Need synchronization of X and Y axis by offset

I have a tape library XY assembly with two servo motors onboard (I am using two open-source Elm servo boards with step-dir input). The servos are both fixed to the base, meaning the x axis does not have a servo moving around with it, and therefore whenever the Y axis moves, the X axis also moves, because it is riding along a stationary belt. In other words, the X servo must compensate for any Y axis movement, approximately 1 step X for every 6 steps Y. I understand that software compensation is not a perfect solution for this, but I think EMC2 or LinuxCNC as it is now called should have a compensation feature somewhere, right?

How could I fix this problem with LinuxCNC?

Are there any examples of something like this that come with LinuxCNC so they can be directly experimented with?

It seems to me that something like this came up before on this forum, a machine was built with the x and y motors mounted to the frame and the carriage was moved by cables.
All moves required both motors to compensate for the other.
I believe the idea was to keep the mass of the carriage as light as possible for fast acceleration.
Unfortunately I cannot provide a link for that, perhaps someone will chime in with it.
But it may give you a starting point.

Rick G

Here are photos of the device, your forum is way too slow to upload anything.
s9.postimage.org/mfy3gtupb/image.jpg
s13.postimage.org/68l3kdznb/image.jpg

There is actually one left for sale,
hxxp://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=380375049171&ssPageName=ADME:X:RTQ:US:1123
I am 90% sure they will take an offer of $80. They have been trying to sell it for a long time. It will probably be gone shortly with this posted here.

There are problems with the assembly. It cannot be used as a precision positioning device, the backlash is awful. I had to disassemble it and use PL polyurethane (find some PL premium at the Home Depot in the large 29oz size tubes) to form a more rigid connection between the acetal nuts and the carriages. Nothing really sticks to acetal but the polyurethane at least expands to fill in all gaps. Even so, the nuts have some pretty bad backlash. This is good for fast positioning but the the assembly is not rigid enough for anything most CNC people would be interested in. It is good only as an educational tool until the problems are ironed out.

In case anyone is interested, here is the pinout of the tiny servos on the device:

With the gear end of the motor facing up, the motor at the top and the connector below it,

1 3 5 7 9 11 13
. . . . . . .
. . . . . . .
2 4 6 8 10 12 14

1 Motor -
2 Motor -
3 Motor +
4 Motor -
5 Motor +
6 Motor +
7 Encoder -
8 N/C
9 Encoder -
10 Encoder +: 3.3V to 5V
11 Encoder -
12 Encoder Channel A quadrature output
13 Encoder -
14 Encoder Channel B quadrature output

OT but the good news is that the forum slowness (its been bad for about 2 days) is fixed!