Servo driver

[karl1]

I have several Advance Motion Control Brushless Servo Controllers (about 24)that have been sitting in my "junk" box waiting for me to use on an electronic project. The data sheet: http://www.a-m-c.com/download/datasheet/bd25a20.pdf shows that the motor these drives use hall effect sensors for feedback. I was wondering if I could use a Gecko Drive G-REX100 to provide step/direction to these drives, and use the Quad. encoder inputs of the GREX to feed back the drive motion as a closed loop, to say, MACH3 software? The AMC Drivers are sturdy and proven design (and free), and it just seem logical to me to spend money on a better servo motor with encoder to upgrade my project from steppers to servos. Any advice would be welcomed.

- Karl

[ger21]

Mach3 doesn't support closed loop operation with a Grex (not yet, anyway).

Also, I don't think those drives accept step/dir signals. But I could be wrong.

You might be able to run them closed loop with Mach3 and a Galil card.

[yantra3d]

What about the pixie p100? Would that work for this application?

[karl1]

Thanks Ger1 and Yantra3d; this being my 1st post I knew someone probably would know a solution; Yantra3d, I think the pixie 100 will work for me. Now to select the servo motors and build something!

[yantra3d]

Well, I'm new to this myself and trying to figure things out so I wouldn't take my word for it w/o other's input!

I happen to have a very similar drive (BE15A8) I purchased w/ a small 12k rpm motor that I planned to use with a small hi-speed spindle @ 2.2:1 for milling wax at 25K rpm. I was planning on using one of the p100's to control the spindle speed, etc... and from what I can tell - it should do the job.

However, I'm not sure what the best method would be for connecting multiple p100's to drive a 3+ axis system via mach/pc. Could a multi-port pci card be configured in mach3 to send signals to each of the connected p100s? Or is there a break-out board option of sorts that you could connect the p100's to effectively?

Can anyone offer any info on the use of this p100 card? Any shortcomings worth mentioning? Loss of drive options or lack of loop control, etc??

[karl1]

Ahhh, I think I have the answer! I bought a rather inexpensive 4 axis parellel breakout Printed Circuit Board for my PC's parallel port; with about $10 USD in parts and $8 for the board, I was able to kludge together a project over a weekend. I am using it to presently drive stepper motors (another junk box treasure), but I figure it will breakout the signals sent by MACH3 to the parallel port - and give the step and directions from MACH3 and send it to the Pixie 100. The Pixie, thanks to your find, would then translate this info to the PWM signal for, in my case, the AMC Servo Driver. I found the breakout board at http://pminmo.com/; Phil has been very helpful to me and his service is fantastic. Although Skyco (pixie) shows out of stock until the 16th, I can now plan to use these "junked" drivers . And with the Pixie 100 comming in at <$70 per axis, I'll have a 4 axis system bodged together befor April.

Thanks,
Karl

[yantra3d]

Well, I figured out the breakout board part for driving multiple pixies for 3+ axis setup... The small 9-pin connector on the board threw me off I guess.

Skyko says that you can send an error line back through the port so does this mean that Mach3 controls the loop, and could then be considered a "closed-loop" system?

[ger21]

Originally Posted by yantra3d Skyko says that you can send an error line back through the port so does this mean that Mach3 controls the loop, and could then be considered a "closed-loop" system?

No.

[ger21]

Originally Posted by karl1 ......I can now plan to use these "junked" drivers

You can always send them to me if you don't want them. I almost bought a set on Ebay a few weeks ago.

[yantra3d]

So, I'm confused then about the open/closed loop configuration.... in this thread/review here he says at the end of the review, "I plan to convert all my machines to servo based systems and move away from open loop steppers".

Are you saying that it's not possible to create a closed-loop system with the pixie or are you saying that the error line to mach has nothing to do with the loop configuration?

Hey Karl, you may want to check this post out here. There's a video of a servo being tested with the pixie board and he claims to be using the same drive that you have.

[ger21]

Mach 3 can not use feedback and make corrections. Most servo drives close the loop in the drives. Mach just sends step and direction signals. It doesn't know if you're using steppers or servos.

I think, though, that if you use a Galil motion card with Mach3's Galil plugin, the Galil can make those corrections.

[Mariss Freimanis]

A closed-loop system compares the actual position against a commanded position and uses the difference between the two (position error) to move the actual position towards the commanded position.

At some place in any system something generates the commanded position and 'trusts' the actual position will match the commanded position. It really doesn't matter where that happens; Galil card, Rutex or Geckodrive, or in software.

I used the word 'trust' because closed-loop operation depends on feedback sensors (encoders in this case) and incremental sensors are themselves open-loop devices. How do you really know a 500-line encoder has accurately indicated precisely one revolution? You can't; you have to trust it. How do you really know a 10-microstep drive has caused a step motor to turn precisely one revolution? You can't; again you just have trust it has.

Quadrature encoders and step motors are incremental open-loop transducers. Incremental transducers can only add or subtract "1" from a sum on every motor step/encoder count. Any momentary error in either persists until it is homed (zeroed) again. The output of a quadrature encoder is exactly the same sequence as a full-step sequence needed to drive a step motor. An eerie coincidence.:-)

If you are a truly mistrustful soul, an absolute encoder offers some comfort albeit at a very high cost ($$$). An absolute encoder is self-correcting in that it outputs the absolute actual postion value for every location. If for some reason one particular feedback actual location is in error, the odds are the surrounding positions will be correct.

Absolute encoders aside, a closed-loop servo system or an open-loop step motor system depends on 'trust' at some point in the system. Both employ open-loop transducers, just that one is a sensor while the other is a prime-mover.

Mariss