Best belt drive ever! (If I do say so myself)

Mike,

You mentioned that the idler bearings should be positioned .002" to .005" above the top of the upper belt surface. For people not using super-straight 8020 and the like, it could be very hard to ensure that the suface supporting the belts is parallel to the linear guides within such a small tolerance, especially over longer distances.

I thought about spring-loading the idlers, but the tension would have to be greater than the force exerted on the belt, or the belt would lift the idler up instead of the idler keeping the belt down. But a 425 oz-in stepper with 4:1 belt reduction and a 10-tooth T5 drive belt could exert more than 300 lbs on the belt. I don't think the belts would like being pushed on with that kind of force 24 hours a day. Ditto for the linear bearings and rails.

Any suggestions on how to address this issue? Thanks.

Quote:

---

But a 425 oz-in stepper with 4:1 belt reduction and a 10-tooth T5 drive belt could exert more than 300 lbs on the belt.

---

I believe that Mike recommends a 15 tooth minimum for maximum belt life.

I wouldn't think it's a problem having a slightly looser tolerance. It might affect precision by a very small amount (.002-.003??), but if you can't build to within .005, then it shouldn't be noticeable. I would thing the spring loading would be a very bad idea, not to mention quite complicated, especially if you want 200+ lbs springs.

Quote:

---

Originally Posted by jalsina

Mike,

You mentioned that the idler bearings should be positioned .002" to .005" above the top of the upper belt surface. For people not using super-straight 8020 and the like, it could be very hard to ensure that the suface supporting the belts is parallel to the linear guides within such a small tolerance, especially over longer distances.

I thought about spring-loading the idlers, but the tension would have to be greater than the force exerted on the belt, or the belt would lift the idler up instead of the idler keeping the belt down. But a 425 oz-in stepper with 4:1 belt reduction and a 10-tooth T5 drive belt could exert more than 300 lbs on the belt. I don't think the belts would like being pushed on with that kind of force 24 hours a day. Ditto for the linear bearings and rails.

Any suggestions on how to address this issue? Thanks.

---

Well, that .002-.005 is just ideal for tooth engagement. As the gap gets bigger, it's easier to pull the belt out of engagement down below, then ZZZZZrpp, and you have to reset everything. The closer you get the belt to the linear bearings, the better. spring loading wont kill it, I think, but as you say, spring force needs to me more than your max linear drive force.

Quote:

---

Originally Posted by HomerSimpson

I created a better video that provides better views and more explanation here:

---

Homer, do you have any updates on your belt drive system? Inquiring minds want to know!
Cheers!
pr

Quote:

---

Originally Posted by jalsina

Mike,

...But a 425 oz-in stepper with 4:1 belt reduction and a 10-tooth T5 drive belt could exert more than 300 lbs on the belt. I don't think the belts would like being pushed on with that kind of force 24 hours a day. Ditto for the linear bearings and rails.

Any suggestions on how to address this issue? Thanks.

---

So, yes, I would increase the number of teeth to a minimum of 15, and then your motor can put 200 lb into the belt. That's not so bad. Though we like to limit it to 100 on 32mm wide belt in a bang-bang actuator, a CNC is a different animal and can be run at higher because you really don't need millions of cycles at any one spot like an actuator.
Just because your motor can put that out of course does not mean it will ever be asked to. Someone can correct me, but CNC accelerations are lower than you might expect, down around .2G. So your moving weight times .2 plus cutting force is what's expected of the motor. Simple math, and unlikely on a router to ever be 200 lb. at the belt, I'm thinkin'

Mike,

Thanks for your reply. The 10-tooth pulley was picked for high resolution, not force, but high worst-case force is a result. With spring-loading, it would be an issue, but without it, it's not.

You're right -- it's hard to imagine a wood router ever needing that kind of oomph.

John

Quote:

---

Originally Posted by PaulRowntree

Homer, do you have any updates on your belt drive system? Inquiring minds want to know!
Cheers!
pr

---

Paul,
Nothing major to report. Machine is working well. The only update I have is that the edges of my timing belts were beginning to fray as they were rubbing on the 90deg corners on the aluminum blocks that are attached to the linear slide, as well as the corners on the extrusion that the bottom belt rests in. So I machined fillets on all the edges and polished them to minimize this problem. Also, I had originally put the bottom belt in the extrusion with double backed tape, but I found with higher loads this would pry up in spots. So I tool the tape off and have epoxied the belt down to the aluminum extrusion. I'm playing with different epoxies to see what works best as it needs to bond well to the rubber material of the belt (and the aluminum of course).

I went to the woodworkers show here in Portland today and got inspired to upgrade my machine with a) a 5th axis for a lathe and b) a laser so I can engrave with it. Of course it will probably take me ages to do it but half the fun is in the planning and building.

HomerSimpson: to bond your belts to the extrusions, you could use something like 3M VHB tape. It is used to bond exterior windows to skyscrapers. Some rack and pinion routers use it to bond the rack to the frame. Call 3M technical service. They deal mostly with industrial customers, and are very helpful and very competent. They shipped me more tape than I will ever need as a "sample." They also warned me to apply it carefully, as the only way to remove it once applied is to cut it off with a thin wire. In your case, that might not be possible due to the vertical sides of the extrusion.

Yes, VHB is what we use, and most definitely go slow when laying it down, preferably with a block to smooth it. Bubbles can be dealt with by slicing across them with a razor blade. Clean the belt with acetone, though even then you can peel it off of the VHB, but the drive will never be able to shear it off of the tape.

I called 3M as you suggested but they didn't recommend VHB for my neoprene belts. They sent me a few different things but I haven't had a chance to try them yet:

2216 two-part epoxy
6035 low fogging adhesive transfer tape
9458 adhesive transfer tape used with adhesive 300
9471FL adhesive transfer tape with adhesive 300lse

what is the belt material you have used to adhere the VHB to?

I have been a member of this site for almost 4 years trying to see what will work best for a CNC Plasma Table built on a budget and slowly purchasing items. I have recently decided to try Mike's design; however, I have a few questions that I cannot seem to find within this post. For my plasma design I am using Thompson Rod. I have attached drawing of my idea.

I am not sure of belt width; I am planning on using a T5 belt but as far as width do not see anything on width to load. My gantry weighs about 75#'s and there will minimum load when cutting. I have been looking at using 300-400 oz in motors with a 4:1 reduction, all I am looking for is a maximum travel speed of 125 ipm.

weldmetalman,

I'm not sure about the load carrying capacity of T5 but as a reference, my 3mm HTD belts are 9mm wide, my gantry is probably 45lbs, my steppers are nema23s with a similar torque as what you have, and my gear ratio is something like 3.4:1. I can reliably machine at 250in/min and I've run it without machining, at 2000in/min. That should give you a reference for what is possible.

Thanks for the reply.

Quote:

---

Originally Posted by weldmetalman

I have been a member of this site for almost 4 years trying to see what will work best for a CNC Plasma Table built on a budget and slowly purchasing items. I have recently decided to try Mike's design; however, I have a few questions that I cannot seem to find within this post. For my plasma design I am using Thompson Rod. I have attached drawing of my idea.

I am not sure of belt width; I am planning on using a T5 belt but as far as width do not see anything on width to load. My gantry weighs about 75#'s and there will minimum load when cutting. I have been looking at using 300-400 oz in motors with a 4:1 reduction, all I am looking for is a maximum travel speed of 125 ipm.

---

Might wanna rethink that 125ipm max for a plasma. Thinner metals can be cut upwards of 250ipm with the hypertherms.

Quote:

---

Might wanna rethink that 125ipm max for a plasma. Thinner metals can be cut upwards of 250ipm with the hypertherms.

---

You are correct in your statement, If I didn't know what I was doing and was going to try to cut thin gauge at 120 amps; however, there is a knob on all plasma units to turn down the amperage as well as different size cutting tips designed to cut at lower amperages. So I should be just fine cutting at a maximum speed of 125 IPM, just as long as I know how to setup a plasma cutter correctly. Thanks for your help but that's not the type of help that I am looking for.

Quote:

---

Originally Posted by weldmetalman

You are correct in your statement, If I didn't know what I was doing and was going to try to cut thin gauge at 120 amps; however, there is a knob on all plasma units to turn down the amperage as well as different size cutting tips designed to cut at lower amperages. So I should be just fine cutting at a maximum speed of 125 IPM, just as long as I know how to setup a plasma cutter correctly. Thanks for your help but that's not the type of help that I am looking for.

---

Weldmetalman-

Maybe you misunderstand my response. I never said you "couldn't". If you could do a job twice as fast, why not?

I never questioned your know how, regarding the use of a plasma cutter. I didn't mean it negatively either. I simply figured i'd throw out a consideration to not "physically" limit your maximum speed of travel for some headroom later down the road if you choose to cut at higher feed rates (or atleast have a decent rapid speed). "HomerSimpson" even suggests what is possible with similar nema 23's. Physically Limiting yourself with a gear ratio that can only produce 125IPM on a plasma table is just limiting yourself down the road. Time is money! :)

Has anyone tried to use a closed belt loop (teeth facing out) running like a tank-tread on the fixed belt, instead of the extended open belt-loop design that Mike has developed? It would have a longer belt that would give more stretch/bounce, but it might be simpler to design. Reverse flex may be an issue too...
Any thoughts or experience?

I think it would work, but the devil is in the details. There are reverse flex issues with Mike's drive as well, but these are dealt with by using idlers of sufficient diameter. However, with a tractor belt, the belt profiles must match perfectly, with no backlash. In Mike's design, backlash can be adjusted out by changing the driver to idler distance such that there is no free play at the tooth interfaces that are doing the work (even though there may be free play on the non-working side of the tooth). With a tractor tread, the belt cannot be adjusted for backlash, so the teeth must fit perfectly. Furthermore, construction may not be simpler. You still need a minimum of three wheels, including one that moves to adjust belt tension. The shorter tractor belt might be cheaper than the full-length belt, but it will also wear out more quickly, and when it does, it must be replaced because there is no adjustment for backlash.

Has anyone tried putting flanges on the outside idlers so that the upper belt can't wander off the lower belt? That might eliminate the need to put the belts in a groove or channel. Would it work? Opinions?

Quote:

---

Originally Posted by jalsina

Has anyone tried putting flanges on the outside idlers so that the upper belt can't wander off the lower belt? That might eliminate the need to put the belts in a groove or channel. Would it work? Opinions?

---

I tried. The problem was that this only contains the belt right at the idler. The belts will become misaligned away from the idler and then as the idler assembly moves toward this misaligned section, the forces will be too great for the idler to bring them back into alignment.

I'm not saying it won't work, but my implementation of it didn't work, which was basically putting oversized washers on the idlers to try to contain the belts.