high-precision rotary positioning

[polikmus]

Hi There,

I'm designing a high-resolution positioning machine. It has to move a load between 0.001 deg/sec (in fine mode) to 30 deg/sec, and position repeatability of +/- 0.001 deg. Max torque driving the load is 200 Nm. Backlash & hysteresis are concerns. The movement range is nearly 360 deg total. The motor & gear-train have to have a Ø1" through hollow bore for wires running through the axis. The motor has to run on 24 VDC!

I have two ideas.
1) use a small direct-drive motor (for the hollow shaft) attached to a 160:1 ratio harmonic drive. Have a high-end encoder at the output. I'm concerned about the flexibility of the gearbox affecting my fine controlling. I've only found one motor manufacturer (applimotion.com) that makes 24 VDC hollow shaft motors. Unfortunately, they make frameless motors, which I've never designed with before.

2) use a large direct-drive motor attached to a high-end encoder. I have not found motors that can meet the specs (most use too much power, and their low-speed torque is too small for their size). I’m also concerned about what it takes to control a motor to such small position increments and such low speeds.

Does anyone know how chip-making machines or similar systems control such fine movements? I suspect that they use direct-drives.

Any help or comments on the two design choices, and any motor manufacturer leads, would be appreciated.

Regards,
Polikmus

 

[KENAT]

We use products from Brooks automation for manipulation of wafers.

They may be able to help.

KENAT, probably the least qualified checker you'll ever meet...

 

[eromlignod]

Have you looked at Daedal?

http://www.daedalpositioning.com/products/Rotary_Positioners__5504__30_32_80_567_29.html

Don
Kansas City

 

[drawoh]

polikmus,

What is wrong with frameless motors?

Do not use an assembled Harmonic drive. Use the reducer kit. You need a shaft going from the harmonic drive to whatever the motor is. The shaft locates the harmonic drive wave generator, so it must be accurate and rigid. The frameless motor's armature must be fixed onto a rigid, accurate shaft, and darn it, you have one!

You have a compact layout and no shaft couplings. How many shaft couplings have a 1" diameter bore?

JHG

 

[mikek10]

You don't want a separate encoder in my opinion - why introduce extra components - complexity, cost, potential for backlash / error ?

Better to use a servo with built in feedback.

It would be a trivial task to do this if not for the 24V constraint. What is the reason for this? If its for cost you may well find this is a false economy.

 

[drawoh]

mikek10,

He can use a through hole encoder on the output, or even on the input. Another possibility is to use a frameless stepper motor. There is no need for an encoder. Certainly, there is no need for a coupling.

JHG

 

[Timelord]

Polikmus,

A large hollow shaft direct drive servo and a high end encoder are definitely the way to go. Drive train compliance is real important when micro radian accuracy is required. Harmonic drives are notoriously compliant because a flexible wave member is required. I have worked on several high end gimbals for directed energy weapons and direct drive was the only way to achieve the accuracy required. Drive friction is also very important to the point where unsealed bearing were required because of the friction of seals. Frameless custom motors are the norm for this type of work to assure a stiff connection from the motor to the load. If dynamic accuracy is not required you may be able to get by with a gearbox and some very sophisticated controls, but the encoder must be on the final driven axis.

Timelord

 

[polikmus]

Thanks all for your responses.

To me, the inside of motors and harmonic drives are black boxes. I'm not confident enough to design using their subcomponents. No one at my company has done so either (that doesn't mean it wouldn't be fun trying!). With frameless motors, I'd need to find out how to mount, glue to housing & shaft, assemble, keep metal bits out, interfacing motor with encoder & drive, etc.

@Kenat: Brooks seems like a big company. I doubt they will tell me the mechanism they use inside their machines. I'll try that as a last resort if I can't find a method here.

@eromlignod: I'll look more closely at Daedal later. The ratings are a little low. I'll check their max speed, life, and voltage input. My application is actually one of these tables turned on its side, so I also need to check their backlash.

@drawoh: I've only worked with Harmonic Drives, LLC out of new york. By "reducer kit", I think you mean their component sets or simplicity models? I actually tried designing a frameless motor to harmonic component set once, and boy was that complicated! It was the shaft coupling that got problematic, unless you're thinking I make my own longer harmonic drive input shaft (I'll have to see what the manufacturer thinks it takes to assemble my own shaft into their wave generator). I'll also keep this as a later-resort if I can't find another method using housed motors and housed gearboxes.

@mikek10: A servo system I've designed before has a motor-encoder-drive from Danaher, with its own PI-loop for controlling its speed. That motor attached to my gearbox with my encoder at the output side. I ran a PID-loop, controlling the motor speed in order to get smooth output-encoder movements, in order to compensate for backlash and gear errors. I assume you mean to use only the output encoder and run that through one PID-loop total. I may sound ignorant here, but doesn't a (non-stepper) motor REQUIRE an encoder to time the pulses to run? How will having backlash or "springiness" in the gear train affect this? 90% of our customers run 24 VDC. It would be sooo much easier to find a motor with 230 or 400 VAC!

@drawoh: I did not know that stepper motors come in hollow shaft. It seems obvious, but I've never seen these before. I'll ask my reps.

@Timelord: Compliance is a concern of mine. I hope that during fast runs, I can slow it down before reaching the destination, then switch to fine-control mode and move slowly with an over-damped system. So, “custom-made” is the key here then. Unfortunately, I need to present a bread-board prototype by May. My design also calls for a gimbal. The velocity will have to be smooth only during the fine-mode slow movement. It seems wasteful that I’d be using only the bottom end of the motor’s speed range. I’ll ask my vendors for custom possibilities.

 

[steveh]

I would try Tsudakoma. They make excellent,accurate rotary tables.

http://www.komaprecision.com/tsudakoma/Tsudakoma Main.htm