Hi all,

I have just fitted the ballscrews and steppers ( x and Y ) to my mill and i'm getting some backlash.

the mill is a RF45, i have class5 screws and nuts, 2 angular contact's at the fixed end's that are preloaded. they have no play and don't move any where. the nuts are fixed in there mountings by 6 screws and they don't move ether. both steppers are 960oz, with olden couplings? to the screws.

when testing for backlash..... X .02mm/.03mm, Y .04mm

any idea's where this backlash could be coming from?

thx in advanced.

Man I hope I can get that tight when I get mine in. Are your ballnuts preloaded? I think I read somewhere to expect backlash even with ballscrews unless they're double-nutted. Sounds like you are running single nuts.

hi MackMax,

yep, i have preloaded single ballnuts "zero backlash". well i hope so, they cost enough:eek:

Hmmm, I was thinking that in order for the nut to be preloaded it was a double nut. My bad. I know the double nuts are expensive and I surely can't afford those. If yours are supposed to be zero backlash I wouldn't have a clue to find that slack you have according to the info you provided.

Rick

i may of got the terminology wrong on the preloaded ballnut bit. another way of saying it would be........"it's got big ball's in it"

I am hoping to retrofit my x2 with double ballnut. Not the one piece but two pieces opposing each other. You guys think it will work?

Alex

Hi all,

I have just fitted the ballscrews and steppers ( x and Y ) to my mill and i'm getting some backlash.

the mill is a RF45, i have class5 screws and nuts, 2 angular contact's at the fixed end's that are preloaded. they have no play and don't move any where. the nuts are fixed in there mountings by 6 screws and they don't move ether. both steppers are 960oz, with olden couplings? to the screws.

when testing for backlash..... X .02mm/.03mm, Y .04mm

any idea's where this backlash could be coming from?

thx in advanced.

I know this will sound silly, but something is not as tight as it seems. The screw people will blame the bearing people who will blame the screw people who will blame you as the builder. Look for Vince's bport cnc conversion log and there is a great portion where he fights exactly the same thing.

i may of got the terminology wrong on the preloaded ballnut bit. another way of saying it would be........"it's got big ball's in it".
I believe the proper term is it has been fitted with oversize balls to get to zero backlash. When we talk oversize, we are talking .0001" oversize balls.They keep going up by that amount untill there is zero backlash.

I am hoping to retrofit my x2 with double ballnut. Not the one piece but two pieces opposing each other. You guys think it will work?

Alex

2 nuts working in opposition is a great way to do it. If you allow for some adjusability or spring load them, you will be able to compensate for any wear in the future.
It can be more expensive, but if you have the room and the money, I like it better than the oversize ball route. Just my opinions of course.

2 nuts working in opposition is a great way to do it. If you allow for some adjusability or spring load them, you will be able to compensate for any wear in the future.
It can be more expensive, but if you have the room and the money, I like it better than the oversize ball route. Just my opinions of course.

not too expensive really. I got quite a few ballnuts I bought over the course of my cnc hobby life and all I need to do is drill and tap the other end of the cncfusion kit where the piece bolt onto the table. Put the other ballnut on in teh opposite direction, that's it. I am getting some taps off ebay, what a bargain. 3 15/16" tap for just USD30.

Alex

are you sure your backlash is not coming from the motor itself? I know my small 269 oz steppers have a bit of backlash in them when they change directions. it takes about 3 steps to fully change direction within the motor itself.

TOTALLYRC,

i'll have a look at Vince's log. thx

i had a little play with the X fixed end and found the A/C bearing is sitting flush with the bearing holder. so i faced off a tho or 2 so when i put the holder back on it would make dam sure the 2 bearings are being pushed together. that should make the outer of the bearings nice and tight. only thing is, i dropped my DTI and snapped the end off:nono: so a cant do any thing till monday.

Fixittt.

i will have i look at the motor when get a new DTI

The way to measure backlash that is inherent in the nut/screw/table interface is to put the tip of an indicator held in the spindle on a vertical face (side of a vise, angle plate, whatever) then grab the table and shake it back and forth. You'll find out in a hurry how much slop is in the axis. I just totally rebuilt a retofit RF45 that utilizes Thomson preloaded ballscrews and servos. I bought it secondhand from a guy that didn't know the difference. When I disassembled the X and Y axes, I discovered that the builder had not locked the nuts to prevent them from rotating in their mounting block. Good components do not make up for poor implementation and attention to detail. The X and Y axes went from .008" and .006" backlash respectively to .0015". Most likely this is being introduced by the bearing and not the screw/nut interface. Regardless, I can live with it now.
Joe

Where/how did you measure this?

Phil

are you sure your backlash is not coming from the motor itself? I know my small 269 oz steppers have a bit of backlash in them when they change directions. it takes about 3 steps to fully change direction within the motor itself.

Where/how did you measure this?

was this for me?

if it was, this is what i did to measure the backlash.

i fixed a slip to the bed of the mill. put a clock in a collet and went -x till the clock read zero. i then backed away 10mm using 100mmpm. then used slips till the clock read zero again. the difference in the two measurements was my backlash. i repeated this over a distance of 20mm, 30mm and so on. all came out .02/.03.

scudzuki

totally agree with you. after all, the dovetails on a RF45 are not exactly ground to perfection, more like cut out with a bastard file. not the smoothest I've seen, but do i lap or scrape. from what i've read on the zone that debate will go on for ever.

When I disassembled the X and Y axes, I discovered that the builder had not locked the nuts to prevent them from rotating in their mounting block. Good components do not make up for poor implementation and attention to detail. The X and Y axes went from .008" and .006" backlash respectively to .0015". Most likely this is being introduced by the bearing and not the screw/nut interface. Regardless, I can live with it now.
Joe

Funny this comes up today!!! I bought a 15" travel Thomson Superslide (2 linear rails with integrated carriage + likely the same preloaded ballnut / screw as yours) a little while back, and just got around to playing with it today. Never had been apart, so must have come this way..... I could move the table a good .005-.007" by hand, with the ballscrew locked tight. That can't be right.... Unscrewed the ballnut from the slide table, and suprised myself to find 2 ballnuts with an adjuster in the middle. 1 set screw on top, took that out and still couldn't figure out how to tighten it up.. Nothing online I could find... The 2 wavy washers were loose to the touch. Then I accidentally flipped the ballnut over (it's square, so didnt' flip by itself until the ballscrew accidentally flexed) and sure 'nuf, there was another set screw on the bottom! Doh! Loosened that one and wow, the ring turns and you can adjust it... I have less than .0002" of backlash on the screw now! :) Adjustable preload ballnuts are awesome! I get a solid 400+ipm and I've had it up to 1300ipm. If you can find stuff like this surplus, go for it!!! :)

No it was for Fixittt as per the attached quote. I was just curiuos how he measured backlash in the motor itself. I have difficulty seeing how the motor can make 3 steps without the shaft turning.

Phil:)

was this for me?

No it was for Fixittt as per the attached quote. I was just curiuos how he measured backlash in the motor itself. I have difficulty seeing how the motor can make 3 steps without the shaft turning.

Phil:)

I thought it was just me Phil.:confused:
The only way the motor could have backlash internally is if the armature was loose on the shaft.
If this is the case throw it away as fast as possibly. If what Fixxit means is that he uses the computer to input steps one at a time and it takes several steps to make the motor move, then he may have a backlash lash setting or something else in play. Depending on the number of micro steps involved, it may be inpossible, with the measuring device involved, to measure the first couple of steps.

I am not trying to piss off Fixxit, I am just trying to wrap my sleepy head around this idea. Time for more coffee.

(chair)Bad Mike(chair)

Here is my findings on the backlash in the steppers I have.
First off, I have 8 269 oz steppers from Xylotex. I run 8 and 10 microsteppings on my two machines.
I had spent alot of time tracking down backlash. what I finally ended up doing was taking a DTI and setting it up on the rear shaft of the stepper, Tip of the DTI on the edge corner of the flat. Set mach3 to .001 in steps and stepped one direction until the DTI needle moved. Then stepped the other direction until the DTI needle moved again the other direction. It takes 3 steps to get the needle on the DTI to move in the opposite direction. being the 3rd press of the jog button to physically see motion on the DTI. This holds true to all the steppers that I have. So it very well maybe the stepper design limitation, I do not know.
I hope I explained this so you can understand what I am talking about......

That's odd... I have the same motors (Xylotex 269oz) and can do .0001" movements (1/25 microstep) and the needle moves immediately with each step.. What driver are you using? It may be that the driver can't accurately move the motor 1/8 or 1/10 of a step, some drivers are better at it than others. I have a Xylotex board I could try and hook back up to see if that does it.. I replaced it with IM483 drivers ($20 used on eBay) to get more resolution and more power a while back. The good drivers (Gecko's, IMS drivers, Centent, etc..) have some crazy lookup tables in them that really help get accurate microsteps. *Physically*, there can't be any backlash in a stepper motor shaft as it's solid with the part that moves.. The only "slippage" can be magnetically if the steps aren't complete.

- Just set up the IM483 with 1/256 microstep and it still registers movement though I'm not sure how accurately it's moving since that would be 0.000009765625" per microstep with the 2TPI ballscrew. :) I can't enter that into Mach's manual jog box. lol

Hello;

You have re-discovered a little detail that only a few people know about: Micro-step position depends on the direction of the movement, when using the same micro-step to coil currents table for both movements, and that is precisely what most of the drives on the market do.

The two ways to solve that problem are:
First; Using positioning feedback to correct micro-step coil currents (closed loop drives).
The second way is: Using different current tables for forward and reverse motion on programmable open loop drives.

A simple test fixture to show this effect could be made by attaching a laser pointer to the motor shaft, and micro-steping the motor (for a full step interval) in both directions while marking, on a sheet of paper, the position of the laser dot (projected on a the paper at about 10' distance) at each micro-step number, like shown in the Mardus-Kreutz test #3 (http://www.youtube.com/watch?v=ghUNdTDp69s) video in Utube.

A real open loop precision drive should have custom made micro-stepper tables (because the directional angle difference depends on each motor's parameters). That is relatively easy to do by using an encoder attached to the stepper shaft and adjusting A and B coil currents to get the same micro-step positions when moving in both directions, then constructing a current vs micro-step number table for each direction.

You will need an open source micro-stepper drive in order to implement the above suggested changes...

There is another little detail that only a few people know. Micro-step size is generally not the same for every micro-step (even when moving in only one direction), it all depends on the motor magnetic/mechanic design parameters. That is why using micro-stepping generally does not increase positioning accuracy.

Kreutz.

Hey Fixxit, not for any thing, but it could be you indicator is a little sticky. The step positioning of a motor is +- a little bit but I think I would try with a tenths indicator, if you or a friend has one. I have one that I picked up somewhere and I only take it out for special occasions.

UPDATE,

got my nice new DTI yesterday and did some test on the x axes. first thing was to put the clock on the end of the floating end of the spindle and turned the fix end by hand. doing this gave me a reading of 0.10mm. not knowing if the ballnut or the A/C bearings was to blame, i took the whole thing apart.

(this is where the "call me stupid" bit comes into play)

one of my A/C bearing was round the wrong way(chair). i think i did this the last time i put some shim between the bearing's, just after that i broke my DTI. so i wouldn't of known till now. any ways...... worked out how much play there was in the bearing's and made up a peace of shim to match.(making sure to put them back in the right way round this time :D) so hopefully that will take all the backlash out of the fixed end. i then put the table back together, but this time when it come to doing the jibs. i put the DTI on the X axes and while moving the table i tightened up the jibs. i did this till the DTI stopped moving. i now have 0.01 of end float.

I'll have a look on the Y axes tomorrow and see if I've been just as stupid there.