Problem with VF2 accuracy

[X7racer]

I have a VF2 machine and recently was asked to make some parts that need exact tolerances (+- .0002). So I did some accuracy test holes with different drill bits and lines with some end mills. I found that on my X axis I am consistently off -.001 from the edge and between a series of holes will be as much as -.0007. On the Y axis it will be consistently -.0095 and between a series of holes will have -.0005 .
I used to use a mechanical click edge finder, and now use a laser edge finder, I don't believe finding the edge is a problem, and have done tests to confirm that.
I also use Mastercam X, tolerances and compensation is set to Zero.
In the VF2 control is there something I should be setting such as tool diameter or tool compensation, or is there something else not set right or not working right?

[HuFlungDung]

Drilled holes are not accurate enough to conclude anything from testing their positions. At minimum, you should perform circular interpolation. Myself, I would perform a circular interpolation of about a 2" diameter hole in aluminum at a 20 ipm feedrate. Then set up a high accuracy indicator in the spindle and sweep the hole to see how much out of round it was.

You could run another test where you would machine a nice 'diamond' (ie., a square with its corners on the quadrant lines) and measure what size you get with a micrometer. Keep the feedrate moderate (like 20ipm) for one test, then bump it up to 100ipm and do the tests again. This should give you a feel for what the machine will do under typical conditions.

[Geof]

.......You could run another test where you would machine a nice 'diamond' (ie., a square with its corners on the quadrant lines) and measure what size you get with a micrometer. Keep the feedrate moderate (like 20ipm) for one test, then bump it up to 100ipm and do the tests again. This should give you a feel for what the machine will do under typical conditions.

Hu you're brilliant.

I wanted a way to check just how good (or bad) my GR510 and TM3P are and this is so simple.

[X7racer]

Hmmm. nice idea, but. maybe I am not explaining enough. The job I have requires me to drill 64 .052 and 16 .125 holes in a varing degree of pattern.
Each hole has to be within +-.0002 tolerance from the edge and from the other holes as laid out in the plans. That is why I tested it with the same size diameter bits.
I did try the circle interpolation and measured it, the difference is .0007 out of round.
Being that this machine is 7 years old and was previously used in a high production runs, I am wondering if it needs to be recalibrated or is there adjustments that can be made in the control?

[Geof]

Have you checked it for backlash and the backlash compensation Parameters?

I don't know if it is the 'correct way' but I have checked backlash by clamping a piece of ground shafting vertically on the table; face one end and drill a hole through for a bolt.

Dial into this from the spindle and make that location the work zero.

Then just write a simple MDI program to move away a couple of inches then back to the work zero from +/-X and Y, and dial from the spindle after each move.

If there is backlash the machine will not return to the same position.

[big_mak]

If you are trying to drill holes within 0.0002" in position, you will be wrestling with this job all the way though, and no amount of machine calibration will help you out. You will need to finsh the best possible method to put the hole in the right spot period!!!

No matter how accurate the machine is, simply drilling the holes, even drilling and reaming will not get you in the true position.

Work on gettin grepeatable results drilling holes in position, then you can tweak what ever it needs to be tweaked with regards to backlash comp and such.

[HuFlungDung]

I cannot honestly say I've ever had to meet those kind of specs. But if the machine was a good one, I'd think you'd have to bore the holes to finish.

If you have time to fool with trials, you could drill undersize, then find an .047 endmill and a 3mm endmill to counterbore the drilled holes, hopefully bringing them into true position. Follow this with reaming to size (and fool around trying to get a reamer that will consistently ream dead on).

If a machine has a certain degree of backlash, you still might beat the odds by using unidirectional positioning and feed only between holes. That means, all positions should be approached from the same direction so that no axis reversals occur between holes. When you do have to start the pattern over, you program a dummy point that is past the origin (if the first hole is on the origin) so that the machine must come back in the same directions for the secondary operations. This unidirectional positioning used to be some sort of an option on the old Bandit cncs

Rapid moves introduce the possibility of servo overshoot if the tuning is imperfect, and this will leave the backlash really as unknown and uncompensated correctly. So a moderate feed speed from one hole to the next one may help obtain repeatability of position.

All theory, but to hit +/-.0002", you've got to consider every trick in the book. Also, don't try to test fit the holes with gauge pins, as you need .0002" clearance to slip fit the perfect pin, so that makes conventional testing meaningless. You'd need non-contact air gauging or something to determine the size.

[pants]

0.0002" tolerance is too tight. Your best chance is to first get your HFO to Relevel the machine, check geometrics, then ballbar it to get the basic compensation values and level as close as possible and then use laser calibration equipment to fill out the ballscrew compensation tables. You will need to do this in a temperature controlled enviroment, only moving in one direction at feed and pray to god your buildings got good foundations.
Im assuming you dont have linear scales on your machine?
Good luck

[PinnacleMachine]

There's no way drilling will get that accurate. As mentioned above drill undersized then bore to finished dimension or if your plate is thin enough drill then mill the holes finished . Or drill start holes and have it wire EDM.

JP