NICE! scary fast and ingenious looking drive design.
Printable View
NICE! scary fast and ingenious looking drive design.
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.Quote:
Originally Posted by turmite
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?
Hi Mike,Quote:
Originally Posted by turmite
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.
Oh that is nice!Quote:
Originally Posted by Mike Everman
bob
Hello Mike,
Any closer to a diy solution?:D 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.
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.Quote:
Originally Posted by turmite
As to debris, not tested beyond my initial test that shows brushes do not work. A little air jet does, though.
Please forgive my failing memory! Those who've used both are extremely rare. Were you at mdyn? Thanks for the kind words.Quote:
Originally Posted by dmatsumoto
Mike is the engagement of the two belts good enough to mount this upside down on the underneath of the machine?
Mike
Quote:
Originally Posted by Mike Everman
Two ways to go there, the static belt is always forming the rack.
1. Make the dynamic belt a smaller circular belt, and the static one is then mostly uncovered. The circular belt manages itself, or can be managed with another idler above the pinion.
2. We have made a rolling tool that pinches the walls of the extrusion that holds the belts. A bit of interference works great and the dynamic belt peels off and is rolled back in with a little added friction as payment. This works well on a side surface, but vie not tried it under, but it should be fine.
dmatsumoto-
If you are looking to make an inexpensive FDM machine, take a look here:
MakerBot Industries - Robots That Make Things.
or here:
RepRapWiki
I have not tried either of these yet, but I hope to try a makerbot extruder head on my CNC Mill sometime.
David Carr-
I have been working on a team designing a highly advanced prosthetic arm, and it uses several harmonic drives. They are very cool to work with but insanely expensive. I think I know where you are going with the belt-style harmonic drive. If you have two belts, one glued onto the inside of a bore, and another belt turned inside out with two fewer teeth engaging the bonded in belt at two opposing points, you could theoretically make a poor-man's harmonic drive. Perhaps there are some clever tricks like what Mike has done to tension the belts where they engage the "rack" belt.
Keith
Thanks, Keith, I already have a MakerBot Cupcake and have an UP! printer on the way. I just wanted to dream about making my own for fun. I probably won't ever find time, but it's still entertaining to look at all of the great things people are putting together.Quote:
Originally Posted by NEATman
Hi Mike!
Well, I can report that I've implemented 3 axis of your servobelt and I'm loving it.
I've hit a little snag though and I was hoping you could help me out here as you probable have lots of experience with this.
Namely the two large rollers that are also tensioners, how to you calculate where they should be in relation to the pinion? I have drawn everything and even managed to CNC a few "plates", as I call them - it locates all bolts that carry those little rollers as well as holes for attaching to the stepper but I've only ever managed to have one setup that's slop-free, the other two are rather sloppy. I'm machining these plates out of MDF and the one performs rather nicely, but the other two are nothing to write home about. I've tried with slits instead of holes to have some adjustability but the tension on the belt quickly un-tensions the tensioners (again, the two larger rollers either side of pinion). I can send you a picture of what I've done and maybe you can spot the problem areas?
Or you can explain to me how you go about calculating the distance between pinion and rollers so that when the belt goes around the pinion and comes back down to the two rollers it meshes perfectly and without slop, nicely tensioned. Or is there another way? Move the whole stepper up and down to get the slop out by using slots to mount the plate to the stepper? But that would get undone rather quickly I imagine, just the same as with the slots for two big rollers.
Attached are three (!) images, one from the front and one from the back where you can see the two slits where big rollers attach to the plate. The third image will make it easier to understand the problem.
I shall spare Mike having to repeat his reply (from PM) so here it is for everyone:
The rollers should be fixed, hard mounted in a spot that touches but does not push down hard on the top belt. There's no calculation necessary when you mount the motor to slotted holes, though it is nice to know that you have room to move the motor up to tension the belt. We only use about 20 lb tension.
You need to understand that all linear forces you create translates into shear forces on the rollers and motor. That motor mounting especially needs to be super tight and secure.
Looking good so far! Now all that you need to do is insure that the plate is fixed with respect to the track, and slot your motor mount holes and pilot diameter, then pull up on the motor with <motor weight>+20 to 30 lb tension, and lock the motor.