Linear Motion Double Ball Nuts

[TomS]

If you have (or know someone who does have) a 3d printer, it's pretty quick to print a sleeve for a removal/install tool. I have a set of the double ballnuts and my backlash on a small mill is just under and over a thou. I can't comment on long term because I've only had them a few months. I bought mine through Automation Technologies but Chai machines/supplies them. I suppose you could adjust the preload between the two nuts with shims, but I haven't looked into it as my backlash is minimal. They are the exact same assembly as the single nuts, so larger balls are also an option. I actually ordered the single nuts and was planning on upgrading the balls down the road, but they sent the double buts with a note that said "free upgrade" which was nice, but I already had the saddle milled and ready for single, so had to mill more material to make room for the double nuts on my mill. I'm very pleased with the setup and with a very small amount of backlash compensation set in Mach 3 it's perfect/zero.

As a side note, I removed the balls from mine just to count and make sure they we're correctly filled. I had no real problems (using a little grease and tiny screwdriver) repacking all the balls. Took a little bit of patience, but not that difficult. If you have backlash issues and it's not the ballnuts/screws and everything including your screws bearing mounts/ preload is all right, I'd personally suspect the couplers. Triple check your set screws. I used to use a lovejoy type coupler initially and switching to a oldhams is the way to go. They are pretty easy to mill/make using steel or aluminum for the hubs and delrin for the "disc". Mine came out nice then snug, but you can always hand file the center discs sides to ensure an absolutely perfect fit with zero backlash. The lovejoys just don't seem to be as precise or solid in my experience, and for those using the low cost aluminum "flexible" couplers... throw them away... almost anything is better.

Good to hear that you ended up with near zero backlash. That's what I'm trying to achieve. I thought about going the larger ball route but from what I've read it can be a guessing game on what size balls to use. Then if you guess wrong it's another set of balls and the time to tear down your machine and go through the process again.

My backlash could be coming from the nuts, couplings, mounting brackets, or bearing preload or a combination of all these. I will check each one as I disassemble my mill and make corrections as required. BTW my couplings are double diaphragm type.

Tom S.

 

[coherent]

There are a number of folks who have gone the larger ball route and a little searching should find specifics as far as exact size in which nuts and eliminate the trial and error process.
The double diaphram couplers are normally much better than the standard flex couplers . I've only used them on large RC helicopters years back but never for CNC applications mostly because the good ones I've seen are pretty pricey.
Regardless of the type, one thing I have had issues with is clamp type vs set screws on a shaft flat. I've never been able to completely eliminate eventual play/backlash over time with the clamp type, including using threadlock and ensuring everything is extremely tight. There must be a trick to securing clamp type couplers that I'm not aware of.

 

[jbolt]

There must be a trick to securing clamp type couplers that I'm not aware of.

Add set screws to the clamp style couplers. If the shafts have flats use blunt nose set screws, it the shaft is round use pointed set screws and drill a detent in the shaft the same point angle as the set screw.

 

[TomS]

Add set screws to the clamp style couplers. If the shafts have flats use blunt nose set screws, it the shaft is round use pointed set screws and drill a detent in the shaft the same point angle as the set screw.

That's what I did. As I recall you were the one that mentioned this when I was building my machine.

Tom S.

 

[chevydyl]

I don't have flats on my screws where my pulleys are fastened, just two set screws 90 apart, I have around 12 lb ft and 3hp on those pulleys, with quick acceleration and 250ipm rapids, never lost placement.... they are also not super tight...

 

[TomS]

Finally got some time to disassemble my mill and install the double ball nuts. I installed the nuts on the screws and the Y and Z nuts are smooth in both directions when turned by hand. The X ball nut turns smooth in one direction with the same resistance feel as the Y and Z nuts but has slightly more drag in the opposite direction. The movement is smooth but is this something to be concerned about?

Tom S.

 

[TomS]

Didn't want to hijack mrbasher's thread so posted my response here.

Got my mill back together and operational today. After fiddling with AC bearing preload I ended up with about .0005" backlash on X and Y. The Z axis has about .003". Not sure if it makes sense to spend more time trying to get it closer to zero as my motor standoff and AC bearing housing design is a pain to disassemble and reassemble. Next step is to adjust the gibs. Played with the Y gib a bit but not having much luck getting consistent dial indicator readings. There's always tomorrow.

Jim Dawson suggested, "If you can preload the Z-axis in one direction or the other with a counter balance, you should be able to get that to zero backlash."

Jim - thanks for the suggestion. I may give it a try. My .003" Z backlash could be lack of sufficient bearing preload, flex in the head/column assembly, or something else I'm not considering. I'm leaning towards adding a .005"-.007" shim between the bearings and see if the backlash goes away. If not then it's most likely flex in the head/column.

Tom S.

 

[JimDawson]

The X ball nut turns smooth in one direction with the same resistance feel as the Y and Z nuts but has slightly more drag in the opposite direction. The movement is smooth but is this something to be concerned about?

I would just run it for awhile and let it ''wear'' in a bit. The drag will change with a bit of use as things seat in. Once you get a few hours on the machine go in and make minor adjustments as needed. I would write a program that exercised the machine through the full travel of all axis and just run it in a loop for a few hours. Make sure your auto lube system is working correctly, excess oil is better than not enough.

 

[TomS]

I would just run it for awhile and let it ''wear'' in a bit. The drag will change with a bit of use as things seat in. Once you get a few hours on the machine go in and make minor adjustments as needed. I would write a program that exercised the machine through the full travel of all axis and just run it in a loop for a few hours. Make sure your auto lube system is working correctly, excess oil is better than not enough.

Thanks Jim. I've got a warm up gcode file that runs the three axis to their extreme limits. Still adjusting gibs and assembling a few other things. Once that's done I'll run it for a few hours as you suggested.

Tom S.

 

[TomS]

Using the Bridgeport method for adjusting gibs I was able to reduce the X axis table deflection to .0035" (Bridgeport states .001" is the target). Any tighter and the 1600 oz. in. stepper motor stalls. Having never measured my table deflection before I'm not sure if .0035" is as good as it gets but it does seems a bit too much.

I've played with the velocity and acceleration settings and the motor seems happy at velocity of 150 and acceleration of 50. Base setting was 100 velocity and 18 acceleration. Could the motor stalling be a current setting issue? Or I could leave it as is and test cut a hole to see how round it is.

Tom S.