[murphy625]

I was doing some research on encoders. I want to be able to replace my current servo/resolver with a VFD (normal motor) and just hook an encoder directly to my ICM2900 breakout box. I figure I can send the +/- 10 VCS from the ICM2900 directly into the VCS inputs on the VFD.

But, while doing the research, I learned some interesting things. Not all encoders are created equal.
I figure I need the incremental type, something around 500-2000ppr (my app does not require tight tolerances).

My problem is that there are different outputs on these and I dont want to get into the 24volt stuff because it requires the installation of resistors on the ICM2900..
Anyhow. While reading the BEI site:
http://www.beiied.com/PDFs2/H25_Incremental_Encoder.pdf

What is an 8830 output configuration? What should I stay away from?
Im looking around on ebay for them..

Thanks,,
Murphy

[unterhaus]

A differential ttl output encoder will drive just about anything. Not sure what a VFD would take as far as encoder inputs, but I generally think you want to stay away with encoders with only one track and analog output encoders.

[Al_The_Man]

What do you mean by VFD normal Motor? as a VFD generally refers to a 3ph induction motor variable frequency drive, generally reserved for the spindle drive.
You can drive one direct from the ICM2900 you need a spare axis analogue output. Do you intend putting an encoder on this motor?
If the encoders you are looking at specify differential output then you should be OK as most use a 8830 or one with similar specs, there are multi voltage ones but most work on 5v.
There are a few models of Heidenhain differential encoders that do not have the line drivers and have a sine wave output. I personally have found that it does not pay to buy used encoders off ebay, make sure they state new,I have been burnt a few times.
Al

[murphy625]

What do you mean by VFD normal Motor? as a VFD generally refers to a 3ph induction motor variable frequency drive, generally reserved for the spindle drive.
Al

AL,

Yes.. I want the option of being able to replace my AC brushless servo motor (480v 1.3kw) with a normal 3 phase motor that is driven with a VFD. (the VFD accepts a velocity command signal of +/- 10V and has additional settings for ramp, current, etc.). The VFD also has an encoder input connections on it but I don't think I need them for what I want to do.
I would mount the encoder somewhere else on the machine and run the connections back to the ICM2900. Would this not allow the computer to read the encoder counts and send the required 10v signal back to the VFD?

Does this sound right to you? My machine is NOT your typical CNC machine and I don't require pin-point accuracy. I only need to control its motion on a long pair of rails.

Also,
When you order an encoder and it has one of those darn military connectors, do they provide you with the other side also to attached to your own cable?
http://www.bei-tech.com/products/enc.../shaft/h25.htm

At $450 I would think so but I thought I'd ask.

Thanks,
Murphy

[Al_The_Man]

I can pretty well assure you, you are not going to get precise positioning with this set up, I would imagine this is going to be pain to tune to prevent 'hunting' of the motor once it is approximately in position, Although some modern VFD's do lend themselves to almost servo like performance. But with no feed back to the VFD itself is probabally going to run in the velocity mode of operation.
I have purchased some BEI quite a while ago and they came with the MS connector.
Al

[murphy625]

Al,
My project is a large cart that rides back and forth on a pair of tracks. If it stops smoothly within 1/4 of where its supposed to be, I'll be happy.
What is your opinion on this? (I dont need CNC tolerances).

I guess where I am a little confused is on the feedback control...
Your saying this wont work? I need to know why.. As a servo/motion control newbie, I dont see why it wont work..

When I tell my cart to move 1 inch or 50 feet, I dont care if the timing and position are pefect as its traveling down the track. As long as its smooth and stops where its supposed to within that 1/4 inch.
(eventually, I will probably use proximity switches to verify position before continuing operation)
My drive train is a motor (currently a servo motor), that is connected to a 10:1 dual output shaft gear box. The output shafts go to rubber wheels that ride on a track.

Any comments? If you still think I am heading for trouble, I'd like to hear why..


Thanks again AL
Murphy

[Al_The_Man]

You may have to put quite a long accel. and decel. for the axis moves otherwise I can see alot of overshoot and 'hunting' with this set up if you try and position it to a rapid stop.
3ph induction motors are not servo's. Alot also depends on the quality of the VFD as some are better than others at positioning.
Al

[murphy625]

So it should work ???

Yes.. I set my encoder counts for the accel and decel very low. This cart carries objects that just hang below it and they can swing if it moves to quickly. Most times, moves are between 2 and 5 feet but there will be calls for it to travel the entire length (over 50 feet) on a regular basis.
Even the servo that is hooked up now is "tuned" very loose..
The reason I am using the servo is because it was 1. very cheap, 2. I thought it would allow me to make moves quicker, and then an "anti-swing" move that would reduce the amount of kenetic energy that is left in the swinging object after a more sudden stop. (sort of like a skyscrapers dampening device) Anyhow, as it has turned out, I dont think I am going to need that feature as the entire device does not really move quick enough to create any objectionable swing in the load below.

So again, your saying the model is sound but will not achieve the tight tolerances or quick movements that a servo would ?????????????????????????? <<that was a question.

Thanks again AL.

Murphy

[Al_The_Man]

So again, your saying the model is sound but will not achieve the tight tolerances or quick movements that a servo would ?????????????????????????? <<that was a question.

There's only one way to find out for sure
Al

[unterhaus]

The state of the art for induction motor control can do what you want, but I'm not sure it's commercially available. If your vfd is made for positioning control, I don't see any reason why you shouldn't be able to use it. However, you may well have to go to the encoder to get 1/4 inch positioning, depending on how many revolutions of the motor there are per 1/4 inch. If it is one or more, you have a good chance of success. The point is that sensorless controls have fairly bad position regulation, because they are trying to read the position from the windings of the motor. I would make sure that the vfd is intended to work to the tolerances that you are asking of it.

There are plenty of good encoders on Ebay, the only reason I don't buy them for work is that I can't use a credit card. Certainly, it takes some investment in time to understand what the encoder is.

And no, BEI does not include the connector, it is an additional cost. I built my cables last time I bought from BEI, but in general it's better to buy them for work. You're looking at $25 for the mil-std connector alone.

[murphy625]

unterhaus,

did you just say the encoder and VFD will work together just fine and give me that 1/4 inch?

I am using a Reliance Electric GV3000. It is not outdated as they sill list it as a star product on thier webpage. It takes a +/- 10v analog signal, has encoder inputs (which I wont be using in my case because my encoder goes back to the cnc control board).

Dont they use these VFD's and encoders on elevators all the time?

Hmm.. Should be interesting to see what happens when I try it.

Murphy

[unterhaus]

If it works, I'll claim credit

The brochure for that drive claims to be able to do positioning, but I'm not sure that it can do it without the encoder. You may have to put a y in the encoder outputs and share them. Looks like you want to be in vector mode. I would guess that they want a motor mounted encoder.

[Al_The_Man]

The main problem I see is the control source is also seeing the encoder input and unless you program in a large in-position band, the motor may probabally end up constantly servoing, due to the imprecise positioning capability of using a VFD.
Al

[murphy625]

Al,
I think you have a point there...

Thanks for the advice.. I guess I wont know for sure until I get it hooked up that way. (Which may never happen if the servo works out fine)

Murphy