[harlan]

Hello all.

I've been reading for a few weeks now. Lots to digest

My current situation....

We've got a series 1 BP J-head machine that i'm itching to convert to CNC.

We aquired an original Proto Trak controller complete with power supply, servos, brackets, controller, and both ball screws. The control seems to work properly, even after 15 years, but it's a bit clunky from a UI perspective. Untill I get it on the machine, i can't determine the final working order of the setup, but atleast the DRO works and the motors move. But for free, I can't complain

My issue is, whom ever de-installed the setup from the original machine took no care to keep the nuts and screws in one piece. I've got maybe 12 of the original balls, and best i can tell they are all the same size. The screws appear to be ground, with .200" lead. and the nuts are 3 circuit with internal returns.

I measure the balls to be .1367" with my trusty mic.

without special orders, i can get replacement balls, 25 grade, 3.50mm diameter ~.1378"

After all that, my real question becomes, am I asking for trouble with the size difference or am I worried too much about it?

The machine will eventually get a PC based controller, but i need to make the machine move to determine if i can reuse the servo controls.

thanks
harlan

[Ken_Shea]

My thinking would be since it is a used screw, doubtless having some wear that the additional .0001 pre-load would be beneficial.

Ken

[Geof]

Ken, your arithmetic is a bit off; 0.1378" is 0.0011" larger than 0.1367".

[Ken_Shea]

OOPS!
My math is great Geof, it's my eye sight that sucks

Well that may be a problem then !

[mwp]

Well you could try to load the nut with the oversized balls & see if it binds... if it does, I'd take it off immediately & buy the custom balls. I suspect that there is some wear on the screws/nuts since they are used so the .001 increase might just work. Another option is to send the screw out to be rebuilt. The rebuilder will make sure that everything is right (Assuming you use a good rebuilder). Down side to this is that it is pretty expensive. I've gotten quotes of between $1200 and $3000 to rebuild the ballscrews in my mill (56" screw & 24" screw). For now, I will just deal with the .004 lash.

[harlan]

Thanks for the input guys.

For $8 i'll order up a bag of the 3.5's and give it a shot. What's the worst that happens, they don't work and I've got ammo for cube wars at work

any tips for stuffing the nuts? other than patients and a little greese

[harlan]

Well, the 3.5mm balls were a tad too big. The nut would go on, but it was too snug for my taste.

Ordering up the custom sizes. :wee:

[harlan]

for any that care, we have lift off

An email to Bal-tec(www.precisionballs.com) got me a few stocking sizes around what I needed. Ordered up a few sizes but ended up using my first try.

From my measurement with the mic's of .1367", i had choices of .136670", .136779", and .136884" in a grade 25 chrome 52100 alloy.

i ended up with the 2 larger sizes, and ultimately tried the 779's first. after a little clean up with 1000 grit paper on the screw itself for minor nic's from the last owners I've got a very smooth running pair of ball screws

no noticable backlash in the nut along the screw axis, but this was me pushing on it. I'll know more when I get them in the machine.

Always have the next size up if they wear, or have more lash then i'd like when installed.

so, in hindsight, .001 is way to big

[mwp]

I'm glad to hear that this worked for you. I'm interested in hearing what the backlash turns out to be once you have these scews mounted.

[NC Cams]

Be careful when you're looking at backlash from a "ballscrew". You have to look at the whole system.

I'm a former bearing engineer who has gotten into machining and I own a BPT mill with the EZTRAK CNC package.

My research found that the oem BPT ball screw support bearings were offset ground special width 6204 ball bearings. Some had 75lb preload and later ones (circa 1995) had upwards of 125lbs preload. I don't know what the current machines are using.

In my case, you could still lean on the table and show up to 0.001" deflection (depending on size/strength of person leaning). And this was on a like new machine in very good condition.

This is lack of stiffness is due to the fact that radial load bearings are being used with an axial preload. THis is a common trick to tighten up the bearings but it does NOT create a stiff, low friction bearing as if it were a true ball screw bearing.

Again, these measurements were made on a pristine albeit used machine. Older machines could easily be much less rigid.

We first shimmed the bearings to get more preload. WE used thin strips of aluminum foil placed inbetween the OUTER races. WIth this trick, we were able to get nearly NO perceptible movement when leaning on the table.

HOWEVER, the force to turn the handles was quite high. We then removed the oem bearings and installed some true ball screw support bearings (480lb preload, ABEC 7 and 60 deg internal contact angle).

Result: no backlash or table deflection when loading the daylights out of the table. Better yet, even with a slightly preloaded ball screw (0.0001 oversize balls), the force to turn the handles was LESS than the oem bearings.

This all has to do with accuracy of true ball screw bearings as well as the special contact angle they have.

However, we still had some "looseness". A simple gib adjustment took care of the rest of the looseness.

Moral of the story: look at the WHOLE system when you start trying to take backlash out of a ball screw fitted mill table. You'll be glad you did.