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

Hi guys,
I'm not ready for a full release of this, so I'm starting a thread here for discussion. This summer, I'll put a complete product release thread up in the proper place.
This is called ServoBelt. The vid is (almost) self explanitory.
http://www.bell-everman.com/ServoBelt.wmv
It's been tough keeping my mouth shut about it when asked what's the best way to drive a large format DIY NC machine!
Now I can talk about it, as it is patent pending as of Tuesday.
I plan to have parts available online in the fall for people here, though our usual business is OEM customers. It would be a new realm indeed for us to sell parts and not whole systems, but since joining this forum, I have a desire to help!
Enjoy.

Its an isnpiring foruma and I am gonna send the link to my other friends

It is 3M VHB double back tape.

Hi Mike! Thanks for posting. Any updates on the larger belt unit with reduction? Can you post more pictures or the reduction please? Also any recommendations for bearing setup, type, make for the gear reduction?
I am building a 5 axis plasma w/ bellows to cover the linear ways/belts of course!
I am using 800W JVL MAC800 intergrated servos on XY and I am aiming for 3000ipm rapids.
thanks again!

Hi Mike! Thanks for posting. Any updates on the larger belt unit with reduction? Can you post more pictures or the reduction please? Also any recommendations for bearing setup, type, make for the gear reduction?
I am building a 5 axis plasma w/ bellows to cover the linear ways/belts of course!
I am using 800W JVL MAC800 intergrated servos on XY and I am aiming for 3000ipm rapids.
thanks again!

Hi King, or is it Kong? I do have photos I can post soon, we've now got a 5:1 belt reduction for SBLH and have delivered a few pre-production units. We are rolling out the ServoBelt Heavy configurations soon. It broke it's life test rig, which is being rebuilt, but before that it had attained 5 million cycles A-B-A at 300 lb accel/decel forces and looks like it will do a lot more.
I'll get to that, but want to show the laser flat bed cutting gantry we just did using ServoBelt Medium and Renishaw magnetic linear encoders (1um). I'm not at my computer, but will try. It's basically a 4'x4' with 2 meter/sec and 1g capability.
With the linear encoder feedback as the only fb, it is a phenomenally stiff and quiet servo loop. This gantry is targeting routing, waterjet and laser systems.

Attached


IMAGE(http://factorydaily.com/fdattachs/fdattachs6/112617454811709.jpg)


IMAGE(http://factorydaily.com/fdattachs/fdattachs6/112617454911710.jpg)

Hello Mike,
Interesting concept, not to say brilliant,
I downloaded the CAD drawings of a T5 belt, and the belts do not mesh exactly, there is about a 0.5mm gap between teeth´s sides,
How does this affect backlash? or do you just preload one side of the carriage to one side, and the other side to the other?

Yes there is a bit of clearance. Tensioning the belt by pulling up the motor and locking it in place removes the clearance and puts the belt teeth in pre-contact under the first, inner idler rollers.
So, you would not want the upper belt tensioned at the ends, or this benefit would not happen.

I understand how backlash is removed from the two mating belts. What I don't understand is how backlash is eliminated from the Motor pulley and the belt. There must be clearance unless you have special not clearance teeth. The only other way I can see that there is no backlash in the system is that because of high tension, the friction of the belt alone without the teeth engaging is enough to move the truck assembly. But as soon as there is a significant load, the frictional forces will be overcome and backlash will occur.
Just think of a flat belt with no teeth. There is zero backlash.
I am skeptical that there is no backlash under load conditions only because every source of belt drive information I have came across has claimed GT2 belts king for precision positioning applications where little backlash is desired and high load capacity is needed. AT-zero type belts and pulleys, are also desired for precision applications based on information from Brecoflex company.
To tell you the truth I am drooling over the servobelt concept and would be happier if I could get away with T5 belts and pulleys as they are cheaper and seem to be much easier to get a hold of in wide widths.
Can anyone please tell me I am completely wrong so I can be at peace and not worry about getting quotes from companies that don't care about selling their stuff to little guys.

Hmm good question, i was wondering about that myself...
I guess you only rely on the friction of the belt and pulley not to slip and cause backlash.

Hmm good question, i was wondering about that myself...
I guess you only rely on the friction of the belt and pulley not to slip and cause backlash.

Well I am definitely glad I am not the only person who was thinking about this. I'm going to do a belt drive like this for my next machine and I am going to use GT2 timing belts from sdp-si or AT5 timing belts from brecoflex, if they respond to my quote request...

You might get more luck calling them directly and asking for a distributor's contact.
It happens with the big companies not interested in us hobbyists...
Good luck!

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

That's an interesting idea. So in a sense it would be like a tractor tread. There is no reason that would not work, but I suspect you would not be able to eliminated the backlash between the two meshing belts the same way that the servo belt does unless you got two belts to mess without backlash. That is a pretty good idea now that I think about it.

Yes that could work,
you could remove the backlash between the two belts the same way Mr. Everman suggested just be able to move the motor pulley one belt pitch to tension the section between the motor pulley and the glued belt, then you would need another tensioner to stretch the belt over the idler circuit.
It is the same concept, a little more machining to do, but less belting needed.
Now the thing is, how to remove the backlash between the belt and the pulley's teeth.
from my measurements there is a 0.014" gap between the teeth of a stock pulley and the teeth of the belt on a T5 belt.
Unless you machine that pulley to tighter meshing, we are stuck with relying on friction to eliminate that backlash.
Then again, how many timing belts are used today to couple a servo motor to a ballscrew, there is backlash there too. much less linear backlash though because of the screw pitch.
Any ideas anyone?
Best regards
Fernando

Attached


IMAGE(http://factorydaily.com/fdattachs/fdattachs6/112617455011711.JPG)

Yes that could work,
you could remove the backlash between the two belts the same way Mr. Everman suggested just be able to move the motor pulley one belt pitch to tension the section between the motor pulley and the glued belt, then you would need another tensioner to stretch the belt over the idler circuit.

It is the same concept, a little more machining to do, but less belting needed.

Now the thing is, how to remove the backlash between the belt and the pulley's teeth.
from my measurements there is a 0.014" gap between the teeth of a stock pulley and the teeth of the belt on a T5 belt.
Unless you machine that pulley to tighter meshing, we are stuck with relying on friction to eliminate that backlash.

Then again, how many timing belts are used today to couple a servo motor to a ballscrew, there is backlash there too. much less linear backlash though because of the screw pitch.

Any ideas anyone?

Best regards
Fernando

Fernando

How are you measuring the 0.014" gap? Is it when the belt is flat or when it is wrapped around the pulley? This could make a difference.

Cheers

Don

When it is wrapped around the pulley it is about 0.014"
When it is flat the gap measures about 0.021" but this does not affect because you can remove that backlash with the motor pulley adjustment.
Fernando

Have I missed something fundamental in Mike's original set-up ?
As far as I can see, the top belt, outside of the outer idlers, has no function regarding backlash nor tension of the "active" part of the belt.
Once the belt is outside this area it may be completely loose, for all it does. In which case it could be just a free loop over the top, and doesn't even need any further idlers to guide it.
Please explain what I have missed.
John

The Servo motor actually has to turn the top belt with a toothed pulley in order to get movement.. The backlash in question exists where the top belt is being turned by the servo. If there exists any backlash between the pulley and the belt, you'd in theory have "backlash" in the movement of the top belt on motor reversal, which is ultimately pushing / pulling against the lower belt.. Unless I'm completely missing how it operates. Either way, very cool design, and I'm sure there's an answer coming!

When it is wrapped around the pulley it is about 0.014"

When it is flat the gap measures about 0.021" but this does not affect because you can remove that backlash with the motor pulley adjustment.

---------

Now the thing is, how to remove the backlash between the belt and the pulley's teeth.
from my measurements there is a 0.014" gap between the teeth of a stock pulley and the teeth of the belt on a T5 belt.
Unless you machine that pulley to tighter meshing, we are stuck with relying on friction to eliminate that backlash.

Fernando

What if you were to put some kind of coating on the pulley about 0.007" thick? Some kind of plating or anodising. Just a thought. It would be basicly the same as making a new pulley to the specs that you want or need.

I know its just a crazy idea but sometimes one crazy idea brings forth other ideas that might work.

cheers

Don

Hi Guys, been a while since I've posted some ServoBelt stuff.
http://www.youtube.com/belleverman#p/a/u/1/zTeBokV6Iv0
I did a little write up and another vid and pic on this one at mikeeverman.com
Seems like we're shipping different systems every week these days!

question , the alunminum extrustion in the video, which brand is it
not 8020? or bosch?

Yes, it is Bosch. I am constantly amazed at how straight their extrusion is!

Yes, it is Bosch. I am constantly amazed at how straight their extrusion is!

what the height and width ot the extustion, 1.5 x8 inch
or is 2 pcs stacked on top of each other
tell me what profile ver you used im been look for wide version
for my needs also.

The extrusion is 90x180mm heavy. The bearing rails are 20mm.

well I have finally got my plasma table built using drive belts ,using the info from here and there.
I have had great results so far, in fact had to make sure the dti was working properly!
But it was..
have a 1500 x 3600 table, and when running it up and down the full length of the table was getting as low as 0.04mm out.
Am not running linear bearings at all, just plain bearings running on a "C" beam i had ground up.
But i am a true belt convert.!!!
The cuts I an getting are great.
In fact the only thing letting me down at the moment is the Z axis, which is ball screw.
I hope to get some pics sorted out later.
But for those of you wondering if they really can work in real life.....
Yes they do
As to how well they will cope with the dust.. only time will tell.
But at the moment mine are doing great, even with water all over the rails, from the water table.

Hey Mike,
I was looking at your ServoBelt stuff over the weekend and it is very impressive, it makes me want to change my major from Biology to M.E. Anyhow, I was reading a few of your threads and I noticed that people have asked about your products being available to us common CNC hobbyist. Do you have any updates on products and pricing available for us?

First post on this thread but am completely awed by this design. Now I am hoping it will work for my app.
I need to build a machine for making some of my products and want to build it heavy. The spindle could be anything from one of the atc chinese spindles to a 5hp Colombo kit.
The machine is a moving table design with a movement of about 24" on the Y, 60" X and 20" Z. The table will be the Y axis. This should make the weight distribution more even on the X and Y.
I currently carve laminated wood in 3d at 100ipm. I would like to carve much faster than that. Will this belt app work (diy) for a heavy X (200 lbs) and a matching weight on the Y? The Z should be around 75 lbs.
Mike

NICE! scary fast and ingenious looking drive design.

First post on this thread but am completely awed by this design. Now I am hoping it will work for my app.
I need to build a machine for making some of my products and want to build it heavy. The spindle could be anything from one of the atc chinese spindles to a 5hp Colombo kit.
The machine is a moving table design with a movement of about 24" on the Y, 60" X and 20" Z. The table will be the Y axis. This should make the weight distribution more even on the X and Y.
I currently carve laminated wood in 3d at 100ipm. I would like to carve much faster than that. Will this belt app work (diy) for a heavy X (200 lbs) and a matching weight on the Y? The Z should be around 75 lbs.
Mike

Weird, but I seem to not get alerts of posts here.. Thanks to recent posters for the kind words. Still no solid ability to supply DIY'ers, sorry about that. I'm still thinking it will happen.
Mike, I'm not sure I understand about the X being 200 lb, is that the estimated weight of Z stage and spindle package? The Z should be a screw, I'm thinking.

Hi Mike,
Thanks for the post. The combined weight of the x carriage and the complete z axis will be about 200, maybe as much as 300 lbs. The z axis will most likely use r/pinion.
I am curious if this will move the combined weight of the x and z axis at the speeds I want/need?
Mike

Since its belt drive, how do they factor in the wearing and stretching of the belt over time?..
And under load there must be a factor of stretch?

Well, im guessing it has that clearance for a reason, maybe prevent the belt climbing on the pulleys teeth.
I dont know if it would work. have no way of testing that right now. maybe there is another type of belt and pulley profile that does not have this clearance, GT perhaps? dunno
Best regards
Fernando

Does anybody have a cad drawing of the GT profile belt? I asked this earlier in this thread as to the mesh of the two belts and the reason for choosing a trapazoidal profile versus the superior (especially in terms of backlash and power transmission) curvelinear profiles like GT.

The reason I'm trying to find an accurate profile drawing is to sketch the involute (is that the correct term?) of the belt teeth as they pull out of the "rack". This would depend on the diameter of the last idler which looks small in the original design and gives me some concern regarding backbending of the belt.

The main advantage that I can see with a T profile belt is that if will self mesh while a GT type does not (I believe) thus requireing the rack to be either machined or possibly epoxy cast using the belt as a mold.

Neat idea of turning the drive belt inside out BTW. I really like the whole concept of this flexible rack and pinion design.

Cheers

Mark

Mark
if you are referring to the lousy PAINT sketch i uploaded, it was just a concept drawing, nothing is to scale, and as you can see, the belt is less than perfect! heheh
Of course every idler arrangement should comply with the minimum backbend diameter needed for a particular belt.
if you are talking about the first original design then, no need to worry, the last idler is small, but the belt is already straight, so no backbend there.
Regards
Fernando

I was refering to the original servo belt design and the last idler prior to the driven pulley. Not sure on the spec and I guess it depends on each belt but would think the back bend radius would be larger than the smallest rated pulley.

The concept "tractor drive" paint drawing is a pretty cool version on the original. Although you are relying more on the belt to rack engagement than the original as it is only dealing with the stretch of the belt where as the later has the whole drive force acting on it.

I like these designs as to me they are a version of a rack and pinion but with better tooth engagement due to the larger number of "pinion" teeth on the belt engaging the driven pulley.

The acurrate drawings I'm trying to make are of the tooth profile (personally am planning on using GT) so that I can plot the involute of the belt tooth as it leaves and rejoins the stationary rack.

Cheers

Mark

I dont know if it would help,
but gates has a 3d model database for all their pulleys including GT2
http://www.gates.com/designview/form...ocation_id=866
You can find them there, but you need to register.
Ill attach an 18 tooth pulley for a 5mm pitch GT2 belt in .stp format which I can open in autodesk inventor, other formats are available in that same page up there.
Now providing we can sort out the backlash in the motor pulley,
Can someone come up with a nice neat way of adding a reduction ratio using timing belt and pulleys?
Regards
Fernando

Attached

Thanks for that Fernando

I already have cad files for the pulleys but I suspect that they are not the mirror image of the belt tooth profile for the same reasons I'm trying to draw the mesh with the "rack". I'll try using this profile just out of interest to see what I come up with.

Here's a link to my build thread.... about halfway down the page you ca see how I plan to do my reduction gear.

http://www.factorydaily.com/forums/showthread.php?t=53824 (Yet another Aussie biuld log)

Cheers

Mark

Hello Mark,
I saw your design, clever, but im afraid that setup might not work if you are thinking of adding the stationary belt to act as a rack, unless you eyeball exactly there the drive pulley should be. There is no adjustment in it to aid in removing the backlash between both belts.
unless I am missing something.
Regards
Fernando

Yep your 100% right... at the time I posted that drawing I was still unsure if I would make a mating rack... the design has changed somewhat since then and the motor mount and drive pulley assembly now is adjustable as required by this design.

If I had lots of spare cash I would get one of those tming pulleys with a planatary gear reduction inside them instead of the belt reduction... bit too expensive at this point.

Cheers

Mark

Mark,
YEa i was thinking the same thing, use a planetary gearbox and get on with it... simpler, cleaner, and probably less backlash.
What kind of timing pulleys with a planetary gearbox inside, i had never heard of them. Can you point me on where to find them?
Fernando

Came across this awhile ago... think they are one off type items not shelf stock and from memory were close to $1000 (don't quote me on that its just a figure that would scare me away even though they are pretty cool)

http://www.motiontech.com.au/assets/...r%20Pulley.pdf

They are made by Mectrol this was on the local distributers site. not been able to find the original site yet.

Scratch that ... just found them. They changed there name in 2005

http://www.onviollc.com/dotnetnuke/P...7/Default.aspx

quoted backlash is higher than I remember but still not much given the reduction it gives.

Cheers

Mark

Pretty innovative.
Although other than saving some space, i dont see much advantage to a planetary gearbox ($500 aprox) fitted with a timing pulley on the shaft ($30)
I think its cheaper to go with the planetary gearbox and stock pulley.
Regards
Fernando

For the questions regarding backlash in the motor pulley and the belt. In one of Mikes documents, he writes that he gets his T5 timing belts from Brecoflex. Brecoflex offers 3 times of timing pulley tooth profiles. They come in Normal ("lots" of tooth clearance), SE ( reduces tooth clearance), and Zero (No tooth clearance).
I believe it is safe to say Mike Everman uses zero clearance pulleys to achieve zero backlash in his system.

Hi Mike,
Thanks for the post. The combined weight of the x carriage and the complete z axis will be about 200, maybe as much as 300 lbs. The z axis will most likely use r/pinion.
I am curious if this will move the combined weight of the x and z axis at the speeds I want/need?
Mike

Hi Mike,
It's more about acceleration load on the belt than speed. While we've taken this T-5 belting to 150 acceleration force (32mm width), I would not use it over 100. So, with your 300 lb payload, you'll want to limit it to .3g acceleration/deceleration. At 100 lb, any one spot will be good for approximately 1.5-2 million cycles (start and stop on this exact point) before tooth failure occurs, and the steel reinforcement shortly thereafter. Plenty of life for a random motion. we test it as a bang-bang actuator, which is worst case. Life is a non-linear thing, so for comparison's sake, a 25 lb acceleration force will do 30 million hits on the same spot.
Speed will not be an issue, since your rotary to linear conversion is 75mm/motor rev. Where you may get in trouble is inertia matching your motor to the reflected inertia of the mass in motion. You can push the limits on that with ServoBelt. With great amps, 200:1 will work, decent amps, 100:1 (at a guess).

Mike,
Really nice drive you have there. I think I'm going to take a crack at a rotary version. Think very poor man's harmonic drive.
You mentioned a lubricant on the belts, what do you sugguest?
Thanks,
David Carr

Best lubes are spray teflon with wax, and food-grade chain lube (mineral oil, I think).
Good luck with a poor man's harmonic. Been chasing that one for a long while!

I just wanted to put up a little photo of how basic it gets on a linear. This is the smaller version with 15mm bearing rails, meant for nema 23 size motors only.

Attached


IMAGE(http://factorydaily.com/fdattachs/fdattachs6/112617455111712.jpg)


IMAGE(http://factorydaily.com/fdattachs/fdattachs6/112617455111713.jpg)

I just wanted to put up a little photo of how basic it gets on a linear. This is the smaller version with 15mm bearing rails, meant for nema 23 size motors only.

Oh that is nice!
bob

Hello Mike,
Any closer to a diy solution? Another question that hopefully you can answer. What about wood chips and dust? Would these type drive systems need covers or will the engaging belts and a brush keep them clean enough?
Mike

This is my first post on CNC Zone, and I'm glad to see that there is so much activity, particularly from guys like Mike Everman. I remember the first time I used the LowBoys and KAOS configurations and was impressed by the performance, quality, and ingenuity.
I'm here looking for possibly-economical ways to build a FDM machine. The servo belt approach is really cool, and I'm hoping I can use something like that for my little hobby project. I'll have to keep up with all of the activity here -- you guys are making really great stuff.

Hello Mike,
Any closer to a diy solution? Another question that hopefully you can answer. What about wood chips and dust? Would these type drive systems need covers or will the engaging belts and a brush keep them clean enough?
Mike

Sorry, still trying to get this in enough OEM hands to get to play with DIY on it. I'll make sure I do! I'm trying to show how simple it is with my pics, and help anyone that wants to try it, at least the intrepid souls that can machine the parts. There's nothing much you can't work out on a napkin.
As to debris, not tested beyond my initial test that shows brushes do not work. A little air jet does, though.

This is my first post on CNC Zone, and I'm glad to see that there is so much activity, particularly from guys like Mike Everman. I remember the first time I used the LowBoys and KAOS configurations and was impressed by the performance, quality, and ingenuity.
I'm here looking for possibly-economical ways to build a FDM machine. The servo belt approach is really cool, and I'm hoping I can use something like that for my little hobby project. I'll have to keep up with all of the activity here -- you guys are making really great stuff.

Please forgive my failing memory! Those who've used both are extremely rare. Were you at mdyn? Thanks for the kind words.

Add new comment

Images
More information
  • Files must be less than 2 MB.
  • Allowed file types: png gif jpg jpeg.
Documents
More information
  • Files must be less than 2 MB.
  • Allowed file types: zip rar.