Hi,
I'm new to this forum and if this has already been covered in detail - please forgive me!

I have a "Chinese" CNC machine, controlled by a properly purchased/licensed copy of Mach 3 (R3.043.062).  The software is on an XP PC with 3.66GHz processor and plenty of memory, using the LPT to drive the steppers.

I've been using this setup for a while with no noticeable problems, but a recent, bigger, project has turned up 2 problems:

• Occasional warning message "Scripter compiler error in:M3.m1s".  I found some responses on this forum to this error - but when I look in the file I see no spurious characters or other corruption.  I've not got around to replacing the .m1s files with .mcc as suggested, but will try that tomorrow
• I am noticing cumulative position errors, principally in the Y-axis.  I have a couple of test patterns - the one that reverses direction to return to it's starting point, cutting in both directions, does not seem to produce a position error - the one that cuts in one direction only, then a "fast" move back to the start ends up with the 0.0 point a couple of mm displaced.  I have a third test that cuts a curved path then fast move back produces an even bigger error - but in the Y-axis only!  In each case, the cut repeats 10 times, but with the Z-axis zero set well above  the material so there is no cutting load (and the spindle is off)

I don't know if these problems are connected - it's difficult to see how they would be.  I have checked that the Y-axis stepper is firmly locked to the leadscrew.  The feed speed is set quite slow and doesn't seem to make any difference, nor does the cut depth for each pass if I actually cut into the stock (softwood) - or if I keep well clear of the stock surface.

I'm finding this very frustrating - any suggestions?

Many thanks,
Mike.

Maybe missing steps on rapid moves on the Y axis?  Try slowing down rapids.

The other one I have never seen but I have read that the saving as .mcc works.

Thinking about the "direction" of the error - it could be explained by "slippage" in the rapid moves back to start the next cutting pass.  I'd been concentrating on the possibility of excess load on the cutting pass.  I'll have a look at that in the morning.

Thanks,
Mike.

OK - time for an update!

• The "scripter compile error" message was, as others have found due to a corruption of the original m3.m1s file.  In my case, it wasn't the addition of "random" characters, but the addition of some space and line-feed characters - more difficult to spot!  I deleted them and no more error messages.  For some reason, I was not able to save the result as .mcc, but the error has not appeared again since - so I'm happy.
• The cumulating position error was a bit more tricky to resolve.  I reduced the maximum speeds, but that didn't help.  In the end, I fiddled with the pulse-widths.  I couldn't find the original information for my controller, but I'm pretty sure it said to set the pulse width to 10uS.  I found the manual for a similar controller, which said 5uS - but that didn't work at all!  8uS produced rough motion, 10us same results as before and 12us - perfect!

Now to get on with getting some work from the machine!
Many thanks,
Mike.