[heavycnc]

Hi Ger21,

Something I plan to do different from everman's design is slow down through a planteary gearbox, the shaft for the pinion, so that my pinion sprocket can be a larger diameter. The purpose is to engage more teeth. From the pictures it looks like everman's pinion has 7-9 belt teeth locked in. I'm going to use a 3.5" pinion which will double the amount of teeth engaged and use a 50mm wide belt. I also spoke with brecoflex application engineers about the bigger T10 belt. It has some play in it which cancels that option. These two things should increase the payload capabilities

[ger21]

I wonder is that 100lb per axis or 100lb for both X axis

It's 100 lbs per 32mm belt. So, if you have two belts, you can accelerate 200 lbs at 1G, or 400 lbs at .5G.

You can also use a wider belt.

[turmite]

It's 100 lbs per 32mm belt. So, if you have two belts, you can accelerate 200 lbs at 1G, or 400 lbs at .5G.

You can also use a wider belt.

Gerry, what does the 1g and .5 g translate to in ipm? I have been approached by a company needing 3-5 machines just like the one I am about to start, and I really want to use this belt system, but I have to know it will work under this kind of load, and that I can keep it fairly dust free?

Has anyone found a solution for dust mites? I am talking lots of dust with
3d wood carving.

Mike

[Bear5k]

Not Gerry, but G = 9.8m/sec^2. This is acceleration/deceleration. It means that at the end of 1 second, you are going 9.8m/sec (aka ~386ips) and will have traveled 4.9m. If you cut acceleration in half, then halve all of the prior numbers. For a CNC machine, this is still blisteringly fast, and you will be speed-limited somewhere else in your mechanicals (e.g., moment arms).

[PaulRowntree]

I see that people are using T5 belts here, and that Mike has said it works with T5. The tooth/groove profiles are not exactly the same width according to sdp-si. I calculate a mismatch of 0.28mm. Why does this mismatch not cause slip/backlash?
Cheers!

[ger21]

I see that people are using T5 belts here, and that Mike has said it works with T5. The tooth/groove profiles are not exactly the same width according to sdp-si. I calculate a mismatch of 0.28mm. Why does this mismatch not cause slip/backlash?
Cheers!

I think that most belts and pulleys need that slight bit of clearance to get longer belt life.

Are you basing the .26mm on the belt profile alone? If so, then yes, there's a very noticeable gap when meshing belt to belt. When I wrap the belt around the pulley, though, the fit appears to be very good.

Until I actually get to test this, I really don't know if backlash will be an issue. It doesn't seem like it will be.
If it is, though, Mike has said that it can be reduced by turning the OD of the pulley down, so that the belt rests lower in the grooves, eliminating any slop in the meshing of tooth to groove.

[jsheerin]

Gerry,
I think Paul is talking about where the belt mates to itself, away from the pulley. However, I think if you preload the belt by having a vertical adjustment of the pulley, you'll pull the belt inward towards the pulley on both sides which will load the teeth on one side of the pulley against the teeth on the lower belt in one direction and the teeth on the other side of the pulley against the opposite teeth faces. That should prevent backlash.

[ger21]

Yes, I started to say that, but left the rest out.

When they make pulleys, does every different size pulley have different size teeth? That would make sense, as the smaller the diameter of the pulley, the smaller the spacing between the belt teeth as you wrap it around. If the teeth size don't decrease as pulleys get smaller, than it would seem that the smaller the pulley, the tighter the fit. Anyone have the answer to this?

[PaulRowntree]

jsheeerin, that is exactly what I was wondering, and your reply makes sense. It would seem that positioning the belts during the preload would be a critical part of the process. Thanks!

One more question ger21 : why do you have two idlers on each side of the drive pulley? I understand the diameter issues for your delrin rollers, but what is the function of the outer skate bearings?

Cheers!

[ger21]

I'm sure that Mike mentioned why somewhere in this thread, but I don't feel like looking for it. I'm just copying what Mike did. It's either to get more teeth engaged, or to keep the belt from flopping around. Or both.

Thinking about it, it's probably the first. Without it, all the load would be carried by the 2-3 teeth directly under the large rollers. The second one spreads the load out a bit more.

[Mike Everman]

Gerry, very nice build! Looking good. Sorry I haven't been by in a while, I see I could have chimed in in a couple of places. One comment on the idlers: minimal runout is very important, it is why we do not put bearings into a made roller. Best to get some cheap dual row ball bearings from VXB and just use them as is. Likewise pay close attention to the runout of your pinion.

Comments/answers in no particular order:

1. Don't use Kevlar! Kevlar is not good for back-bend stress cycles. We tested it under moderate load, and it did not last through the night. Steel is the way to go.

2. Tooth shear load is roughly equivalent to the recommended tensile maximum for the belting (T5=300N/10mm width), so increasing the number of teeth in the pinion is putting effort in to the wrong place, and making it more sure that you will need a reduction pass. Better to have more idlers down below to get more teeth involved there, but really, there is "more than 1" teeth in engagement under each idler. The pinion at the minimum recommended number of teeth is best, which for T5 is 15 teeth (.94" Pitch Dia).

3. Yes, there is tooth clearance that is very noticeable on the belt to belt interface, but moving up the motor pinion, takes up this clearance.

4. T10 is capable of twice the load of T5, per 10mm width. Our T10 version is 50mm wide vs the 32 for T5, so we are spec'ing it at 3x load capacity.

5. Yes, pinion teeth get thinner as diameter goes down.

6. With regard to the tooth clearance at the pinion, which would be the only source of lost motion (I hesitate to say backlash, which I interpret as a "zero stiffness deadband") you can solve this as someone pointed out, by turning down the OD of the pinion (works to a degree), or plating the pinion (don't build it up too much, or it won't lay into the tangent points of the pinion which, while working, makes it springy). Breco has a near zero clearance version of the pinion stock, but beware, some batches of belting interfere with a half wrap, some do not.

7. Each T5x32mm belt is good for 20-30 million cycles if run back and forth to the same spot at 50-25 lb linear force (15 tooth pinion, so approximately 25-13 in-lb torque, respectively. Take a look at the graph I'll attach. Saying it's only good for 100 lb is just me keeping it in the "sweet spot" of life vs. cycles. CNC's do not hit the same spot on the belt like the life test, so feel free to run it at higher levels, like 150 lb (75 in-lb).

[ger21]

One comment on the idlers: minimal runout is very important, it is why we do not put bearings into a made roller. Best to get some cheap dual row ball bearings from VXB and just use them as is.

Is this a factor in precision, or load capacity, or both?

1. Don't use Kevlar! Kevlar is not good for back-bend stress cycles. We tested it under moderate load, and it did not last through the night. Steel is the way to go.

Too late! I'll use the Kevlar for the flat belt, and pick up some steel.

But I may be redoing the whole thing to make it stronger. I can go with these.
10 Bearing 1621ZZ 1/2"x1 3/8"x7/16" Shielded:Ball Bearings

I'm assuming there's very little load on the outer bearings, correct?So I can go with 1/2" shoulder bolts and these bearings for the inner Idlers, and the 5/16" bolts with skate bearings on the outer and I think I'll be good.

I just need to find someone to bore out my pinions for me. I don't have the equipment to do is myself accurately enough. I hacked the one in the pics just to get it together.

[ger21]

Thanks for all the help, btw.

[heavycnc]

Gerry if you want I can give you price to have the machine work done for your pinions/plates and any other machine work needed for your cnc.

My email [email protected]

[Mike Everman]

Hi Jerry,
The outermost idlers are taking little load and the outer diameter matters not. The inner however have a lower limit of I think 30mm dia, which is a back-bending limit spec. We use 35mm OD dual row ball bearings and two of those stacked to get just under 32mm width.
These are just the cheapest way for precision runout and reduction in tolerances that lead to runout. Clamping the inner races axially or bond them to the pins they are on for best repeatability. Why is a long story, but suffice to say that slop at the ID of these idlers directly hurts repeatability.

Be sure to get your belts down in a groove in the table, too, or build up a sidewall some other way as thick as the two belts together. It's the best way to keep the upper from a walking out disaster.

General notes:
Don't push down on the idlers, height should be .002-.005 above the belt top.
Don't clamp the ends of the upper belt. It needs to be able to find it's happy place when you tension the belt at the motor.

[Bear5k]

Don't clamp the ends of the upper belt. It needs to be able to find it's happy place when you tension the belt at the motor.

Mike -
Since I'm seriously thinking of setting one of these up for a test axis, may I ask a clarifying question here? Do you clamp it once the motor and top belt are tensioned?

In the spirit of "no stupid questions, only stupid people"...

Wouldn't having the top and bottom belts interlocked up and down the length of track basically combine with the idlers to create a clamping force? In other words, do the end clamps provide any meaningful force if you:

• Have the bottom belt taped down,
• Have the top belt aligned so the teeth mesh, and
• Put the carriage in place so the front and back idlers are in place and preventing the rest of the belt from being put under tension?

Or is there more fine-tuning to setting the tension on the top belt than just pulling upward with 25# of pressure? The knuckle draggers are curious...

Thanks!

[Mike Everman]

Hi bear,
No fine tuning required, and no tension on the free ends of the upper belt. Just something to keep it from slithering off. If you've gotten a length of belt on the edge of a table, slither is the only way to describe how it gets to the floor. Ha

Your bullet list is the essentials of it, yes. All this for a single purpose that isn't terribly obvious. is a drive method that has inherent damping amd stiffness without a lot of stick-slip. This is a very good combination. It allows you to tune the servos hot for the best path following (for the dollar, of course!)

[HomerSimpson]

Mike,
I'm glad you've checked in. I too have built your drives system but with stepper motors (post 178). I'm using 3mm HTD belts that are 9mm wide and dual row wide ball bearings with an 8mm ID for the main idlers. I've machined slots to put the belts in to prevent them from tracking sideways.

To tension the belt between the pulley on the motor and the pinion bearings, I push up on the motor bracket and tighten 4 screws which mount the bracket to the assy which holds the pinions and slides. The belt is then very tight, however, when I drive the motor back and forth at 200 in/min and an acceleration of 5-10 in/min^2, which is relatively tame, the belt loosens up in just 1 or 2 cycles. No matter how hard I tighten the belt, it still loosens up, even though it appears the motor mount has not moved.

Any suggestions you have would be very helpful. Thanks.

[Mike Everman]

Hi Homer,
Well, the HTD rounded tooth form is less forgiving of ride height changes. run an indicator from the carriage to the belt mounting surface to see what you're getting there.
Tightening this rounded form a lot is going to make for high wear, especially if your gap is too much. Ways I can think of around this are: pushing down with the idlers harder, more of a pinch, and less belt tension at the motor pinion.
So after you get a feel for the running parallelism, then let's talk.
If it's pretty parallel, which it should be, then you may have to shim down your idler holder piece for more interference where belt meets belt.

[tjskcnc]

I was hoping for a DIY version of this coming out, but it appears that, after several years, this is not going to happen anytime soon.
During this time I came across 2 other designs similar to this.

The first is from Joe's 4x4 Hybrid forum, where he was testing a belt drive system several years ago, probably about the same time this post was started. He has a video showing this system working on a 10' rail and designed to fit the current mounts for his 4x4 hybrid, running rapids at 1000ipm. Further testing had rapids over 1700ipm. The Yellow and Green parts are MDF. He apparently abandoned the belt system in favor of R&P.

The second is a setup from Rockcliff CNC. It doesn't have the second set of idlers, but we're DIYers and could probably set up something to work from it, and it is advertised at $95.