PDA

View Full Version : Low cost 0 backlash transmission - Ball bearing + trapezoidal screw



LabSkis
17-03-2011, 07:15 AM
Hello,

Despite some search, I haven't found anything on this forum about this technique described (in french) here: http://cncloisirs.com/Construction/Entra%eenementVictorSendas
Here is a picture from this site which can help you to understand:


3841

It seems very simple, a ball bearing rolling on a trapezoidal screw and quite efficient in terms of backlash reduction & transmission without friction.

Have any of you tested this technique? What are the feedbacks?
I am wondering what is this "nut" in which the ball bearing is maintained?

Thanks for your answers!

M250cnc
17-03-2011, 11:17 AM
My view is that most if not all leadscrews are soft "Not Hardened" the profile of the leadscrew is not made to be used in the scenario you describe.

So in the standard situation the most expensive part "The Leadscrew" runs in a bronze bush/nut "The Cheap Part" so the bronze bush is replaced at minimal cost preserving the integrity of the leadscrew.

The situation you describe is a bearing is used to run down the thread. Now for the problem, the bearing "The Cheap Part" is hardened and it is gonna wear the expensive leadscrew in no time at all as its harder than the leadscrew.

Now you could use a ballscrew which is hardened but then its not much more trouble to use a ballnut and get an even better solution.

Good English BTW
Phil

BillTodd
17-03-2011, 03:26 PM
Extensively discussed here:
http://www.cnczone.com/forums/linear_rotary_motion/66381-you_gotta_see.html

see also :

http://www.cnczone.com/forums/linear_rotary_motion/13593-different_better_kind_screw.html

Bill

M250cnc
17-03-2011, 04:04 PM
Well i looked at the 3 page thread, nobody mentions the downsides i brought up, i will not be looking at the 51 page thread.

Phil

Jonathan
17-03-2011, 11:58 PM
Have any of you tested this technique? What are the feedbacks?
I am wondering what is this "nut" in which the ball bearing is maintained?


I have done it on my Y axis as 'Chip' said it's in my thread and videos on youtube. There's a good shot of it in the long video about wind turbine blade mould if I recall correctly.

http://www.mycncuk.com/forums/showthread.php/2288-1.7*0.74*0.4m-Mill-Router-building.../page11

(I've measured the backlash and recorded it somewhere in the thread - pun not intended!)

I used 16mm threaded rod with a bearing that happened to fit - can't remember the dimensions offhand but I'll check if you want. I didn't put any sort of insert on the bearing - it's just running directly on the screw. At the moment it's working perfectly well, however the screw is visibly worn. Not much, but noticeable. Initially the backlash was very very low (practically zero - clearly you're never going to actually get zero), now it's a little more but still good. Better backlash than a cheap ballscrew at least.


The situation you describe is a bearing is used to run down the thread. Now for the problem, the bearing "The Cheap Part" is hardened and it is gonna wear the expensive leadscrew in no time at all as its harder than the leadscrew.

Easily solved - press a brass/bronze insert into the bearing. I did this for an M12 threaded rod (was going to use it on Z) and machined the insert to fit the thread profile better. I think if you want it to work for a long time then that's the way to go. I'll do it at some point on my Y axis, but for now it's adequate - better than the X-axis nuns.

M250cnc
18-03-2011, 10:59 AM
Easily solved - press a brass/bronze insert into the bearing. I did this for an M12 threaded rod (was going to use it on Z) and machined the insert to fit the thread profile better. I think if you want it to work for a long time then that's the way to go. I'll do it at some point on my Y axis, but for now it's adequate - better than the X-axis nuns.

If say you were using a bronze nut "You Don't Use Brass" that was 12mm long that would have a contact area on the thread of (12x3.142)x12 of 452mm approx, using a bearing this way would have a contact area of approx 20mm that's 22 times less contact area not to mention threaded rod is low grade with a poor surface finish so will wear out the bush in no time.

The fact that this idea is not commercially available says it all really.

Phil

Jonathan
18-03-2011, 06:58 PM
If say you were using a bronze nut "You Don't Use Brass" that was 12mm long that would have a contact area on the thread of (12x3.142)x12 of 452mm approx, using a bearing this way would have a contact area of approx 20mml

True, you're not going to be able to put anywhere near as much force on it - and the pressure is much greater due to the tiny contact area. Probably only a couple of mm, not 20, since it's only contacting at two lines (in theory). Could improve this with more bearings.



The fact that this idea is not commercially available says it all really.


Yes it does. Except one thing - the sort of machines we're making are generally not used for heavy cutting (like commercial machines). I think this method excels when you're only cutting woods/plastics so the forces are low. ACME rod would certainly be better, but I just used stainless rod as I know that eventually I will probably buy a ballscrew.

It could be good for a PCB machine where low backlash is going to help with accuracy and the cutting forces are tiny.

M250cnc
18-03-2011, 07:17 PM
True, you're not going to be able to put anywhere near as much force on it - and the pressure is much greater due to the tiny contact area. Probably only a couple of mm, not 20, since it's only contacting at two lines (in theory). Could improve this with more bearings.

I was being generous, so its even worse than i thought it was.



Yes it does. Except one thing - the sort of machines we're making are generally not used for heavy cutting (like commercial machines). I think this method excels when you're only cutting woods/plastics so the forces are low. ACME rod would certainly be better, but I just used stainless rod as I know that eventually I will probably buy a ballscrew.

It could be good for a PCB machine where low backlash is going to help with accuracy and the cutting forces are tiny.

To drive a small PCB machine a belt drive like they use on the RepRap machines would do the job better, be easier to implement, would last longer and be far stronger. You also do not need angular contact bearings for the fixed end of the screw either. That is the HD 10mm pitch belts BTW.

Phil

Jonathan
18-03-2011, 07:30 PM
To drive a small PCB machine a belt drive like they use on the RepRap machines would do the job better, be easier to implement, would last longer and be far stronger. You also do not need angular contact bearings for the fixed end of the screw either. That is the HD 10mm pitch belts BTW.

Belt would certainly be a lot easier, not to mention faster. The 'Rapman' machine at school was pretty poor to be honest, mainly due to not being put together very well though I think! It only used 5mm belts and the pulleys were laser cut from acrylic.

M250cnc
18-03-2011, 07:37 PM
Belt would certainly be a lot easier, not to mention faster. The 'Rapman' machine at school was pretty poor to be honest, mainly due to not being put together very well though I think! It only used 5mm belts and the pulleys were laser cut from acrylic.

Your not agreeing with me are you. :eek:

I use these belts on my lathe the Z is about 300mm BC

I was cutting some tool steel and a bit too hard it pushed the bar in the lathe chuck but the pulleys are about £30/40 each, built it has awesome power. Backlash is 0.05mm

Phil

Jonathan
18-03-2011, 08:04 PM
Your not agreeing with me are you. :eek:

I use these belts on my lathe the Z is about 300mm BC

I was cutting some tool steel and a bit too hard it pushed the bar in the lathe chuck but the pulleys are about £30/40 each, built it has awesome power. Backlash is 0.05mm

Phil

Sorry maybe I came across wrong there - I meant to agree with you!
I meant to say that the laser cut pulleys and only 5mm wide belts were not good.

I've used XL pulleys and 1/2" (some bigger) belts on my machines. HD belts would be nice...


...better than the X-axis nuns.

Just noticed that typo I made - oops!

M250cnc
18-03-2011, 08:14 PM
Jonathan

I was joking, as we always seem to be at loggerheads.

Phil

Sent from my HTC Desire using Tapatalk

LabSkis
19-03-2011, 08:35 AM
Thank you all!
Jonathan,
You have done exactly what I intend to do.
Except that I was going to use an ACME rod.

Where did you find this thing on the picture, enclosing the ball bearing?
Did you get it machined?

3857

Laurent

Jonathan
19-03-2011, 10:10 PM
Where did you find this thing on the picture, enclosing the ball bearing?
Did you get it machined?


I cnc milled it on my machine. I can send you the drawing/Gcode/make another if you like. I made it that slightly odd shape to fit in a tight space on the Z-axis before changing the router design somewhat and using it on Y. If I did another for the Y I would leave more material there for the bolts, and put a spring/belville washers on the bolts so that it rides over imperfection in the screw,

LabSkis
26-03-2011, 08:44 PM
OK. Thanx for your proposal.
But I have nothing around here to get this thing machined.
For the moment, I think I will stick to 0 backlash nuts like the one you find on dumpster cnc.
Yet, do you think it has a real advantages on those nuts in terms of backlash or friction? Despite the Acme being worn by the roller bearing?