[cnczoner]

Hi all,

I'm building a small gantry (24" x 18" travel), that will move a super light load (~2lbs). I really need .002" resolution, but would like to get 0.001", and as many IPM as possible. For the Y-axis, there are 2 linear rails (one on each side), and I want to drive both sides simultaneously to prevent "racking". I'd going with belts, as I can use one stepper with a long shaft to both sides (eliminates out-of-sync issues if two motors), each with it's own belt. Ballscrews will get expensive as I'd need to get two, then use 2 steppers, two controllers, etc. This drawing shows one side of my game plan... http://tinyurl.com/yexcmg5

So now the calcs... as I understand it, microstepping will not let me get more resolution, but is just used for some smoothness between full steps ... is this correct? Or can I use say 1/2-microstep for a bit more resolution? Assuming I can't, each step will have to move the belt 0.001". With 200 steps/rev, each rev will be 0.001" x 200 = 0.2", which is now the circumference of the pulley. So that back-calculates to a pulley diameter of 0.0637" ... LOL!!! Even if I went for 0.002", that's a pulley diameter of 0.1275", so still LOL!

Plan B is to move the stepper elsewhere and use another pulley system to the long shaft to reduce the stepper motion. Is there a simpler way?

Cheers,
-Neil.

[vger]

Neil,

Microstepping will indeed increase the resolution. However, it will also reduce your max speed and torque. The drive method you have pictured it similar to that of copy machines and flatbed scanners. I'd be tempted to move the motor out and up so the belt could be attached closer to the centerline of the rail and bearing. Long lengths of timing belt are available, so the belt does not have to be a closed loop, it can have it's 2 ends attached directly on the centerline of mass of the gantry. With rapid starts and stops that will keep you out of the pendulum effect.

With a 1" diameter pully you would get about .0157" per step with full step. The smaller the pully size, the more resolution you will get, but there is a point at which accuracy will suffer due to tooth depth. Using 1/16 step drive would get you down to .001" steps with a 1" pully.

How fast are you looking to travel?

Steve

[Mariss Freimanis]

vger,

"However, it will also reduce your max speed and torque."

That isn't universally true. You get 100% of the motor's high speed potential if the drive morphs from microstepping to full-stepping at speeds higher than 3 revs per second.

Mariss

[cnczoner]

I'm a bit confused now. I read another post stating/indicating/implying that we can't stop between steps to get higher resolution. Mariss, since this will be Gecko 201's with Automation Direct 276 oz-in stepper motors, can you confirm that I can stop between steps (1/2 step)? And what about 1/4 step or 1/8 step?

It seems I'll be writing my own software now, so I can easily single-step from source to destination locations, but then microstep at each end to get the precise position. There will be no side force when stopped at each end (this is for a pick and place system).

[Mariss Freimanis]

cnczoner,

Not true. You can stop anywhere between full steps and have exactly the same torque. Motor torque is the vector sum of the phase currents; if those currents are sine and cosine then the vector sum is a constant (sin^2 + cos^2 = 1). Motor position is the vector angle.

Mariss

[vger]

Yes, with microstepping you can stop at the fractional steps. With a 10 microstep drive, just send 10 pulses to get a full step, 5 pulses for a half step... and so on... The accuracy of the microsteps is dependant on the motor and the drive. Don't expect exactly 0.1800 degrees per step, maybe a 5% error in microstep angle, but back to right on every 10 steps...

Steve

[cnczoner]

Wow, great to know, thanks.

BTW, for speed, I'd really like to get 200ipm, but it's been tough to figure out what's realistic because most of the stepper datasheets I've seen don't list max rpm. I'm using those 276 oz-in steppers for now cause I have some, but I may change to Keling's 495 oz-in units if torque at high rpms becomes the limiting issue. Avoiding servos for cost reasons, but never say never.

Cheers,
-Neil.

[ger21]

I'm using those 276 oz-in steppers for now cause I have some, but I may change to Keling's 495 oz-in units if torque at high rpms becomes the limiting issue

If you do, you may end up worse than before. Usually, smaller motors will have more torque at high speeds than larger ones. In a lot of cases, increasing motor size to get more speed will actually give you less speed.

BTW, for speed, I'd really like to get 200ipm, but it's been tough to figure out what's realistic because most of the stepper datasheets I've seen don't list max rpm

Max rpm would depend on the application, because steppers lose torque as rpm's increase. An unloaded stepper on a bench may spin at 3000 rpm. But in actual use, it may only work up to 500rpm. Look at the torque curve to get an idea of the torque it'll have at a given speed, and calculate if that torque is enough.

Typically, you start with how fast you want to go, and how you're going to drive the machine. Then you calculate how much torque you need to accomplish it, and choose a motor that fits.

[cnczoner]

Was not aware of that, but I just checked the Pacsci datasheet for a couple examples and that really seems to be the case. I would've thought that a double-stack motor would be the equivalent of two single-stack motors for twice the torque at all rpms.

So that raises the question -- for a given rpm, how does one get more torque? Surely there must be some other motor option...??? I'm keeping servos out of the question for now.

BTW, I found a couple motors with 400 steps/rev and 800 steps/rev, but their website/quoting process implies that they won't come at hobbyist-source prices.

[ger21]

So that raises the question -- for a given rpm, how does one get more torque? Surely there must be some other motor option...???

Do you mean a given stepper rpm, or a given screw rpm?

There are lots of different motors available. You can search their torque curves and fins one that fits your application. Or, you can design around the motor.Use a belt drive to spin a screw faster and the motor slower, or use different pitch screws.

If you want more torque at higher stepper speeds, look for motors with higher current ratings, around 6-7 amps.

[cnczoner]

In my case either, as my stepper to screw (or belt in my case) rpm is fixed, as dictated by my resolution requirements.

This is still confusing -- I now know that using a larger (more holding torque) motor has similar torque to a smaller motor at higher rpms, but you're suggesting a motor with more current. Isn't that the same as a larger motor? Or just some specific larger motors?

[ger21]

I'm not aware of any smaller high current motors.

Keling has a 400 oz Nema 34 rated at 6 amps. KL34H260-42-8A No torque curve, but probably similar to an Oriental PK 296-4.5A, which is 440 oz and 6.3A. They also have the 465 oz KL34H260-60-4A

[BobWarfield]

cnczoner,

Not true. You can stop anywhere between full steps and have exactly the same torque. Motor torque is the vector sum of the phase currents; if those currents are sine and cosine then the vector sum is a constant (sin^2 + cos^2 = 1). Motor position is the vector angle.

Mariss

So to be clear on this, the holding torque at any microstep is the same as at a full step?

However, as I understand it, you might not get going again, because the torque available to move to the next microstep may be way less per links like this one:

http://www.micromo.com/n391003/n.html

How should we be thinking about this, Mariss?

I have always looked at it similar to the table in that link. If your application requires, say, 1/2 the available torque from the motor, you can't rely on more than a 1/2 step when microstepping.

If that's not true, it's good news, but I would like to confirm it.

Best,

BW

[Crevice Reamer]

Hi CZ.

With the right drive/motor/PSU combination, you should be able to get good resolution AND very fast rapids.

Many are confused about picking motors. Basically, you must match the motor to the PSU, and to the driver. It's a package deal.

First decide on a driver. Morphing drivers like Gecko G540 or G203V will carry the most torque up to the highest speed--IF you power the motors at their best Voltage. Powering a motor below it's best Voltage will cause loss of potential high-speed torque.

So you must also pick a motor whose best Voltage and Amperage is compatible with the driver's max Voltage and Amperage.

For example: You mentioned the Keling 495 oz. motor. This is a 85V, 3A motor. It will provide 495 oz inches of torque at rest--and considerably less at speed. Running this motor with say 50V PSU will cut the high-speed torque available almost in half. To get the Max torque at speed, you would have to use an 85V PSU. Even a Gecko G203V cannot handle that much Voltage.

Now if you were to choose the G540, a 48-50V PSU, and the Keling 54V 381 motors, you would surely get everything you wish for.

http://crevicereamer.com/Page_2.html

CR.

[cnczoner]

Crevice Reamer, That's a not-so-intuitive approach. I think about this in the opposite way ...
(1) Since I know my resolution, I can figure out my belt ratio (back to the motors)
(2) Assuming I know the force required to move my head/gantry assembly (which I don't currently because there's some info missing for my linear rails, etc), and my desired speeds, plus my belt ratios now, I can determine my motor torque requirements at various rpms.
(3) Search for a motor to achieve the specs above.
(4) Add whatever driver is required to move the motor at close to it's limits.
(5) Add whatever power-supply is required to move the motor at those specs, and of course staying within the driver's capabilities.
(6) Get whatever PC is required to provide enough steps to the driver from there.

The approach you suggest seems almost like trial-and-error.

[Crevice Reamer]

The approach you suggest seems almost like trial-and-error.

Well, no--It's called find what is working successfully and emulate it. You don't HAVE to reinvent the wheel.

Your way will work if you get reliable data. The trouble starts when you believe torque speed curves to be reliable data. These are specific to the motor drive combination, and like statistics they can be fudged and outright lie.

Also, bigger motors are not necessarily better. You will learn these things after you have wasted enough money.

I'll suggest just one more thing before I butt out: You seem to be missing only one bit of data to solve your equation--the force needed to move your gantry. Get a fish scale, pull the gantry with it and voila you will have that measurement.

Good luck with your project.

CR.

[ger21]

Ideally, you'd also want to calculate the force required to accelerate at a given acceleration rate. Acceleration can make a bigger difference than top end velocity..

I guess I never really read your first post very well. You might want to consider a 3:1 or 4:1 gear reduction. It'll give you much more force, and still give very fast speeds.

Guys are building rack and pinion machines with similar reductions and getting 700-1200ipm. Belt drive is basically the same as rack and pinion.

[cnczoner]

So my final plan is to use a 1.019" pulley to drive the belt that directly moves the linear block. That pulley will need a 1/16 reduction to get .001", but I think I'll try 1/4 reduction and try 1/4 microstep to get .001" and see how accurate it is. If not, it should be relatively easy to experiment with different ratios. I just need to find a decent encoder to check accuracy. Thanks.