sudden change in registered axis position

[shoneill]

I've run into an unusual problem.

I'm using LV 6.1, and the motion software drivers which are bundled with it.

I am controlling a 3-axis robot via NI-7344, all three axes equipped with feedback encoders (resolution 1 micron).

I'm using a micrometer switch attached to the Z-axis to measure the height of a substrate while scanning in the X&Y directions (move to position, move down until triggered, register position).  I'm using high speed capture in order to capture the switching position.  I've been achieving reproducibilities in the range of 0.5 to 1 micron.

So far so good.

I set up a scan last night, and it worked - mostly.  Half way through the scan, the Y-position seems to have shifted itself by around 1 to 1.5 mm.  The position information returned by the driver software is the same as it SHOULD be, but I can see visually that it's simply in the wrong position.  A shift of my substrate can be pretty much ruled out (never 1mm).

It seems to me that the driver software had a "tick" and forgot the previous encoder settings (or "lost" 10,000 counts somewhere).  I'm running Win98 (I know, I know), and the PC has been running for around a week now.

Anyone got any ideas?

Shane.

[Jochen]

Shane,

could you please provide some more information:

1. Which NI-Motion version are you using?
2. Are you using servos or steppers?
3. What motion mode are you using (linear, arc interpolation, contouring,...)?

Jochen Klier
National Instruments Germany

[AnalogKid2DigitalMan]

All mechanical connections secured and motor couplers (if any) tight?

[shoneill]

Ak2DM and Jochen,

Thanks for the reply.  I've since found the problem, and lo and behold, exactly what I thought it couldn't be, it was!  The substrate holder shifted.

Hang on a second.... (wipes egg from face).  OK.

So I guess you're correct to ask if everything is tightened.  It wasn't.  That's what happens when you don't fix the samples in place.

I couldn't check the substrate until the measurement was finished (takes several days) so this is why I'm responsible for a nuisance post.

Doh!

Shane.

[Jochen]

... and I imagined myself already designing test scenarios and urging R&D to examine the board's firmware. Thank goodness there are still some issues out there that don't require weeks of investigations...

Have a nice weekend,

Jochen