[jalsina]

It depends on what you want to mount on it. You could place a rectangular plate on top of the two vertical plates and fasten it with vertical screws into the top surfaces of the vertical plates. If you countersink the screws, you would have a nice, horizontal surface that you can mount other things to. Some thought would be required to avoid interfering with assembly/disassembly/maintenance of the slider/belt drive unit.

[greybeard]

I was thinking of moving the plate which currently carries the motor even further in, so that it is virtually rubbing(but not quite!) along the belt.
This would mean a sub-assembly to carry the motor itself, but I was thinking of the tracking of the belt as the main problem, rather than the number of parts.

As to mounting the whole assembly, I suppose the first decision is which moves, gantry or work.
I, too, am going through the "how big do I need" rather than "how big to I want" or even "how much room have I got".
My current machine, still the Mark 1, though rebuilt many times, has a cutting area of 16" x 30" on a 22" x 36" bed, and this can just accommodate the dulcimers I'm now trying to make.
That's OK if I rip up the 8x4 ply into the sizes I need, but then have to do several operations to trim, and then clean up afterwards.
With an 8x4 + bed, I could think in terms of a single operation, drilling, milling and cutting out all part of a single programme.
If I did go to this size, I'd be forced to stick with a moving gantry, but if I stick to the sort of programme I'm following at the moment, cut and laminate the parts, then drill and mill, I would definitely go towards a moving table.

Sorry it was such a ramble, but that's me

[Zebethyal]

It depends on what you want to mount on it. You could place a rectangular plate on top of the two vertical plates and fasten it with vertical screws into the top surfaces of the vertical plates. If you countersink the screws, you would have a nice, horizontal surface that you can mount other things to. Some thought would be required to avoid interfering with assembly/disassembly/maintenance of the slider/belt drive unit.

Probably more a case of me trying to over engineer everything - I seem to have something against drilling and tapping aluminium, and prefer instead to go with through bolts, I have no idea why - possibly paranoid about stripping the threads.

@greybeard - I understand what you are saying, a multilayer design (similar to some of my earlier designs) or some other sub assembly would indeed be required - maintenance/belt replacement might be interesting though.

I also keep thinking - an 8x4 table would be great, but how often will I actually use it for whole sheets of MDF/Plywood?, 4x4 will probably be more than enough.

I am also aiming to build a machine with about a 350x510mm cut area mainly for PCBs and smaller items - fixed gantry design. Bigger machines always seem to take up so much extra room if designed as moving table (i.e. you need about an 8x4 machine for a 4x4 cutting area) yes it is a lot more rigid but at the cost of space (often a premium).

[Mauser]

I'm curious about where you get the numbers to decide that R&P is cheaper.

Instead of flanges, most of the drive parts stores have smooth pulleys with flanges on them which could be used for your idlers.

For "Attaching it to something", if one of the blue plates was the same part as the "Skate" underneath it, that's a pretty direct connection (But perhaps not as adjustable as you might like.)

[Zebethyal]

I'm curious about where you get the numbers to decide that R&P is cheaper.

Instead of flanges, most of the drive parts stores have smooth pulleys with flanges on them which could be used for your idlers.

For "Attaching it to something", if one of the blue plates was the same part as the "Skate" underneath it, that's a pretty direct connection (But perhaps not as adjustable as you might like.)

With regards costs, lets just consider a single drive on a single axis for the moment, assume that the axis is 2m long.

This is the Rack and pinion version being used for comparison (reverse engineered from the CNC router parts unit):

Ignore for the moment any items common to both systems - stepper motor, reduction gears and drive belt, long trucks and the flat steel they run on, as I would be using the same on both systems.

I need a plate of some sort to mount the reduction gears, I can buy a 244x244x6mm plate for about £8.00, or a 10mm plate of similar size for about £12.00 (I need a plate of about 200x100 for the R&P gearbox and 2 plates at about 170x90mm for the current belt drive version), this could pretty much be ignored as well as again required for both.

Oilite flanged bearings for the drive shaft (2 for the R&P, 4 for the current belt drive system) about £1.00 each.

I can buy a 1.5m length of rack for about £40.00, pinion gear about another £10.00 (rack does not need to run the full length of the table, just as far as the pinion needs contact with it for the length of movement in that axis, 2m length would be £50.00) and a couple of pounds for a turnbuckle and spring (not shown in the picture).

I am ignoring the clamps for the rack, as this could just as easily be bolted direct to the extrusion, and the belt system may yet still need guides (sort of evens out).

Drive gear for belt system: about £5.00 for 25mm, can't find a price at the moment for 32mm gears, but assume about £2.00 more.
Open lengths of belt can be had for the following prices:
2m of 25mm is £22.50, 2m of 32mm is £28.82
tractor belt 25x575mm £13.34, 32x575mm £17.08 (or a second belt of similar length to the static belt)
Skate bearings I can buy at about £1.00 for 10, so not really worth bringing to the argument, if I want additional flanges on the idlers, penny washers are a few pence in price.
M8 studding £0.70 for 1m, trade this off against any bolts in the R&P design.
Spacers are made from 3/4" aluminium round bar - about £9.00 for a meter, less than 250mm worth of spacers, so less than £1.00 here.
Some of the other designs use several plates of 10mm and some larger bearings (all potential additions to the cost of the belt version).

So at the end of the day it pretty much boils down to the cost of the rack and pinion gear £50,00 (10+40) or 60.00 for 2m (10+50) vs the drive gear and belts £40.48 (5+22.5+13.34) for 25mm tractor, 50 (5+22.5+22.5) for 25mm 2 x full length, £52.90 (7+28.82+17.08) for 32mm tractor, £64.64 (7+28.82+28.82) for 32mm 2 x full length.

Not a huge amount in it, but the belts will wear much faster than rack, and the Rack version is much simpler to make.

If you are wondering where I am getting my prices from - Belting Online, cheapest UK supplier I have found so far, rack and pinion prices are from Zapp Automation, you may know of cheaper suppliers closer to your location.

[JustPlaneSweet]

I have taken a slightly different approach to this drive. Instead of depending on the slide to maintain the 'pinch' between the idler rollers and the stationary belt I have used 5/8inch cam followers on a plate to pinch a stationary belt attached to one leg of an angle extrusion. Photos below. The photograph is showing the X-axis upside down on the bench. The THK rails and angle are attached to the the bottom of the sliding table.

I have a 6:1 reduction to the 15 tooth T5 drive pulley. Using a Gecko 251X drive with the 10:1 microstepping gives me 4064 steps per inch. My X-axis is giving me 650 inches per minute. I am using a 10mm wide T5 belt rated at 75 pounds tension.

The SolidWorks images show a step to retain the lower bearing. This was added after I made the prototype. The prototype has a plywood cover to retain the bearing. The module is machined from MDF and painted black.

Stan

[Mike Everman]

True to your handle, that is just plain sweet! Have you run it repeatedly yet? I am concerned that the belt is going to track out of the idlers. I've found that flanges on pinion and/or on the idlers did not control the tracking well (in our setup). Your backing idlers are a good bit.
This is an appropriate treatment. I've pitched the small circular belt several times, but for our product, it's all about cycle life, which leans in favor of a full length upper belt.
Also, be very mindful of runout at the pinion and main idlers. It makes a substantial sinusoidal error in motion. "Substantial" may not be in the range that concerns you, of course. We use big cheap dual row bearings instead of sleeving smaller ones to eliminate one source of runout.
Nice work!
Mike

I have taken a slightly different approach to this drive. Instead of depending on the slide to maintain the 'pinch' between the idler rollers and the stationary belt I have used 5/8inch cam followers on a plate to pinch a stationary belt attached to one leg of an angle extrusion. Photos below. The photograph is showing the X-axis upside down on the bench. The THK rails and angle are attached to the the bottom of the sliding table.

I have a 6:1 reduction to the 15 tooth T5 drive pulley. Using a Gecko 251X drive with the 10:1 microstepping gives me 4064 steps per inch. My X-axis is giving me 650 inches per minute. I am using a 10mm wide T5 belt rated at 75 pounds tension.

The SolidWorks images show a step to retain the lower bearing. This was added after I made the prototype. The prototype has a plywood cover to retain the bearing. The module is machined from MDF and painted black.

Stan

[JustPlaneSweet]

Thanks for the feedback Mike.

I have a full length spacer attached to the table right beside the angle (not shown) that keep the belt from riding 'up'. The cam follower plate keeps the belt from riding 'down'. These could be faced with UHMW for life if necessary.

Runout of the idlers and pinion is set by commercial tolerances for the cam followers and the prebored pulley. The aluminum treads were turned ID and OD in the same setup.

I have not had a chance to run it much yet. So far it looks pretty good.

Stan

[greybeard]

Stan, could you add a bit more to your description, as I'm having a bit of a problem seeing how the "cam followers" pinch the angle extrusion carrying the stationary belt against the idlers.
Perhaps it's just that I'm not familiar with the concept of cam followers, but if they are mounted on the same plate as the idlers(the darker brown plate in drawing 1 ?), how to they push the angle+stationary belt onto the moving belt+idlers ?

Regards
John

[JustPlaneSweet]

John,

The drive does not clamp the two belts together. I probably should not have said 'pinch'. It only needs to keep the belts in mesh. One of the earlier posts said to allow a slight gap between the belts and the idler wheels. The cam followers are arranged to allow the angle plus two meshed belts thickness between them.

When everything is in mesh the the light brown and dark brown plates are pulled apart with 20-30lbs force and locked down. The tension in the belts wants to pull the belts out of mesh but the two cam followers on the back side of the angle keep everything together.

Putting all the cam followers on the same block maximizes the rigidity of the setup. The clearance is taken out when the timing belt pinion tension is applied.

Stan

[greybeard]

Thanks Stan. All is now clear
John

[mipop35]

Sorry about the question, perhaps is repeated. it's a long thread and I couldn't get it.

How is the "rack" belt glued to its back rail?

Thanks,

Diego

[Mike Everman]

It is 3M VHB double back tape.

[kingkong111]

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!

[kingkong111]

[Mike Everman]

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.

[CNC43711]

Would the servobelt use Steel or Kevlar tension members in the belt?

[ger21]

Mike stated earlier that steel is the better choice.

[Low-Alloy]

Has anybody considered having the bottom, static, belt be mounted on the moving table and the motor and dynamic belt be mounted stationary to the frame? I don't know how well it work, but in some designs, like a large moving table, I would think it would be easier for speed reducing/torq increasing pulley setups and larger motors to be mounted to the machine frame instead of traveling back and forth.

What do you think?

[PaulRowntree]

Has anybody considered having the bottom, static, belt be mounted on the moving table and the motor and dynamic belt be mounted stationary to the frame? I don't know how well it work, but in some designs, like a large moving table, I would think it would be easier for speed reducing/torq increasing pulley setups and larger motors to be mounted to the machine frame instead of traveling back and forth.

What do you think?

I don't see that it would be a problem. The machine I am designing now will be built this way, for exactly that reason.