But sprockets and spurs, piecea cake.

Real gearshops don't even put teeth on the drawing, it's pointless. We don' gotta show you no steenkin' 3D models !

Click to expand...

I was able after a few phone numbers to get to an engineer at Gates and he told me the exact OD I am suppose to machine the shaft to match the .500 pitch that their belt has if I make it for 14 teeth. If anyone goes through that it is 2.174" I already tried it once at 2.200" and still was a little off. Will try it with their specs. Thank you to those that gave links to companies who make gears. I was looking for an H style but different companies have other style letters/numbers close to same specs. A few had 14 teeth pulley with an OD of 2.176". For getting the belt of the original pitch of the spindle, it would not have helped me either. Somebody already replaced/repaired the big pulley in the mill's head and had it with a belt of .500 pitch. It was a lot easier to make another rear pulley. For some reason, gates or Jason did not make a belt to match the splines of the mill spindle. That's why I had to modify the spindle. Make new splines to match new belt or buy a pulley that matches the belt. Since I cut splines from time to time I thought it wouldn't be very hard. Never did a timing belt gear/pulley though. It was always splines or involute gears. So anyways, finding that OD was the real problem for now. Will try it out and see. Thank you for those who wanted to help.