[Lasershop]

V2XT Boss 8, SurfCAM 7.0. Just got my mill up after not using it for a bit. Have some stuff to do that is fairly high precision. Replaced all the axis drive cards 2 years ago with new EMI units. I milled some holes for pins that should have been .188" diameter using a .125" end mill. Coming out to be about .175" diameter. This is just plastic. Just got done rechecking all axis backlash and it is excellent for this old machine, about .0005" both ways. Lead error is about the same. New screws a few years ago. I use good NC cutters also. How do I compensate for the motion inaccuracy? Somewhere within Surfcam or the machines software? I have never "tuned" the new cards, have no idea how but they seemed to work well so far.

[9 1/2]

Any chance your plastic (not knowing exactly what sort of plastic you are machining) is retracting after the cuts? 4 thou a side?

Just a thought that the issue may not be hardware?

[Lasershop]

Any chance your plastic (not knowing exactly what sort of plastic you are machining) is retracting after the cuts? 4 thou a side?

Just a thought that the issue may not be hardware?

No. Its acrylic I use for the bed. Machine is set up for coolant also. Actually did the same holes 3 times.

[machintek]

Read the program and see what it is telling the machine to do.

Did you check backlash in a dynamic state?
I usually overcomp the backlash and cut a circle with a light spring cut. I can then sweep the circle with a dial indicator and measure the steps where the axis reverses direction. This is the amount an axis is overcomped and needs to be subtracted from the exaggerated value just used.
It is a lot easier to measure a step than a flat (if zero comp was used).

Use cutter comp and adjust the tool diameter so it comes out correct.

Was it a G79 that is such a beautifully simple was to circular interpolate a hole? You could try that.

You have many options.

George

[Lasershop]

Read the program and see what it is telling the machine to do.

Did you check backlash in a dynamic state? What does that mean?
I usually overcomp the backlash and cut a circle with a light spring cut. I can then sweep the circle with a dial indicator and measure the steps where the axis reverses direction. This is the amount an axis is overcomped and needs to be subtracted from the exaggerated value just used.
It is a lot easier to measure a step than a flat (if zero comp was used).

Use cutter comp and adjust the tool diameter so it comes out correct. That seems like the easy out. I have to do some linear milling tests to find out if the error is accumulative.

Was it a G79 that is such a beautifully simple was to circular interpolate a hole? You could try that. I do all my mill programming within SurfCAM. I wouldnt know G code if it hit me in the head. Surprisingly though I wrote all of my software for my laser post about 7 or 8 years ago using Galil based programming and C++. Its a stand alone program. I can easily adjust the counts per inch with mine. Makes a difference depending on winter or summer. Some of the laser jobs I get I have to hit +- .0005. I was hoping I could find the count per inch variable within the SurfCAM or V2XT software just to play with it.

You have many options.

George

RonO

[machintek]

There is always a difference between a atatic state and a dynamic state due to stiction. The coifficient of friction to start maotion vs staying in motion. Hence why I overcomp and mill a circle, and measure the steps.

George

[Lasershop]

There is always a difference between a atatic state and a dynamic state due to stiction. The coifficient of friction to start maotion vs staying in motion. Hence why I overcomp and mill a circle, and measure the steps.

George

I did some linear tests on the mill when I found the incremental jog toggle. I secured my digital calipers and measured several lengths of movement jogging the machine a set distance. The V2XT appears to be pretty accurate and in excellent condition. I an convinced at this point the post on my SurfCAM is not entirely accurate as far as counts per inch. Like I said before, we get pretty anal about the accuracy of the lasers and it needs to be adjusted from time to time. My big machine is usually within a thou’ over an eight foot distance, but not every day. Think Ill just have to lighten up on the mill and not expect as much from it. I did over (under?) compensate my tool diameter and its better on my one inch by three inch test piece but I don’t think that will help on a longer distance. Can I ever expect to nail geometry within 4 digits, or even 3 with this mill?

[NC Cams]

We were able to get our Extrak (same table, screws, ball screw bearings and drives as V2XT) to mill round circles withing a few 0.0001" of roundness BUT it took a lot of money and work.

First you spend about a day or so iteratively adjusting the gibbs so as to remove slop and stick-tion (start up friction).

Then you pull the ball nuts and have them reloaded with oversize balls to take out nut slop.

The BIGGEST contributor to accuracy was achieved via replacing the ball screw support bearings. Do a search for "Extrak" on this site and details as to what we did can be found.

With properly preloaded angular contact bearings replacing the factory modified special 6204's, you can EASILY get low three digit accuracies with the X and Y axis of an Extrak and/or V2XT.

IF you spend additional funds and in the right areas, you can get low four digit decimal accuracy. In our case, we're well within 0.0002" on profile when we mill our cam grinding masters with our Extrak - on milling circles, we're under 0.0001" at the inevitable step that occurs at the 4 direction change points.

The electronics of the BMDC found in our ExTrak and V2XT will easily do the accuracy. The iron, however, needs a lot of TLC in just the right areas.

[Lasershop]

Wow! Thats pretty serious! Fortunatly I dont use the mill to make money. Its an in house machine just for whatever but I love it just the same and Ill probably never get rid of it. I did replace all 3 motion cards mostly out of the fact you cant find them any more. Had one go bad but just did all three. Curious though what you think the mechanical upgrades you did cost.

[NC Cams]

The Ball nut reloading cost about $150 or so - I don't recall. This was done by Roberts Ball Screw in Dansville Michigan.

The gibb adjustment was performed by me and I spent a couple of days getting it figured out and then adjusted to move without drag or slop when I turned the handles. This takes time and patience as you have to check, recheck and readjust over and over to get it as slop free as I was looking to get.

The ball screw bearings I used to replace the OEM 11060203 bearings were the real killers. What my bearings cost is not a "real" number. I had connections and got the bearings surplus. Why? THe OEM bearings are lame but adequate for a glorified drill press. For true ball screw work, they aren't even close - they are not rigid and they have too much slop.

Because I was looking to go to the extreme, I used TRUE machine tool grade preloaded ball screw bearings. The "feel" and accuracy are UNBELIEVEABLE.

The posted price of the bearings I used was something like $830/set and 2 sets are needed - 1 each of X and Y. You could use something a bit less costly and get similiar performance. I gave the alternative P/N's in my Extrak postings that can be found via this site's search enginer.

7204A5TYDUHP4's by NSK would probably do almost as well as the ones I used and the 7204's should be more affordable - even these would be FAR superior to the OEM bearings for both X and Y.

If you're doing 3 D milling with the V2XT, the 11770118 bearing that supports the Z axis should be replaced with a 7014A5TYDUMP4 as this is far superior in accuracy, preload and rigidity than the OEM bearing. Thi bearing is a true ABEC 7 machine tool bearing - the OEM is a modified set of 6014's selectively picked to obtain ABEC 3 runout specs. YOU can change the P4 to a P5 and use an ABEC 5 instead. These might be a bit cheaper.

You will have to make up a small spacer to compensate for a width difference between the 7204's and the 11060203's on the X and Y axis - laser cut steel shims or preloading wave washers from McMaster Carr would suffice.

If you have a BMDC, you may/will probably need a servo tune to really get the thing to really shine, especially in cutting ROUND circles. Brian at BPT Machine in Carol Stream Illinois (630-784-9942) did mine. I think I paid time and materials for the service call but don't recall anymore.

If Bridgeport had taken this sort of care and attention to detail when they built the V2XT, Haas whould still have a competitor for the TM1 at a fraction of the cost. Hardinge could do so as well if they still had the BMDC and were interested in selling XT's instead of just the Traditional mill but I suspect they don't care that much anymore.

Cost wise, the tune up I did it was far less than a TM-1 Haas and the machine now matches that in accuracy - I already had the mill and the knowledge about "how to do" the mods came from my experience as a machine tool bearing engineer. The cost was worth it as the thing is deadly accurate and a true joy to operate anymore.

Better yet, when you do something with the machine that any number of people say "can't be done", the benefit is priceless.

[Lasershop]

.... when you do something with the machine that any number of people say "can't be done", the benefit is priceless.

Very good info. Thanks! I have a lot of patience with tuning any of my machines. It takes what it takes. Sometimes minutes, sometimes days. Definetly going to try to duplicate your mods on mine.
Thanks again, RonO