Control Logix, Servo Problems

Ok.. here is a bit of back ground about our hardware and software

We are using:
Control Logix 5000
AB Ultra 3000 SERCOS Servo Drive
8 Axis SERCOS Card
AB Servo Motors (forget model)

This is what I am trying to do. I am trying to use the MRP command to define the position a servo is at, as its zero position. This sounds pretty simple, and it is, except for one little nasty problem banghead .
After the zero is defined we use the GSV command to get the values for the ACTUAL and the COMMAND position of the motor. The actual position is exactly 0, but the command position is 0.00020481.

AB tech support says this is a rounding error to do with windows bonkhead (AKA lets pass the buck to a company everyone already hates). But I dont understand how windows can have a problem with defining 0. zero is zero no rounding needed.

Has anyone heard of this problem before?

any help would be appreciated!!

Regards,

Steve

At a wild guess I'd say you're setting actual position with the MRP. What happens if you set command position? Do you get a similar offset in actual position?

I suspect you'll never get the two to be exactly the same.

One 'gotcha' I've encountered with the MRP is that it only has a DN bit (no PC bit) and it comes on in the control structure before the action is complete. The MCD is similar.

This is just a WAG, but the difference could be the minimum error required to generate enough holding torque to keep the motor from turning.

Gerry:

Yeah, I probably should have been a little bit more clear. When I am doing the MRP, I am defining the command position as zero, and im also using absolute.

That is what I find frustrating about it mddr . I tell it, that the command position is zero, then use a GSV to monitor it, and it says that its off by some small amount.

I have not used the analog equivalent to the drive im using now, but im told that it does not have that problem. So maybe its a SERCOS problem, I dunno .

Slightly off-topic, but...

I've never been able to rationalise the need for the option of command or actual position when doing the re-define. Since either choice will not cause axis movement and (ignoring your observations for the moment) sets both attributes to the same value, why have an option?

This is not a windows problem!

When you define a position, you are specifying a absolute position in floating point. Now what if the motion controller can not get EXACTLY to that position because it operates in counts? I bet if zero was defined to be an integer number of counts the problem would disappear. Some motion controllers do not even work with engineering units and dodge this issue. They let the user worry about these rounding errors.

An Rockwell could define 0 as a integer and a fraction of count but then the servo/PID would never be able to get to this position and the integerator would wind and unwind trying to get at this position.