Lead Screw to Timing Belt, many questions.

[p00ky]

Hi there,

A few years a go I built a cnc mill following some YouTube videos. I went to town on the electrics & skimped on everything else, Slowly but surely I have replaced a lot of the machine, Aluminium Right Angles with skateboard bearings turned into 20mm Precision Rails with matching Housed Bearings etc.
I have always run it on some rather cheap 10mm lead screws with brass nuts, but due to the size of the machine & my rather abusive machine tool ethic they are rather worse for wear and need replacing.
A friend who had seen some rather impressive videos of timing belts on cnc machines has suggested I consider them. I think he's obsessed with speed personally & I'm not sure he understands the limits of forcing a cutter through a block. However, considering their price I'm rather interested but a little lost as far as gearing etc. goes. I've spent many hours trawling through posts on this website and others, but I've ended up with more questions than answers.
So what I have:
3 x sy6osth86-3008BF Steppers (I think their Nema 23 460oz, 2000 steps per rev due to MicroStepper Drivers)
3 x M752 Lead shine Drivers
1 x PcPPS v3 Breakout Board

2 x X axis 1000mm 20mm Rails
2 x Y axis 800mm 20mm Rails
2 x Z axis 500mm 12mm Rails

3 x 10mm Lead Screw with Brass Nuts (slightly bent )

Dewalt Router (approx. 1000w)

Due to a fixed central gantry the cutting area isn't maximised, its approx. 450mm X, 450mm Y, 300mm Z.

What I'm looking at:
25mm T5 Timing Belt (2 x for X axis, 1 either side & then 1 for the Y axis & 1 for the Z).
Timing Pulleys (No. of Teeth & Size unknown?!? This is my main question really)

I'm considering going back to a moving gantry on the X axis to maximise the cutting area. Stealing the Z axis stepper as to have dual steppers driving the X axis, 1 either side (Z axis being replaced with a new stepper and driver). Or using 1 Stepper running both X axis Pulleys through a shaft.

I can see how to implement the pulleys but haven't got a clue about gearing, torque, speed, etc.

Any and all suggestions are welcome.

[p00ky]

Here's a video of the machine as it is right now.

Sorry for a long and tedious first post :P

http://www.youtube.com/watch?v=hR-rlOr1qVw]my diy cnc machine - YouTube

[harryn]

It is easy to be over critical of our own work, and miss out on what has been accomplished. I think in your case you should be quite proud of what you have completed. Of course anything can be made better, it is just a matter of approach.

The new rails are in fact better, but they are still unsupported, so there is a limit to how fast you want to drive this thing.

Belt drive systems can go really fast, but your distances are pretty short. The main challenge with belts is that almost no one uses a wide enough belt to really account for the forces and effects. A system like yours in a belt drive would need to be more like 75mm wide AT 5 to see any real benefit of speed gain vs lost accuracy due to stretch. I like belt drives, but using them for linear motion control is entirely different than using them to transfer a rotating motion.

There are some things which would make your system run a bit smoother and faster though, with not too many changes - change the screws, add stepper motor dampers, and a local table support.

At least in the video, the Y screw appears to be bent, and has to rotate fairly quickly to move the router. Changing to a higher pitch thread will help a lot.

Since your system is a fixed gantry, you can also improve the table support in the vicinity of the cutting to help with that aspect. Some slick plastic blocks touching the bottom of table where the router bit cuts will reduce local defection.

The stepper motors are running in resonance, which is a common challenge. Just do a search in the forum and there is a ton of info on resonance and using dampers to get rid of it. Just the use of these might double your useful speed.

[p00ky]

Hey there, thanks for your reply.
Speed increase isn't my over all objective although a little more while cutting accurately and finishing (while carving) would be nice.

75mm belt eh, not come across that in the UK (although I'm sure it exists).
Your right, dampening would probably be more important, I just hadn't thought of it. I know, I know, years in & I'm still a noob the story of my life. I shall certainly prioritise that.

Support under the cutter is also something I haven't really considered, would it be worth considering moving the long rails closer together as well in order to give greater support? They are as far apart as could be fitted at present.

Supported rail you say, certainly another thing to consider (after dampening, I'm looking a head now). Could my existing rails be altered (instead of being replaced) using bearing bocks that are open on one side (I've seen them certainly), perhaps milling a slot along the length and the welding in a support of some kind?

Anyways, thanks for the post, certainly some food for thought there.

[OCNC]

The new rails are in fact better, but they are still unsupported, so there is a limit to how fast you want to drive this thing.

Belt drive systems can go really fast, but your distances are pretty short.

It doesn't seem like you could easily overdrive this machine. The travel distances are short so generally it would take high accelerations to get things 'bouncing around'. Belt drive systems that go 'really fast' are generally servo driven. With steppers you aren't going to have that problem since their usable operating rpm range is much lower.

A system like yours in a belt drive would need to be more like 75mm wide AT 5 to see any real benefit of speed gain vs lost accuracy due to stretch. I like belt drives, but using them for linear motion control is entirely different than using them to transfer a rotating motion.

My machine uses 3/8" wide xl belts for the x and y axes (26" x22" cutting area). I can rapid at 400ipm and cut at 200ipm. Your comment about a 3" wide belt makes no sense to me.

Since your system is a fixed gantry, you can also improve the table support in the vicinity of the cutting to help with that aspect. Some slick plastic blocks touching the bottom of table where the router bit cuts will reduce local defection.

I think that all you need here is a thicker more dimensionally stable table. That piece of plywood shown in the video looks pretty thin and spongy.

The stepper motors are running in resonance, which is a common challenge. Just do a search in the forum and there is a ton of info on resonance and using dampers to get rid of it. Just the use of these might double your useful speed.

Microstepping drives are specifically intended to overcome the resonance issues generally present in full and half stepping drives. There shouldn't be any resonance issues here and I'm not sure I understand what it is you're noticing in the video that indicates that there are.



Looking at the machine my thought was that the best improvement would be new higher lead screws and matching nuts. Screws that will give you four turns per inch of travel will make a big difference in performance. Going to belt drive will involve a whole new configuration of drive components. Perhaps something to consider for the next machine.

Chris

[robe_uk]

Looking at the machine my thought was that the best improvement would be new higher lead screws and matching nuts. Screws that will give you four turns per inch of travel will make a big difference in performance. Going to belt drive will involve a whole new configuration of drive components. Perhaps something to consider for the next machine.

Chris

I agree with Chris and its probably the simplest upgrade you could do with minimum fuss. Have a look for ACME multistart screws

Rob

[harryn]

Chris - isn't your machine a screw drive ? I thought your belts were just for very short range connection to the screws ?

As far as resonance, I can hear it in the video, or at least it sounds like it.

Using the existing bearings and adding support to the rails - yes it is a good idea. Perhaps start with the X and just router some grooves in a block of wood for that round rail to sit on. It isn't easy to drill through hardened rod like that, and welding would probably distort it.

[p00ky]

Thanks again for you replies.

The spongy looking Bed is actually just a thin piece of ply on the top of a 1 inch thick piece of MDF. When the machine wasn't quite so well calibrated I would tend to overcut by 0.5mm when I was cutting something out, this is a habit I've never really got out of. I change the plate every so often. The video is probably a little misleading.

I have been looking into dampeners (mainly within this forum) today and found that people, even people with Micro Stepper Drivers (mine are micro steppers) recommend adding dampeners. Their an easy enough self build, one guy had made them using pennies as weights So I shall be giving that ago. It certainly wont detract from the performance & shouldn't cost me more than a few pennies .

The resonance is there trust me.

Replacing the higher Lead Screw is a must for sure & is what led me too consider Belts, but no one seems to be saying directly that I should. I have much less wobble on the bottom table Lead Screw as I used 2 Brass Nuts, 1 either side of the table.

All comments welcomed

[OCNC]

Chris - isn't your machine a screw drive ? I thought your belts were just for very short range connection to the screws ?

I have a Roton Hi-Lead screw directly driven by the stepper on the Z axis and XL belts on the x and y. The the rolling gear is based on a Momus with a few modifications to improve rigidity. I'm intending to further modify it by adding a slaved stepper to the second X-axis rail. I have no problems with resonance.

[p00ky]

Looks like a nice and simple setup, no gearing etc. Would be great to see a video of it in action.

[PaulRowntree]

I have a Roton Hi-Lead screw directly driven by the stepper on the Z axis and XL belts on the x and y. The the rolling gear is based on a Momus with a few modifications to improve rigidity. I'm intending to further modify it by adding a slaved stepper to the second X-axis rail. I have no problems with resonance.

Is there any sign of overshoot & bounce on X or Y at your working feed speed and acceleration?

[OCNC]

Is there any sign of overshoot & bounce on X or Y at your working feed speed and acceleration?

There is some unwanted motion. I'm reluctant at this point to give it a name only because I'm not certain yet what is causing it. It could be stretch in the belt, it could be overshoot in the stepper and it could be flexing of the gantry and it could be a combination of all or some of these. I have some new belting that is wider and has wire cords rather than fiberglass (T5 15mm). I'm going to replace that first to see what the difference is. I'm also going to upgrade the power supply from 24V to 36V. Adding a slaved stepper is the last thing I'll do. Keep in mind that while this machine is based on the Momus for it's rolling mechanism it's larger than the Momus and it's natural that it should suffer from scaling up without making any changes to the drive system. It works quite well for a wood router that isn't used to do small fine work.