Need help brainstorming a solution for this mill turn spindle

[Jim Dawson]

Getting back to a clutch/brake system, I found something interesting yesterday. I'm doing some service work on a John Deere riding lawn mower for a friend and it has an electromagnetic clutch/brake for the blade drive attached directly to the motor shaft. This mower has a 22hp Briggs on it, so that means that this clutch is pretty beefy. It might be adaptable to machine tool use, could be interesting to look at. This one appeared to have 3 wires going to the coil(s), unlike an automotive air conditioner clutch that has two wires, so that might indicate that you would have an option of applying the clutch or brake independently. I haven't really looked at it too closely yet.

https://xtremeope.com/page/search-re...words=AM122969

[QuinnSjoblom]

Getting back to a clutch/brake system, I found something interesting yesterday. I'm doing some service work on a John Deere riding lawn mower for a friend and it has an electromagnetic clutch/brake for the blade drive attached directly to the motor shaft. This mower has a 22hp Briggs on it, so that means that this clutch is pretty beefy. It might be adaptable to machine tool use, could be interesting to look at. This one appeared to have 3 wires going to the coil(s), unlike an automotive air conditioner clutch that has two wires, so that might indicate that you would have an option of applying the clutch or brake independently. I haven't really looked at it too closely yet.

https://xtremeope.com/page/search-re...words=AM122969

That's pretty interesting. So it has a bore for an input shaft and the clutch locks the pulley to it? With a second option of locking both the shaft and pulley to the main assembly? I did a lot of looking around when I first started this project and thought I was gonna run dual motor for turn/mill, but eventually decided to avoid that whole mess and just throw a huge servo at it lol. But at this point it's looking like I probably will add a brake.

So anyway, I finished the mill turn. got it all wired up and started testing. Things are never quite as simple as you hope they will be. Couple issues. First of all, the gt3 8m profile has more backlash than I expected. At a radius of 3 inches, it's about 10 thou. Not so good. Not your typical backlash though. Up to a certain amount of torque the belt has enough grip to allow no backlash at all, but when applying enough load to the spindle, the belt "slips" and the teeth bottom out in other direction, so this would be very hard to compensate for. After more digging, indeed the 8m profile is not as good for positioning accuracy and 5m is much better for this. Second issue, this is my first time dealing with servos so I'm learning a lot about how they operate. With average gain settings, the servo felt very spongy when grabbing onto the 5 inch pulley attached to it and applying load by hand. I could push it off position a couple degrees before it would correct and pull back to center. Running the auto tune cycle still came up with numbers that were too low to get a rigid feel. After increasing main gain and I gain up to about 75% of max, it was starting to feel pretty rigid, but can still push the pulley off position a small amount before it corrects. Increasing I gain even further helps but eventually it gets into a bad oscillation. A mill turn spindle isn't exactly a common thing for a servo to deal with. It has to hold a rock steady position while loads are being applied. Usually servos are introduced to load as they are commanded to move which is much easier for a pid loop to deal with. This is the opposite. When holding a steady position and applying load to it, the pid loop needs to see error and correct for it. It's also hard because of the lack of gearing. In my case, a single encoder count is about a half a thou of movement at a 5 inch radius. With a servo on a ballscrew, a single encoder count is 50 to 100 times less linear movement. Much easier to hold a position in that case.

So long story short, I think I'll most likely be dropping down to a 5mgt profile belt to reduce backlash and add the brake. Had to see for myself if this could be done without a brake, but it's starting to look like most likely not if I want best surface finish possible. Luckily the existing design will work with the new pulleys and it has room for the brake. They will also use the same 2012 taper lock bushings. Might even be able to return the pulleys I have.

All that being said, still very impressive to see what this mill turn can do. Wrote a program testing out my spindle swap macro. First had it do some indexing as a 4th (at insane speed lol) then ran the swap macro, had it spin up to 3k as a spindle, stop, swap back to 4th axis, do some more indexing. Very cool

Oh and I found something else really cool about this servo. It has an absolute zero position input. You can wire it up to give it an input signal that tells it to go to encoder zero. So after running as a spindle, i can call that signal in my spindle reset swap to automatically re home the 4th axis to an accurate position. No need for homing sensor

[Jim Dawson]

Very cool!

Is a video possible?

I'm really happy you are doing the R&D for this project. it's nice to know what works and what doesn't, it's going to make my life easier when I get around to doing mine.

[handlewanker]

Hi, the biggest problem with a servo motor solution is that when it's unloaded or at idle it draws little current and only ramps up the current draw when it detects a movement so that it moves the opposite way to counteract it......by that time the tool has moved on and the toolpath is wiggly...….just my opinion.

Having read all 123 posts on the subject on this thread it appears that the solution to locked status for milling and loose status for turning is still a long way off.

For a 4th axis I'll still vote for the worm drive with a spring loaded worm as it can be very finely resolute and/or locked while moveable.....you just can't back drive a 1:60....whatever..... reduction worm drive......IE no backlash.....but you can with a belt drive no matter how tight you pull them and worm drive will also give you a very high reduction ratio if you need it.

If it's a turning 4th axis drive that's needed then the worm can be lifted out of engagement and the worm wheel/spindle is free to rotate at any speed you want...….a separate motor would do the trick to rotate the actual spindle......6.000 rpm + is easily attainable.

It does mean that at the end of a milling program when a turning function is needed you would pause the program and manually disengage the worm drive.....you would of course lose position once the milling was finished and the worm disengaged so it would need to either mill or turn first.

I would think that the drive motor for turning could be permanently attached to the end of the spindle as it would just freewheel when the worm was doing the driving while milling, but that would be a neat solution to positioning the motor directly in line with the spindle......it would mean sacrificing the through hole capability for long work pieces, but a belt drive to the side would solve that.

In my case I would use a 1/4 HP 3 phase 3,000 rpm motor direct drive as it can be speed controlled easily with a VFD and cost's under A$100.
Ian.

[Jim Dawson]

For a 4th axis I'll still vote for the worm drive with a spring loaded worm as it can be very finely resolute and/or locked while moveable.....you just can't back drive a 1:60....whatever..... reduction worm drive......IE no backlash.....but you can with a belt drive no matter how tight you pull them and worm drive will also give you a very high reduction ratio if you need it.

If it's a turning 4th axis drive that's needed then the worm can be lifted out of engagement and the worm wheel/spindle is free to rotate at any speed you want...….a separate motor would do the trick to rotate the actual spindle......6.000 rpm + is easily attainable.

Ian.

Ian, I really like that idea. Simple to implement and should be very effective.

[QuinnSjoblom]

I believe the servo/belt arrangement can work really well, it just needs the brake. As you said, it's difficult to get a servo to hold a rigid position because it's waiting for error and correcting. The belt isn't helping with the rigidity problem either. I think once I drop down to the 5mgt profile and add the brake, I'm pretty confident it will be very accurate and plenty rigid. The belt will have some backlash no matter what, but I think if all turning and indexing is in one direction, it should be capable of great accuracy.
Even the way I have it now isn't terrible, I'm just looking for worst case scenario. The backlash measurements i took were at a radius of 6 inches. A lot of my parts will be 1 inch diameter and less. About a thou of backlash in that case. Also at this diameter the servo wiggle should be pretty minimal. So for parts that size, i think it would work ok as is with no brake at all, but if I can improve on it, i will.

I think before i change anything at all, I'll go ahead and try some index milling just to see. I'm just waiting on my 5c drawtube which arrives Tuesday, then the real testing will begin. After that the next step will be switching to 5mgt. I'll also try that without a brake to see how it looks.

The way you explained the servo being at low amperage when stationary kind of makes me wonder something. If there was some way to put some rotational preload on the spindle, the servo would already be at a higher amperage and holding position with some extra power. Would probably respond much better to a tool coming in contact with the work. Not sure how you could give that preload though.
I think the real solution is just a much more expensive servo with higher count encoder that can hold position better. Many 5 axis machines have a direct drive servo rotating table and it obviously does a great job holding position and responding to load without a brake. Just comes down to good enough hardware.
If the belt was the limiting factor and the servo could hold well enough, I would actually direct drive it by coupling the servo to the spindle. It would be a bit tricky to keep the 5c drawtube functional, but not really that complicated. Downside would be being limited to shorter stock since servo is at the back. But since the servo also has some rigidity problems, I think the best solution is to get as good as I can with the belt and add a brake. The direct coupling would be nice since it would really improve accuracy and eliminate backlash completely, but I don't think you would be able to combine it with a brake to address the servo rigidity problem. If the brake locks it even 1 encoder count off of where it wants to be, it's gonna overload and fault. Would need some kind of dampening in the coupler. In the case of the belt, there's enough give to allow a brake on the spindle without faulting the servo

[spumco]

Regarding the belt backlash, some pulley mfgrs can supply zero-backlash pulleys. One way is to anodize the pulley a bit heavier than usual to increase the tooth height and narrow the distance between the teeth. I believe on a HTD/GT2/GT3 tooth shape that the tip of the belt tooth hits the bottom of the pulley groove slightly before the sides hit. Much less for GT2/3 of course than HTD.

Making the pulley teeth bigger means the belt profile is wedged in the pulley and meshes on the sides before the tip hits. Brecoflex has a pretty good explanation of this. Downside is more noise and faster belt/pulley wear.

You might contact Brecoflex and see if they can supply a zero-backlash 8mm pulley set - won't help with tooth flex and you'll still want to think about a brake - but should help with positional accuracy.

The electric PTO clutches I've replaced on tractors don't have a through-hole - they go on the end of a tapered shaft. You'd probably have to removed the factory pulley (drill rivets), drill it out somehow to fit yoru spindle, and then attach your GT2 pulley and make sure it's concentric.

And 'real' electric shaft brakes from Warner are ungodly expensive.

I'm still thinking a pocket bike disc brake with an air-powered caliper sandwiched between the spindle pulley and whatever spindle thrust nut/collar you have would be the ticket. In my case I have a live spindle inside my tailstock housing, so a simple hinged clamping collar with some friction pad material should do the trick.

[handlewanker]

If it's a brake solution that's needed all you really need is a friction pad against a largish diam drum.....chuck body perhaps......then a solenoid to apply the push at the right moment to do the holding...….maybe a band brake etc.

What you are trying to do is mainly for a milling situation to prevent the cutter from pulling/pushing the work piece around radially.....it gets tricky if you want to mill a spiral as the spindle must be mobile all the time.....but only in one direction.

In that case a one way type clutch would prevent any reversal...…..it's the bounce back from the cutter force that will do the damage.

I think if the work piece is rotating clockwise, looking at the chuck, and the clutch is a one way thing you need to make sure the program does not reverse the spindle during the machining op.

I don't think this can be a complete solution as some parts would need a radial back and forth action to produce the shape required......but if the program was written in such a way that you only went one way...clockwise whatever...... it could work.

Not really my cup of tea as a one way clutch does become a fixture that can be a problem to disengage or take out of the way.

I did a test on my very old 6" rotary table to see if the variable worm depthing can be a solution......depthing of the worm is by the normal method of a cam housing that moves the worm deeper into the wormwheel.

This does work to give you a backlash free solution …...PROVIDED...…..you hold the knob that adjusts the worm depthing and manually move it back and forth to cater for worm wheel wear IE tight and loose spots etc.....not at all practical.

As the adjustment is a rotary movement it cannot be automatically adjusted to cater for the highs and lows etc.....the worm needs to be swung in and out of engagement, maybe only a couple of thou or so, hence my advocating the worm in a separate swing housing that can be radially spring loaded to maintain full tooth to worm contact automatically and ride up and down on the worm wheel......being a worm to worm wheel drive it has an automatic braking function by design for the milling work.

Anyway, that's how I would/am going to make the drive when I get a "round tuit"...….I'm deeply involved in tiling my bathroom at the moment so another iron in the fire is not forthcoming too soon.
Ian.

[QuinnSjoblom]

Thanks guys, lots of good ideas. That's really interesting about anodizing pulleys for tighter fit. I'll look into that. If I switch to 5mgt profile narrower belt, it wouldn't be that expensive to swap the belt now and then from extra wear caused by slightly tighter teeth. Worth losing the backlash for sure.
After thinking a bit more and getting a better understanding of the gains, I'm seriously considering putting the servo behind the spindle with a direct coupling. Obvious advantages are removing all belt variables. Perfect accuracy, zero backlash, rigidity limited to only the servo. My first thought was that this would not allow the use of a brake since it would fault the servo if locking it even slightly off the location the servo wants to be at, but after playing with gains, I found that using a low integral gain gives it a bit of give. You can apply torque and the servo will allow itself to be pulled slightly off position without fighting back with all its power. It still pulls to an absolute precise position, but with a bit of a spring affect. This would allow the use of the brake even with a rigid coupling. I would probably build my coupling to allow the use of a large ballscrew coupler to allow for very slight axial misalignment.
So pros and cons to this, best precision possible, zero backlash, less vibration with turning since no pulleys and belts. Cons would be it's a pretty long assembly, about 20 inches total. This would probably be an issue for a standard layout moving table mill, but perfectly fine for my gantry layout. Another complication is the 5c drawtube. I have some ideas on how to keep the drawtube functional with the coupling, but might still be hard to deal with. Tightening and loosening won't be hard, but turning the tube many turns to change collets could be time consuming unless I come up with something better than what I have in mind for this. Also gonna be limited to about 12 inch bar stock since i cant pass it right out the back, but not a huge limitation for the size of parts and quantity I would be doing. I'm gonna model up what I'm thinking for the coupler/draw tube assembly and see what you guys think, see if it can be done differently to make dealing with the draw tube more convenient. I'll probably model something up tonight and post it for you guys to look at.

Oh and one more advantage to the direct coupling, it gives me plenty of room for the brake disk and caliper. With the belt setup it's all very crowded and would probably require a custom designed caliper to fit. With the the direct coupling, I have plenty of room to use just about any pocket bike brake assembly and just build a mount for it

[mactec54]

Thanks guys, lots of good ideas. That's really interesting about anodizing pulleys for tighter fit. I'll look into that. If I switch to 5mgt profile narrower belt, it wouldn't be that expensive to swap the belt now and then from extra wear caused by slightly tighter teeth. Worth losing the backlash for sure.
After thinking a bit more and getting a better understanding of the gains, I'm seriously considering putting the servo behind the spindle with a direct coupling. Obvious advantages are removing all belt variables. Perfect accuracy, zero backlash, rigidity limited to only the servo. My first thought was that this would not allow the use of a brake since it would fault the servo if locking it even slightly off the location the servo wants to be at, but after playing with gains, I found that using a low integral gain gives it a bit of give. You can apply torque and the servo will allow itself to be pulled slightly off position without fighting back with all its power. It still pulls to an absolute precise position, but with a bit of a spring affect. This would allow the use of the brake even with a rigid coupling. I would probably build my coupling to allow the use of a large ballscrew coupler to allow for very slight axial misalignment.
So pros and cons to this, best precision possible, zero backlash, less vibration with turning since no pulleys and belts. Cons would be it's a pretty long assembly, about 20 inches total. This would probably be an issue for a standard layout moving table mill, but perfectly fine for my gantry layout. Another complication is the 5c drawtube. I have some ideas on how to keep the drawtube functional with the coupling, but might still be hard to deal with. Tightening and loosening won't be hard, but turning the tube many turns to change collets could be time consuming unless I come up with something better than what I have in mind for this. Also gonna be limited to about 12 inch bar stock since i cant pass it right out the back, but not a huge limitation for the size of parts and quantity I would be doing. I'm gonna model up what I'm thinking for the coupler/draw tube assembly and see what you guys think, see if it can be done differently to make dealing with the draw tube more convenient. I'll probably model something up tonight and post it for you guys to look at.

Oh and one more advantage to the direct coupling, it gives me plenty of room for the brake disk and caliper. With the belt setup it's all very crowded and would probably require a custom designed caliper to fit. With the the direct coupling, I have plenty of room to use just about any pocket bike brake assembly and just build a mount for it

Here is a custom Brake I made for a Bridgeport this is all you will need it has ( 2 ) 1.5 Dia Piston's and has return springs so there is no rubbing when at speed, 100 PSI air pressure locks up the disc, if you had of installed the 2 Belt idlers that would of removed the backlash that is why they use them

You can also experiment by just putting a drag on the spindle shaft, just a piece of plastic with adjustable pressure onto the spindle shaft that will get your servo sorted out, the big problem you have is you have no load on the spindle and that big servo can't control such a small load

[QuinnSjoblom]

Here is a custom Brake I made for a Bridgeport this is all you will need it has ( 2 ) 1.5 Dia Piston's and has return springs so there is no rubbing when at speed, 100 PSI air pressure locks up the disc, if you had of installed the 2 Belt idlers that would of removed the backlash that is why they use them

You can also experiment by just putting a drag on the spindle shaft, just a piece of plastic with adjustable pressure onto the spindle shaft that will get your servo sorted out, the big problem you have is you have no load on the spindle and that big servo can't control such a small load

How would idlers change the backlash? The width of the tooth is narrower than the pocket it fits into on the pulley. Not sure how idlers would change that.

Brake assembly looks great. It would definitely be nice to have no pad contact when it's disengaged. Do you have any pics of the inside? I'm curious how it's built. I can kind of picture how it would work but how are the pistons sealed? Are there orings in grooves or something?

The plastic peice preloading the spindle sounds like a good idea, but what about when running at speed? Is thatbonly applicable if it's just a 4th axis? I would imagine the plastic would melt running at 3k rpm. I guess just tighten it down for 4th axis stuff and loosen for turning?

[mactec54]

[QUOTE=QuinnSjoblom;2283874]How would idlers change the backlash? The width of the tooth is narrower than the pocket it fits into on the pulley. Not sure how idlers would change that.[QUOTE]

You would have to check out Mike Evermans web site he has all the specs for his belt drives and he uses belts that have more backlash than what yours has, the 5mm pitch is best for Backlash though and I would still use 2 idlers for better control of the belt tension for what you are doing

[QUOTE=QuinnSjoblom;2283874]Brake assembly looks great. It would definitely be nice to have no pad contact when it's disengaged. Do you have any pics of the inside? I'm curious how it's built. I can kind of picture how it would work but how are the pistons sealed? Are there orings in grooves or something?[QUOTE]

Yes there are orings in this design, I do them for Hydraulic racing brakes as well so do show what and how I do them, that one is very basic

[QUOTE=QuinnSjoblom;2283874]The plastic peice preloading the spindle sounds like a good idea, but what about when running at speed? Is thatbonly applicable if it's just a 4th axis? I would imagine the plastic would melt running at 3k rpm. I guess just tighten it down for 4th axis stuff and loosen for turning?[QUOTE]

Yes just something to try to see how much load you can / need to apply when indexing only, without it making the servo fault, and no you would not use this for when you are turning, if you get creative this could be part of your brake that creates the loading then either lock after indexing and retract for turning

[spumco]

How would idlers change the backlash? The width of the tooth is narrower than the pocket it fits into on the pulley.

True, but if you have idlers you'll get more teeth in contact which should reduce the effective backlash. Increasing the number of teeth in contact by wrapping the belt amost completely around both pulleys means flex is reduced because the load is shared by more teeth - and the effect of the tooth gap is reduced because there's more belt-pulley friction (surface area) to keep it from slipping from one side of the tooth to the other.

Take a belt and engage one tooth (more or less) on a matching pulley - you'll see some backlash and maybe be able to flext the tooth a bit. Wrap it completely around the pulley and it feels rigid with no slip.

Dual idlers gets you maximum contact on the servo (small) pulley without violating the back-bending specs for a particular belt - in a shorter pulley center-to-center distance than a single idler. Belt mfgrs publish minimum idler diameters to ensure the belt doesn't wear out or separate from too small a radius bend. And from what I've read in the vatious belt specs, the back-bending radius is larger than the smallest drive pulley diameter.

You can do it with one idler, but because the idler has to be a fairly large diameter the servo and spindle pulleys have to be rather far apart to get the big idler down in between them to maximize teeth in contact. By this I mean the belt is almost completely wrapped around the servo pulley. With dual (opposing) idlers you don't need the pulleys to be spread as far apart to have the same number of teeth in contact.

For example the minimum pulley for an AT5 belt with back bending is 25T and a 60mm idler. I'm sure you can find similar spes for GT2/3-5M and 8M belts. Model it up and see how close you can get the belt to touching itself withot violating the backbending pulley specs. Check the center to center distance and do it again with two opposing idlers. I suspect you'll find that the center distance will be more favorable for similar number of teeth in engagement.

On the other hand, your physical layout may dictate that a single idler is really a better compromise.

I suggest that before you pitch the current arrangement get an idler and a longer belt and wrap the belt as far around the servo & spindle pulleys as far as it can go without the belt teeth touching each other. Then check your backlash and flex and see if it'll do for your purposes. Shoudl be much less work as a test than completely re-doing it for a direct drive.

And you can still put a brake on the back of the spindle pulley and clamp it with a caliper.

[handlewanker]

For anyone who contemplates using "flexible", or probably more true, resilient belt applications you will still get elongation when you apply a force to a length of elastic no matter how tight you pull it.

It's also very true that if the elastic medium is gross enough you can apply more force to overcome the force that attempts to stretch it even more...…...eventually the elastic medium will tighten enough to overcome the force that wants to stretch it

The force that attempts to stretch the elastic more is the cutter digging into metal in a radial direction and the bounce back when the cutter is out of the material.

The only flexible medium for transmission without elongation under load is a chain and you don't need to apply tremendous force to make the chain taut.....it gets taut with very little end force......and chains can run at high speed if you take for example the timing chains used on some engines.

Of course, the distance between chain sprockets should be as close as possible to prevent the chain from sagging in the centre.

A chain drive would only need to be used for a milling application to prevent backlash and the relatively small tension for chain tautness is kinder to the bearings of the spindle and servo motor.

In spite of that, the servo motor still has to be powerful enough to hold position and it's still variable before the cutter applies a load and will still move before it goes back to position so it's no better than a taut belt drive.

The servo motor with a direct drive solution in this equation is the easy way out for a holding force, but in my opinion does not solve the problem as it's potentially more resilient than the best belt drive.

I read somewhere that a DC brake motor is a very rigid holding force without any servo motor characteristics as the DC current is constant at standstill to the breaking force needed.....IE, it doesn't increase if the shaft moves.
Ian.

[mactec54]

For anyone who contemplates using "flexible", or probably more true, resilient belt applications you will still get elongation when you apply a force to a length of elastic no matter how tight you pull it.

It's also very true that if the elastic medium is gross enough you can apply more force to overcome the force that attempts to stretch it even more...…...eventually the elastic medium will tighten enough to overcome the force that wants to stretch it

The force that attempts to stretch the elastic more is the cutter digging into metal in a radial direction and the bounce back when the cutter is out of the material.

The only flexible medium for transmission without elongation under load is a chain and you don't need to apply tremendous force to make the chain taut.....it gets taut with very little end force......and chains can run at high speed if you take for example the timing chains used on some engines.

Of course, the distance between chain sprockets should be as close as possible to prevent the chain from sagging in the centre.

A chain drive would only need to be used for a milling application to prevent backlash and the relatively small tension for chain tautness is kinder to the bearings of the spindle and servo motor.

In spite of that, the servo motor still has to be powerful enough to hold position and it's still variable before the cutter applies a load and will still move before it goes back to position so it's no better than a taut belt drive.

The servo motor with a direct drive solution in this equation is the easy way out for a holding force, but in my opinion does not solve the problem as it's potentially more resilient than the best belt drive.

I read somewhere that a DC brake motor is a very rigid holding force without any servo motor characteristics as the DC current is constant at standstill to the breaking force needed.....IE, it doesn't increase if the shaft moves.
Ian.

Where in the world did you come up with that chain idea, I think you should rethink that a chain would have 20 plus more times stretch than a GT series belt, belts drives have been used on CNC machines for more than 40 years and some of those machines can hold .0001" tolerance on the cut parts, .0005" is the normal average for direct drive CNC machines

[QuinnSjoblom]

How would idlers change the backlash? The width of the tooth is narrower than the pocket it fits into on the pulley.

True, but if you have idlers you'll get more teeth in contact which should reduce the effective backlash. Increasing the number of teeth in contact by wrapping the belt amost completely around both pulleys means flex is reduced because the load is shared by more teeth - and the effect of the tooth gap is reduced because there's more belt-pulley friction (surface area) to keep it from slipping from one side of the tooth to the other.

Take a belt and engage one tooth (more or less) on a matching pulley - you'll see some backlash and maybe be able to flext the tooth a bit. Wrap it completely around the pulley and it feels rigid with no slip.

Dual idlers gets you maximum contact on the servo (small) pulley without violating the back-bending specs for a particular belt - in a shorter pulley center-to-center distance than a single idler. Belt mfgrs publish minimum idler diameters to ensure the belt doesn't wear out or separate from too small a radius bend. And from what I've read in the vatious belt specs, the back-bending radius is larger than the smallest drive pulley diameter.

You can do it with one idler, but because the idler has to be a fairly large diameter the servo and spindle pulleys have to be rather far apart to get the big idler down in between them to maximize teeth in contact. By this I mean the belt is almost completely wrapped around the servo pulley. With dual (opposing) idlers you don't need the pulleys to be spread as far apart to have the same number of teeth in contact.

For example the minimum pulley for an AT5 belt with back bending is 25T and a 60mm idler. I'm sure you can find similar spes for GT2/3-5M and 8M belts. Model it up and see how close you can get the belt to touching itself withot violating the backbending pulley specs. Check the center to center distance and do it again with two opposing idlers. I suspect you'll find that the center distance will be more favorable for similar number of teeth in engagement.

On the other hand, your physical layout may dictate that a single idler is really a better compromise.

I suggest that before you pitch the current arrangement get an idler and a longer belt and wrap the belt as far around the servo & spindle pulleys as far as it can go without the belt teeth touching each other. Then check your backlash and flex and see if it'll do for your purposes. Shoudl be much less work as a test than completely re-doing it for a direct drive.

And you can still put a brake on the back of the spindle pulley and clamp it with a caliper.

So here's the reason why I don't think the idlers are going to significantly reduce my backlash. You are correct that getting more belt wrap will spread the load across more teeth and reduce tooth deflection but here's the thing, when I was doing my backlash tests, there was basically a hard stop when the teeth would bottom out in either direction. There was a very consistent amount of play when twisting the pulley back and forth. Once it was bottomed out it one direction or the other, applying more force gave basically zero movement. so in my case, tooth deflection is only a very small part of the equation, which i kind of guessed would be the case with my huge 50mm wide belt. You are also right that more wrap is going to increase the amount of torque needed before the belt "slips" and bottoms out teeth in the other direction, but the increase in wrap isn't going to increase that torque amount enough to keep it from slipping under machining loads. Right now I have 180 degrees of wrap around both pulleys. Dual idlers will bring it up to maybe 270 degrees, which means about 50% more torque needed to allow the slip, but not enough to prevent it under machining loads. Also I imagine that backlash will walk back and forth during indexing just from the movement of the belt around the pulleys. So since tooth deflection is already almost non existent with my huge belt, the idlers aren't going to reduce my backlash in a way that's helpful.

All that being said, if I drop down to 5mgt profile which is limited to 25mm wide, backlash will be less of a factor and tooth deflection will be more of a factor, so in that case I would say it would definitely be beneficial to add the dual idlers and I will do so if switching to that configuration, I just don't believe it's going to help significantly with the current 8mgt setup.

At this point I'm still thinking the direct coupling is going to be the best solution. I haven't modeled it up yet, but I thought of a good way to make the draw tube more manageable. If I don't do direct coupling, I'll do 5mgt with idlers. It's actually going to be pretty cheap to try out the direct coupling. I already ordered material and parts for about 50 bucks. I have to machine most of the coupler myself to fit around draw tube, but im going to use a big ballscrew coupler to connect to the servo. I can get the axial alignment pretty good, but I imagine it's still a good idea to not have a completely rigid connection. It's also gonna make it more likely that the servo won't have problems when locking the spindle.

Quote "The servo motor with a direct drive solution in this equation is the easy way out for a holding force, but in my opinion does not solve the problem as it's potentially more resilient than the best belt drive."

Handlewanker, can you elaborate a bit on what you mean by this? Not sure what you mean by it being potentially more "resilient". Whay are the reasons you beleive the direvt coupling wont be as good as a good belt drive? Personally I only see advantages other than limiting the length of stock I can use to about a foot. This eliminates all problems associated with the belt like backlash, tooth deflection, etc. Any specific reason why you don't think it's a good solution?

Anyone else with an opinion on the direct coupling? It's very cheap for me to try

[QuinnSjoblom]

so this is what im kind of thinking for the direct coupling setup. I dont have the supporting structures modeled yet, but this is a basic idea of the mechanics. The wheel on the draw tube is accessible enough to spin it pretty quickly by hand to change collets. It has the added benefit of backing up against the back side of the coupling mechanism so collets get ejected when loosening the drawtube. It will have some knurling as well as holes for a torquing bar. The part that attaches to the servo is a large ball screw spider type coupler to allow for minor axial misalignment. Im sure i can get the spindle and servo running within a few tenths runout, but i think even that much would be hard on bearings without the special coupler. Disk for brake is shown, but not the caliper. I think the combination of the ballscrew coupler and a low integral gain will allow the spindle to be locked with a brake without the servo faulting. if its still too rigid of a connection and the servo faults when locking the spindle, i can get creative with the coupling plates to allow a bit of dampening. In the last pic, the exploded view, the draw tube is below at the bottom, The wheel of it is what is sitting in the middle of the assembly in the other 2 pics. The main hub part that clamps onto spindle is designed symmetric with a bolt at the bottom just like the top for balance. I might try to come up with something to replace the ballscrew coupler that doesnt take up as much length. I could lose a couple inches in length if i come up with a different way to make that connection and give some dampening/ allowance for axial misalignment. Feel free to point out any faults with this design. Still just a rough draft

[handlewanker]

So here's the reason why I don't think the idlers are going to significantly reduce my backlash. You are correct that getting more belt wrap will spread the load across more teeth and reduce tooth deflection but here's the thing, when I was doing my backlash tests, there was basically a hard stop when the teeth would bottom out in either direction. There was a very consistent amount of play when twisting the pulley back and forth. Once it was bottomed out it one direction or the other, applying more force gave basically zero movement. so in my case, tooth deflection is only a very small part of the equation, which i kind of guessed would be the case with my huge 50mm wide belt. You are also right that more wrap is going to increase the amount of torque needed before the belt "slips" and bottoms out teeth in the other direction, but the increase in wrap isn't going to increase that torque amount enough to keep it from slipping under machining loads. Right now I have 180 degrees of wrap around both pulleys. Dual idlers will bring it up to maybe 270 degrees, which means about 50% more torque needed to allow the slip, but not enough to prevent it under machining loads. Also I imagine that backlash will walk back and forth during indexing just from the movement of the belt around the pulleys. So since tooth deflection is already almost non existent with my huge belt, the idlers aren't going to reduce my backlash in a way that's helpful.

All that being said, if I drop down to 5mgt profile which is limited to 25mm wide, backlash will be less of a factor and tooth deflection will be more of a factor, so in that case I would say it would definitely be beneficial to add the dual idlers and I will do so if switching to that configuration, I just don't believe it's going to help significantly with the current 8mgt setup.

At this point I'm still thinking the direct coupling is going to be the best solution. I haven't modeled it up yet, but I thought of a good way to make the draw tube more manageable. If I don't do direct coupling, I'll do 5mgt with idlers. It's actually going to be pretty cheap to try out the direct coupling. I already ordered material and parts for about 50 bucks. I have to machine most of the coupler myself to fit around draw tube, but im going to use a big ballscrew coupler to connect to the servo. I can get the axial alignment pretty good, but I imagine it's still a good idea to not have a completely rigid connection. It's also gonna make it more likely that the servo won't have problems when locking the spindle.

Quote "The servo motor with a direct drive solution in this equation is the easy way out for a holding force, but in my opinion does not solve the problem as it's potentially more resilient than the best belt drive."

Handlewanker, can you elaborate a bit on what you mean by this? Not sure what you mean by it being potentially more "resilient". Whay are the reasons you beleive the direvt coupling wont be as good as a good belt drive? Personally I only see advantages other than limiting the length of stock I can use to about a foot. This eliminates all problems associated with the belt like backlash, tooth deflection, etc. Any specific reason why you don't think it's a good solution?

Anyone else with an opinion on the direct coupling? It's very cheap for me to try

Hi, at best the servo solution is a holding force that relies on the "elastic" magnetic flux doing the holding...…...but it's a holding force that wants to be go back to the same position if you attempt to move it.

That means the servo will move and be forced back against it's holding force that at standstill is minimal but will only increase when it senses movement......but it does move first before correcting it..

The servo is not a rock solid wall that won't move like a dead stop...…...but it will, as designed, move back if you try to move it away from position......this means it will move off position under a load no matter how much mag flux is injected electrically, and then of course with more power being injected it will move back to position......that is a back lash effect you don't want.

I will be the first to say that if a servo can hold position under load it will be the ideal solution......if it doesn't move you don't have backlash.

On the belt drive topic that Old Mac has chimed in on, once again if the belt is tight enough with enough strength to resist stretch then that too is an ideal situation but not as elegant as an inline servo.

I suggested the chain drive as it is a more rigid drive than a belt but not requiring as much tension to stop it stretching more under load...…..I also think the sprocket profile in the gap needs to be deepened to prevent the roller from bottoming and so migrate back and forth......this will ensure the roller makes contact with either side of the sprocket tooth.

If you tension a belt to stop it stretching under load any increase in load will only apply more tension to the belt.

For simplicity I personally prefer to go the worm drive solution with a spring loaded worm to worm wheel contact as I think it has more chance on a smaller scale to prevent backlash under a milling cutter load and has an automatic brake function by it's design couped with the advantage that you can have a huge resolution if required.
Ian.

[QuinnSjoblom]

Anybody take a look at the model i posted for direct coupling? I'm ready to start machining parts. Looking for any input on possible design changes. I did find a way to reduce the length of the ballscrew coupler so that will be about an inch shorter

[Jim Dawson]

I would worry about the end plates wracking under torque.