Low cost Servo Controller

I missed that you intended to use rotory shaft encoders and build a "liner to rotory converter" to in effect get linear table scales. That obviously will increase the accuracy when there is backlash or poor accuracy screws. Check the DRO comments and there has also been a lot of messages about wrapping a wire around a shaft encoder drum.

I think you just took a big step toward a real servo system.

As you are proposing, you will need an interface for your encoders and it is not a significant cost difference if it is done as a plug-in board in a PC or as a separate interface in a black box.

You have already said that you need a servo amp so the only missing part is the D-A converters which are very simple and low cost and you would need them anyhow in your stepper to servo black box. The same applies here, we might as well stick a board in a PC and be done with it.

Bingo! You/we will have a real servo system. The only disadvantage is that we are limited to use software that will work with real servos.

Bertho
(with blistered hands from power screw driver after putting up too many shutters)
===============================

David Howland wrote:

> From: David Howland <dhowland@...>
>
> Bertho,
>
> Yes, back lash is a concern I have. I have heard of stepper retrofits with back lash compensation. The approach I have comes from the fact that rotary encoders seem to be about $15 to $40 surplus. Big DC brush type motors seem to be around the same price, and really large stepper motors seem to be around $70. By re-thinking the servo control loop requirements to take into account back lash it seems like the system performance will be a little slower than a real servo contoller, but faster and more accurate than a stepper system with near the same cost in components.
>
> If you have 3 open loop steppers for $210 or 3 sets of DC motors with the encoders mounted on the table (closing the loop) at $90 to $240 you just need to add the cost of the linear to rotary components to gain the benefit of not needing to worry about back lash compensation. It may be possible to use other linear table encoder technology to improve the concept at a little higher price. We just need quadature data from the tables. The trouble with rack and pinion interfaces to rotary encoders with reference to dirt and resolution are an area I have no experience with. Someone may suggust something better. I don't know what work has been done on DRO stuff. I think the speed and accuracy may be useful on old mills that may have variations in the amount of back lash from one side to the other.
>
> If you believe a high performance low cost servo system is more appropriate and has not already been done, what would be the goals for such a system? I mean of course, after you are finished having a huricane there in Florida.
>
> David Howland