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

[Lester444]

Hello Folks.


Here's my attempt at this belt system. At the moment (Covid-19 days) GT2-6mm belts are all I have so it is what I used. I do know that GT2 does not mate to itself. Just hoping for a reasonable lifespan with what I have. I can live with slow feed rate, shallow DOC.


I will make a 5-ft length version for cutting plywood or plastics, single axis only for now. The green-colored cylinder is a small 700w router.

[junior2016]

This guy used the Bell Everman system on the MPCNC. I like his belt tensioning mechanism.

[TheWire]

Atlanta gears sells split helical pinions to remove backlash, so I don't think helical r&p by itself eliminates backlash.

Back on topic.

I got my belt in, and while I don't have anything to mount it to, it sure doesn't seem like it'll have any backlash. But I'll have to wait until I start throwing 50 lbs back and forth to see. Unfortunately, it's going to be a while before I'm able to test it.

Gerry,
This question for a post thats almost 10 years old, but can you tell me what belt profile you used and if you had long term satisfaction with the belt on belt system?

Thank you,
Jerry

[ger21]

I never built the machine, so I never actually used what I made.
I used the Belts that Mike used, which I think were AT5?

I decided to buy Helical Rack and pinions for the machine, using servos with gearboxes.

[spumco]

Bell-Everman uses T5 and T8. AT5 belts doesnt mesh with itself as the belt teeth are wider then the pulley teeth. (edited for clarity)

T3/5/8, HTD, GT all mesh with varying success. I've not tried anything but T5. There's a ton of backlash when T5 belt is simply meshed with T5 belt, but the Bell-Everman system preloads the drive belt teeth in both directions and you're left with a minimal amount of backlash between the pinion and the belt.

I'm using DMM 750w servos w/10:1 very low backlash (helical) planetary reducers and 32mm wide T5 belts. Gantry weighs around 75lbs, >6'x12' cutting area. Downdraft table.

Pinions are 16T to help belt life a little and the idler rollers are about 40mm dia. Pinions were made from T5 bar stock - cheaper to make our own than 3/ea custom width/bore pinions. Hard anodizing would probably help longevity and backlash, but this is a plasma cutter - not a Kern or Grob.

Using the "LoopTrack" configuration rather than a stationary top belt. We wanted side-mounted belts to keep plasma crap off of them. Rails and belt mounting surface(s) are mounted to aluminum extrusions, and the extrusions are bolted to the steel frame. Extrusions are shimmed out from the frame so the rails are very straight. With the drive belts off, you can move the gantry with one hand and no racking or binding (or not much).

Rack belt is bonded to milled-flat aluminum stock with VHT transfer tape. Also captive at ends with a belt tensioning system - we found that the servos and reducers had enough torque to start pulling the rack belt off the VHT tape and wound up with a few bubbles or wrinkles. We thought the tape was simply peeling up until we ran the head back and forth a few times really hard - more bubbles.

Tension clamps solved that one and helped straighten out the rack belt as the VHT tape settled in over time while the belt was tensioned.

If there's one thing we learned... the idler/roller flanges should be as big as possible without hitting the rack belt mounting surface. The looptrack drive belt really needs help staying on track and the only thing guiding it are the roller flanges. A slight bevel on the flange also helped.

Here it is in progress. Initial testing on Y (short) axis showed about 1400ipm max speed and fairly aggressive acceleration. Much more testing soon as the wiring is installed and commissioning is started.

Overall layout [note - the gantry uprights are so tall because the gantry can be raised (manually). We will be able to cut >24" high weldments or components and let the THC compensate for a non-perfectly level gantry beam]


X-axis transmission


Y axis transmission


Looptrack transmission guts


Y-axis mostly assembled for testing

[spumco]

Hello Folks.


Here's my attempt at this belt system. At the moment (Covid-19 days) GT2-6mm belts are all I have so it is what I used. I do know that GT2 does not mate to itself. Just hoping for a reasonable lifespan with what I have. I can live with slow feed rate, shallow DOC.


I will make a 5-ft length version for cutting plywood or plastics, single axis only for now. The green-colored cylinder is a small 700w router.

Nice job.

If I could make an observation without (hopefully) sounding like an armchair critic...

Your configuration places the drive belt on the rack belt without the pre-loading that the basic Bell-Everman system achieves. The GT2 belts may - probably do - have less backlash when belt-to-belt than T5. But the way you have the idler/rollers configured they're just keeping the drive belt flat against the rack. Tons of engagment length - good - but any slop between the drive teeth and rack teeth is still there.

On the other hand, the centrally-located Bell-Everman pinion, when tensioned away from the rack belt, preloads the drive belt teeth against the opposite flanks of the rack belt teeth - in both directions. I think this is how B-E achieves such low backlash with otherwise sloppy T5 belts.

I'm not poo-pooing your machine. Run the heck out of it and if it works for you then that's a win. But if you find yourself chasing backlash or jerky direction changes you might have a re-think about the drive belt configuration.

[spumco]

Thank you Mike Everman, for sharing your design so clowns like me can build neat stuff.

[JustPlaneSweet]

spumco,

If you look at the fourth image you can see four rollers. The belt is pushed in between the two middle rollers and is looped around a sprocket on the servomotor. The servomotor is pulled tight which seats the drive belt against both sides of the fixed belt. No backlash. The drive belt portion that loops around the servomotor is always tight. The drive belt portion that loops around the back is always loose.

Stan

[spumco]

spumco,

If you look at the fourth image you can see four rollers. The belt is pushed in between the two middle rollers and is looped around a sprocket on the servomotor. The servomotor is pulled tight which seats the drive belt against both sides of the fixed belt. No backlash. The drive belt portion that loops around the servomotor is always tight. The drive belt portion that loops around the back is always loose.

Stan

Stan,

I was referring to post #361 when I offered some suggestions on belt configuration/path. I think the 'fourth image' you're referring to above is my drive system, which follows the basic B-E belt path.

-Ralph

[johntorr]

I am thinking of using this type of belt system for commercial use and wondering if this design is protected by patents, a quick google search shows that that the patents for this type of system has either expired or abandoned as this type of system has been in use since the 80s ( link )https://patents.google.com/patent/US5819584A/en https://patents.google.com/patent/US20090301237A1/en. I am not sure if there is something I am missing or if I would infringe on someones patents if I use this in a commercial application
Thanks
John