[harryn]

Is there a reason for the 3 to 1 reduction ? The reason I ask, is that it would be potentially easier to just use a 34 size stepper and direct drive.

Of course some resolution will be lost, but perhaps not enough to matter for some wood routing. ?

Thanks

Harry

[ger21]

Resolution and force. A 640oz stepper direct driving a 1" pulley will probably provide less than 40lbs of force at 500rpm. And your resolution will probably be somewhere between .005-.01". My 570oz Nema 23 with 4:1 reduction should provide 3 times the force, with .002 resolution.

[Zebethyal]

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.

I am not quite understanding why you feel there would be backlash with the "tank/tractor track" approach, surely both methods would use the same idler setup to ensure both belts meshed correctly end ensure correct tension due to downwards pressure of the idlers, only difference is one has long ends, whereas the other has a loop, it has already been stated that the long ends could to all intents and purposes be left loose and have no effect on the performance.

The only issues I can see are increased wear on the tank track loop (already noted), ensuring that the loop does not hit anything as the top half "wobbles" about and ensuring that the belt is "truly endless" though I am pretty sure that they manufacture them to very tight tolerances anyway - consider all the applications that use them.

Reverse flex would be at its worst by the drive gear and idlers, the loop over the top would be nowhere near as bad.

Please by all means explain further if I have misunderstood.

[jalsina]

Zebethyal, You're perfectly right. When I saw "tractor tread," I imagined a tread in the shape one sees on crawler tractors, not with the loop and backlash adjusting idlers. Only problem I can think of is that since the lower belt will no longer be covered up and kept clean by the long upper belt, so some means will be needed to prevent dirt from settling on it.

[Zebethyal]

You are absolutely correct - the "cover" effect of the top belt is definitely another added bonus for the 2 long belts approach although I am sure dust particles would make their way between the belts on that solution as well - they do seem to get everywhere

I wonder if anyone would care to comment on their experience in this regard - HomerSimpson perhaps, as I believe he has the only machine that is in active use with this belt drive system.

I too have been making use of the ideas posted here to come up with a belt driven design, no idea if I will actually build it, my 4x4 became an 8x4 and has gone through nearly 15 almost complete re-designs, still not settled on a final design, and still not sure if I actually need a machine quite that big - still designing only costs time - much friendlier on the pocket that way

I actually sought out this thread after seeing one of Homer's videos on youtube, to see how popular a means of axis drive this was.

This is my variation and some of my attempts at integrating it with a table

[Mauser]

I've been thinking about doing the "Tank Tread" thing myself, and perhaps the solution to dust is to have the fixed belt on the underside of the machine, or at least facing down.

[greybeard]

"What goes around, comes around" is quite true.
My post #53, 02/01/2009 !

Hi Mike.
Because of the difficulty of finding timing belt by the metre over here(that's now been located), I wondered if there would be any disadvantage in looping the top belt back over the rollers, so in effect the top belt is just a typical short belt turned inside out

I realise the top belt did have a function in keeping a lot of debris out of the lower one, but that's easily sorted.
John

[Zebethyal]

Belting Online is the cheapest supplier of the various types of Timing belt I have found in the UK so far, they also seem to have the biggest range.

Cheaper than anything being offered on ebay.

Hadn't thought of placing the belt on the underside of the machine for keeping dust out - worth further investigation - My R&P solution had the teeth facing down, for this very reason.

[yngndrw]

What is the pricing on this belt system ? I was thinking about using it with MakerSlide (The MakeSlide Project) but both a future 3D printer and a future pick & place machine. I need to price up various options so I know which direction I should be going in.

I'd also like to know what the pricing of the KAOS system is - It is probably outside of what my budget would be but it's worth asking.

Thanks.

[ger21]

What is the pricing on this belt system ? I was thinking about using it with MakerSlide

It's designed around THK type rails, so won't work with other linear rail methods, unless you build your own version.

[harryn]

Resolution and force. A 640oz stepper direct driving a 1" pulley will probably provide less than 40lbs of force at 500rpm. And your resolution will probably be somewhere between .005-.01". My 570oz Nema 23 with 4:1 reduction should provide 3 times the force, with .002 resolution.

Thanks for the reply Gerry. It has been a long road for me to start understanding what microstepping can - and cannot readily do, and I still am not entirely sure sometimes.

If I understand this correctly, then if I were using a very strong stepper motor at 1/10th microstepping and at say 100 rpm, then this is still not as good as a similar torque generated using 3:1 reduction, because I cannot really rely on the stepper motor to hold micro steps.

On the other hand, if instead I am using a servo motor, with the fine tuned steps generated from the etched resolvers, then these "steps" are essentially full steps that can be counted on. In that case, I could use a servo motor with this belt design in direct drive, and be fine - right ?

[ger21]

On the other hand, if instead I am using a servo motor, with the fine tuned steps generated from the etched resolvers, then these "steps" are essentially full steps that can be counted on. In that case, I could use a servo motor with this belt design in direct drive, and be fine - right ?

Yes, that's what Mike is doing.

[Mike Everman]

Wow, I've missed a lot. I'm not getting alerted to new posts, but should be!

Zebethyal, nice work, man!

As to flanges on idlers or pinion: nothing works better than the belt in a full length groove, or tack down some guide strips on both sides. Other tries at controlling the tracking of the belt all ate themselves!

As to strength of belt, which really in this case should be re-stated to strength of 1.x teeth in engagement under the idlers, we routinely run T5 PU steel reinforced at 75-100 lb linear force at a width of 32mm. Go higher for less life. If you use 150 lb, then any one tooth taking an acceleration load will do it less than 1 million times (out and back to that tooth).

I am doing a 4:1 belt reducer as a reduction stage for our bigger T10 drive. I'll attach a pic.

I apologize for not checking in more often! It seems the only thing (still) that I can offer the DIY'ers is my experience with it. We are still strictly an OEM supplier, but I'm happy to help DIY'ers when I can.

[stanglou]

Alright guys,
I have all the parts spec'd for my large router using the belt drive described here, but I cant seem to track down a good source for the pullies. I need the T5 32mm wide pully for the belt drive, and then I need pullies for the reduction. Doesn't have to be t5 for that, want to keep it cheap. So where do you guys get your pullies for?

Thanks
Lou

[HomerSimpson]

here's where I bought my belts and pulleys:

Metric Pitch Pulleys - Industrial Timing Belts & Pulleys Polytech Design Inc. Clifton NJ 07011

[Zebethyal]

Here's another version that makes use of the "tank track" idea and can potentially be mounted sideways. less adjustabile due to the location of the bearings, but easily changed to wider belts if required, simply by changing the widths of the drive gear and spacers along with the number of skate bearings.



The top idlers could easily be moved in and also have skate bearings if the friction against the spacer was considered too great, it even uses a standard size belt at 575mm.

My main issue at present is integrating it with the rest of the model, short of bolting extrusion directly to one of the side plates, I have not come up with a decent solution.

[Mauser]

One thing you might need to deal with is an ability to adjust the tension in the drive pulley in the middle of the tank tread depending on the "Timing" of the connection to the fixed belt, while still keeping the bearings snug.

[Zebethyal]

One thing you might need to deal with is an ability to adjust the tension in the drive pulley in the middle of the tank tread depending on the "Timing" of the connection to the fixed belt, while still keeping the bearings snug.

The intention there was to have some lateral movement possible in the first set of bearing idlers, but I agree with you 100%.

Still very much a work in progress - as I said, I cannot even work out how to attach anything to it to even make it useful

Cost wise, not sure if there is much of a gain here over conventional R&P, by the time you have made use of a wide enough belt 25 or 32mm (possibly 2 of them at full length) you have already exceeded the cost of the rack (which will be longer lasting).

Gearbox items are pretty much the same items and therefore the same cost cost either way, as such the gearing should make the speed on a given axis about the same.

Less aluminium plate required for a CNC router parts style gearbox, or even a more "boxy" equivalent.

Having expended a decent level of effort in trying to make this work for me from a design view point (and much more required), I have not yet convinced myself this is the way to go.

[greybeard]

Hi Zebethyal,
would it be feasible to have the "cheek plates" of your design act as continuous belt loop guides by bringing them inboard of the drive belt from the stepper, and do away with the flanges on the idler shafts ?
John

[Zebethyal]

Something more like this you mean:

It does save on parts and size - still haven't a clue how to mount anything to it though