The New Machine

[Web Goblin]

I would be very surprised if you can get a pulley with less than 10 teeth. 10 teeth is the least I have ever seen anyway.

Ian

[Jonathan]

I would advise against getting a small pulley as I heard the belt wear is greater. Also you may struggle to get enough teeth meshed on a very small pulley.

Only disadvantage I can think of with a bigger pulley is the moment of inertia is greater, but that's probably negligible on a mill.

[fragger6662000]

I would avoid belt drive, extra complexity is extra problems. They will also limit the maximum speed of the machine if you look at my convertion or the hoss machine you would be better to follow that template. The g540 is the way to go i have used some of the drives from the uk and they are not a patch on this setup. Keling was where i go my drive and motors from.

http://www.cnczone.com/forums/bencht...g_machine.html

[Ricardoco]

I would be very surprised if you can get a pulley with less than 10 teeth. 10 teeth is the least I have ever seen anyway.

Ian

Cheers Ian, Me too, In fact many of the sites I checked started at 12 teeth.

I would advise against getting a small pulley as I heard the belt wear is greater. Also you may struggle to get enough teeth meshed on a very small pulley.

Only disadvantage I can think of with a bigger pulley is the moment of inertia is greater, but that's probably negligible on a mill.

Thanks Jonathan, I did Intend to use a 10 and 25 tooth pulley to give me the ratio I needed, and going for anything smaller as you say may put too much load on the reduced number of meshed teeth. So 10 and 25 then.

So the pulleys will be as above.

For those people that are not aware of why I would like to use pulleys then let me explain.

:confused:Using 1605 ballscrews will give me a pitch of 5mm, that is to say 360 degrees in 1 rotation of spindle will advance the axis 5mm.

If i use a direct coupling stepper to screw,

The stepper will give full steps of 1.8 degrees.

360/1.8 will give me 200 steps per revolution.

5mm / 200 will give a resolution of 0.025mm per full step.

This is an odd measurement to use in my head, i have done it though..

Using a ratio of 2.5 to 1 ( 10 and 25 tooth pulleys) will give me a step resolution of 0.001mm per step.

and if i were to use half steps as the stepper is capable of doing, it would be 0.0005mm per step...

thats a fair bit of resolution/accuracy, there are however disadavantages to using pulleys.

its upto the individual who is designing/building and using the machine.

(a special thank you to i2i who educated me on this subject in the first place)

There you go clear as mud!!!!:confused:

Rick

[fragger6662000]

Cheers Ian, Me too, In fact many of the sites I checked started at 12 teeth.



Thanks Jonathan, I did Intend to use a 10 and 25 tooth pulley to give me the ratio I needed, and going for anything smaller as you say may put too much load on the reduced number of meshed teeth. So 10 and 25 then.

So the pulleys will be as above.

For those people that are not aware of why I would like to use pulleys then let me explain.

:confused:Using 1605 ballscrews will give me a pitch of 5mm, that is to say 360 degrees in 1 rotation of spindle will advance the axis 5mm.

If i use a direct coupling stepper to screw,

The stepper will give full steps of 1.8 degrees.

360/1.8 will give me 200 steps per revolution.

5mm / 200 will give a resolution of 0.025mm per full step.

This is an odd measurement to use in my head, i have done it though..

Using a ratio of 2.5 to 1 ( 10 and 25 tooth pulleys) will give me a step resolution of 0.001mm per step.

and if i were to use half steps as the stepper is capable of doing, it would be 0.0005mm per step...

thats a fair bit of resolution/accuracy, there are however disadavantages to using pulleys.

its upto the individual who is designing/building and using the machine.

(a special thank you to i2i who educated me on this subject in the first place)

There you go clear as mud!!!!:confused:

Rick

And hurry up moderator.

But you should be using 8 / 10 mirco stepping for best performance of the motors and why are you worried about the interger as it is a cnc'd machine that will be controlling it.

As before i have done it with this machine, 0.001 res is pie in the sky, £200,000 machine tool struggle to get that with direct drive and servo's. 0.025 per step is what i run and gives me 1500mm/min rapid you would be lucky to get 700mm/min as you would be over spinning the motor and have not torque.

resolution and power work againest each other.

here a like to my Thread on cnc zone.

http://www.cnczone.com/forums/bencht...g_machine.html

[Ricardoco]

And hurry up moderator.

But you should be using 8 / 10 mirco stepping for best performance of the motors and why are you worried about the interger as it is a cnc'd machine that will be controlling it.

As before i have done it with this machine, 0.001 res is pie in the sky, £200,000 machine tool struggle to get that with direct drive and servo's. 0.025 per step is what i run and gives me 1500mm/min rapid you would be lucky to get 700mm/min as you would be over spinning the motor and have not torque.

resolution and power work againest each other.

here a like to my Thread on cnc zone.

http://www.cnczone.com/forums/bencht...g_machine.html

Ok Hi There fragger, the steppers two x 23 & 1 x 34 from motionking, the 'integer' is for me, not the software, I design what would end up on the machine in the first place as i said "This is an odd measurement to use in my head" Now as for accuracy, wouldnt you say, that aiming for the highest accuracy you can in the begining is the correct thing to do.

I would never suggest I would get that resolution at the end of the project but if i allow the push for accuracy to become secondary to simplicity of build i would not be true to myself, and speed/feedrate well this is a hobby for me, im not in a hurry, im still one of those sad individuals that likes to sit there and watch, and i know im not the only one AM I???

Would i not get more torque with a 2.5:1 setup then?

I have read with interest the posts about this machine type on cnczone.

Pulleys are the way i would like to go as i dont like the positioning of the steppers in the other projects I have seen, they may be fine for you and make the conversion simple and speedy but that is not what im after, I have learned a lot from the conversions on cnczone but they just dont fit my requirements.

Stevie Nicks and her famous song told me all i needed to know lol

I do like the way you have mounted your x axis ballnut, that is simple effective and a touch inspired and look much more stable than the other ways i have seen.

I will follow your thread with interest and i may even use some of your ideas (giving credit of course), "if thats ok of course"

Rick
.

[fragger6662000]

No you don't get more torque, stepper motor produce most torque when stalled, as you increase the rpm the torque drops.

I know well know manufactures of encoder scale that produce an 8um pitch and you have all the american stuff.

When i built my first cnc'd x1 i was using a imperial z screw, you never need to know the interger apart from the time you enter it into the control software.

By using pulleys you get
-another source of backlash
-drive flex(belt stretch)
-reduced top speed
-side load on the screw
+ratio change
+allignment flexabilty

i tried belts on a lathe, big disapointment. now directly driven.

if we are quoting "you can lead a horse to water but cannot make it drink" I personally think your logic is flawed, if you want an additonal challenge that is far enough.

As for i am old school i like to watch it, the x1 had a rapid of 400mm/min and it got very old very quickly and that was a smaller machine.

[Jonathan]

I don't see the point of wanting a round number. As soon as you've entered it into mach3/EMC you won't need to know about it at all. I've forgotten what the value is per step on mine...

As fragger said, you will almost certainly be using 1/4 or 1/8 micro-stepping anyway which effectively changes the value but doesn't really gain accuracy.

By using pulleys you get
-another source of backlash
-drive flex(belt stretch)
-reduced top speed
-side load on the screw

1) no, unless your pulleys are awful. Mine have not added backlash.
2) negligible from my experience.
3) What! That's impossible to judge without calculating it properly. You need to get the motor running in the optimal region of the torque/rpm curve and to do that you need to get the ratio right which, unless you're very lucky, requires pulleys.
4) Surely the bearings in the mill are going to cope with that...if not add more? I'd be more worried about the bearing on the stepper motor.

On my router I've found that generally increasing the size of the pulley on the stepper motor has increased the top speed, but of course sacrificed resolution. It's less clear cut on a milling machine as it's heavier and the coefficient of friction is greater.

[Ricardoco]

No you don't get more torque, stepper motor produce most torque when stalled, as you increase the rpm the torque drops.

I know well know manufactures of encoder scale that produce an 8um pitch and you have all the american stuff.

When i built my first cnc'd x1 i was using a imperial z screw, you never need to know the interger apart from the time you enter it into the control software.

By using pulleys you get
-another source of backlash
-drive flex(belt stretch)
-reduced top speed
-side load on the screw
+ratio change
+allignment flexabilty

i tried belts on a lathe, big disapointment. now directly driven.

if we are quoting "you can lead a horse to water but cannot make it drink" I personally think your logic is flawed, if you want an additonal challenge that is far enough.

As for i am old school i like to watch it, the x1 had a rapid of 400mm/min and it got very old very quickly and that was a smaller machine.

I actually felt like you were telling me off then....

Well i will consider myself told shall I....:cry:

Usually people dont have to be so blunt to get the point across, Im not so stuck in my ways that im not open to suggestion, On a recent project I took a lot of advice from the members of this forum and put it into practice, that is why i participate in forums.

I dont really want to just copy you or hoss, but neither do i want to re-invent the wheel.

thats why i dont just guess at this, I've looked at several sources of information and am still looking, here is just one source.

http://www.arrickrobotics.com/pr23.html

I dont disagree with all you say, and there are sacrifices to be made using pulleys however my 260 is not direct drive, it is very accurate, as much as i need anyhow, and i really dont want the steppers hanging of the end of the axis (not sure what you would call a group of axis lol), so do you have any other suggestions?

Thanks in advance for your input.

[Robin Hewitt]

Using a ratio of 2.5 to 1 ( 10 and 25 tooth pulleys) will give me a step resolution of 0.001mm per step.

Hi Rick

AH! Confidence truly is that feeling you get just before you understand the problem :heehee:

.001mm is one micron. Standard ball nuts give you 50 microns, shimmed gives you 20, the only way to get 1 is to use double nuts and spring them. Helps if you can hold the screws in tension with more springs. The Belleville washer is your friend. You can compensate for backlash, but you will forever expect the tool to dig in.

Ball nuts bed in, they may feel firm but you are probably pushing against the dust seals and kidding yourself. The seals can't hold it when the tool loading kicks in.

If you bolt the column to a brick wall I would suggest a 5 micron step, 200 steps/mm is optimal for accuracy and speed. Half step is good, quarter step is vaguely credible, beyond that everything gets too springy.

Getting high accuracy is more a problem of rigidity than anything else. You can't mill to 5 microns because the tool will bend, you can't skim 10 microns, it will simply ignore you. To get 5 microns your tooling needs to cut a mirror finish. IMHO 5 microns or better has to be ground.

Good luck :naughty:

Robin