[777funk]

I noticed when trying to cut 1/4" diameter circles using a 3/16" diameter bit with my techno gantry machine that I'm ending up with slightly oval shapes instead of a true 1/4" circle.

I'm getting 1/4" in the X axis but getting a little bit less (maybe .015" - .020") in the Y direction. It's just enough that it doesn't quite look right.

Is there anything that can be done about this?

I notice there's a little bit of looseness in the router's bearings (PC 7818) but that would give error in a circular (all directions) manner I would think.
I can push on the machine and get the Y direction (I have it set up like a Graph with Y) to move around .010" with a dial indicator base on the table and the needle on the gantry. But I would think just flex alone would allow for this much.

thanks!
Nick

[Montabelli]

look for simple explanations first, is there any possibility of slop in the thrust bearings, on one axis? it is a problem for my machines once in a while. check to see if the error is consistent between short and long moves, if it is something in the mechanical line is probably at fault, if not is the error proportional, and consistent? if it is then I would suspect something is off in the settings for counts per unit.

[777funk]

How would I check the thrust bearings? And if they are the problem is there an easy way to fix it?

Should I expect less than 10 thousandths error from my machine? It's a Techno Isle 040.

thanks!

[Woody111]

Check to see how much backlash you have in X and Y axis. The bearings should be fairly easy to replace they should be on the end cap. Sounds like the bearings are the cause make sure you put them back the same way they came out. That is very important!!!!

[777funk]

Sorry I'm pretty CNC challenged. I know what backlash is but how would I measure it?

[Woody111]

Mount a jo block on the table put an indicator in the spindle moving Xor Y axis towards the block in only one direction touch off on the block and zero your indicator next reverse direction .0001 at a time until your indicator moves off zero count how times it took add them up that will give you backlash. Should be less than .0002 0.0000 is best. If bearings are bad it will be .001 or even higher. Check Y first then X.

[judleroy]

How fast are you cutting the hole? If more than 5-10 ipm try slowing the cut down to that speed or even try a little less 2-3ipm and see if that fixes the problem. That small a hole even slowed down that much will only take a few seconds.
Judleroy.

[RandK]

Step back and cut a 1" or 2" square pocket in some scrap wood and measure the length of the sides with calipers. Work with that until you get the sides and the diagonals to be equal. As Montabelli says check the easy stuff first.

The trust bearings hold the ball screws, on the end plates, in place. Backlash is play in the drive train. The motors/encoders move a set distance but the play makes it less than it should be after a direction change. Like a car with worn steering. I have an 056 Gantry and it's pretty tight, backlash is less than a thou measured with an indicator in the spindle collet.

[777funk]

I can get about 12-13 0.0001" clicks before my indicator moves in the Y direction. I'm guessing that means I have more than 1 thousandth in backlash.

[RandK]

Be sure to measure in both directions of Y. If you've hammered one thrust bearing and not the other, the backlash would be asymmetrical. The machine has a rated accuracy and repeatability of .001mm or .0004". Not sure how much I'd trust jogging .0001". Jogging .001 will give you the same results and is within the accuracy of the machine. Anyways if you are really measuring .0012 backlash, then that's not the problem. I've got a bunch of pretty inlay dots from a far away country that are not cut very round.