A different (better) kind of screw

[cnc2]

Hello Lucien !

Sorry, i meant adapting it for a lathe... i've not been following the thread, i just remembered your roller screw was for a lathe & thought this kind of disengagement was easily adaptable as a module near the motor.

I'm not a machinist yet but i can say: you see backlash everywhere just kidding !
remember you can make it backlash free...look again the profile of the "engagement nut" is a bit trapezoidal....blablabla

Still learning... (do say that for myself)

Thanks !

cnc2.

[rokag3]

We are now on test.
We have some problems
the cam system does not work you can turn the cam with the excentric but not the excentric with the cam (not what we want)
so we have the idea to put a pinion in the center and small pinions on the excentric it work ...but
the screw find a way to bend and we have some back slash .15mm
so we change the profile of the roller
still bending .12mm
so we decide to put a spring on each roller
still bending .08mm
so we add a ring in plastic to prevent the screw to move and we have .04mm but very fast destruction of the plastic ring.
I still do not understand how this 32mm screw can find a way to move .The profile of this screw is very round this does not help.
except that the regularity of the screw is very impressive exactly 5mm per turn checked with the dro 5micron
So we are going to do it but it's going to be more complex than we thought.

To work by hand it's very pleasant but .04mm backslash is too much to work with the glass scale as encoder we must be better.
I will send some pics and movie so far we cannot say we are very happy but we keep on.

[rokag3]

#506 11-26-2008, 10:23 PM
BillTodd Join Date: Aug 2008
Location: UK
Posts: 301


It's all Greek to me ... chasing microns

Last 'blog' from me, I have some real work to do ( although, I'll still be hanging around).. Today I slotted the lead-screw mounts to allow them to sit down a couple on millimetres. Then adjusted the nut mounts (twice!) to lower the roller nut so as to clear the table (I'm now back with the nut frame level with the mounting surface) While the machine was apart, I removed photographed and measured the Y nut. There's a lot more room around the Y nut so it should be an easy task to replace it. Once re-assembled, I was disappointed to find the same horrible springy backlash that I had with the mounting nut loose I spent/wasted a couple of hours trying to find the cause: I re-dialled the lead-screw with my little Peacock gauge (which I use to align video tape heads - it is scaled in microns - 2um/div). I thought I found 20um backlash in the thrust washers, so rather than re-shim them, as a quick fix, I drilled an tapped the end so I could apply pressure on the thrust washers with a bolt. Even when the bolt was way too tight, there was still movement in the gauge - I realised I'd been measuring the flatness/squareness of the end of the lead-screw I then dialled the end-plates (were they flexing?) and the table (was it moving sideways slightly?) and pretty much anything else I could fit the gauge between. Nothing - it must be the nut . In despair, I started to clean up. Then , just as I had put everything away, I remembered something Lucien (rokag3) had said a thousand posts ago: "The top roller should move" (or some strange Greek translation of the same meaning ) Lucien you were absolutely correct! What happens is: The weight of the lead-screw/mountings/handle during assembly, is enough to push the grooved rollers apart slightly, taking the weight off of the top roller. Once fixed, the lead-screw seems OK but, under load, the lead-screw will ride up (it bends) which causes the springy backlash. The fix (in my case) is to loosen the lead-screw mounts and pull them up (to force the screw against the top roller) then re-tighten the mounts. Bingo -next to no backlash an a taut handle. My Y nut design will have fully floating rollers.

we have the same problem

[BillTodd]

Hi Lucien,

Good to hear from you again

Sorry to hear about the backlash problem, but I'm confident you'll be able to fix it.

I would try to adjust and fix the bottom roller(s) so it is in firm contact with the screw. Then, apply spring force between the top rollers (so one roller pulls on the other). This worked well on my Y nut.

Are you sure it is the lead-screw that is flexing? Are you able to test on a 'rig'?

Regards,

Bill

[cnc2]

Hi everybody !
Hi Lucien !

what kind job can be accomplished with 0.04mm backlash ?
And more important for me: What can not be done with 0.04mm backlash ?

I'm just trying to understand how machine accuracy limits machine capabilities.

Thanks & Good luck !
cnc2.

[rokag3]

I would try to adjust and fix the bottom roller(s) so it is in firm contact with the screw. Then, apply spring force between the top rollers (so one roller pulls on the other). This worked well on my Y nut. Are you sure it is the lead-screw that is flexing? Are you able to test on a 'rig'?

Yes we can see that the screw is bending (we have visual access)
i think that the profile of the screw is the big problem. The rough trapezoidal screw of my lathe did not have this bending effect,
I agree with your strategy to have the bottom roller fix and the upper floating.

What you can do with 4/100mm almost everything specially if you activate the backslash compensation.
What you cannot do:using linear scale as encoder (hunting effect guaranted)

[rokag3]

they are not very fresh i hope i will show something much better saturday with a movie of the bending unless we solve the problem .We are working on a much more adapted profile and think to billtodd who think me before we are going to test with a fix bottom roller.Anyway we are going to reach our objective maximum 1/100mm backslash with a load of 100kg (+the weight of the table)at 3hz

[cnc2]

Thanks Lucien ! for the explanation.

Good luck in reaching your goal.

Thanks !
cnc2.

[rokag3]

ευρηκα !!! eureka (in greek) I find it
So springy backslash effect disapear if you put the screw at 0.43 degrees off the axe !!! it seems that there is a sort of wheel alignment problems toe-in ? the fact is that the screw act like the road and the rings like a wheel so the screw as a tendancy to go on the axe of the rings if you are perfectly aligned then you will create a very strong slip effect that we interpret as "springy backslash "(thank for the word bill).
The solution might come from an assymetrical profile of the rings yet to be calculated or an offset of the screw.
I am going to post some picture just after this post

[rokag3]

here are the pics of the day. the 2nd pic is the lost strategy with the cam that work only on solidwork
this very concentrated guy is george the Mr. metal of the team while opening (one more time) the screw trap
We are working only on x axis that's why the y axis is not fitted 1st pic.
George is one of this crazy guy that make crazy work on big vessel in Pirea every big port has this type of very qualified mecanician that will save $$$$ to the shipping Cie. by making the impossible .
pic 4 the alignment of the screw is absolutely perfect (that is the problem by the way) it's hard to make george accepting to desaligning it. So we reach a compromise we will make it working with an offset and find the solution to make it working perfectly aligned.
pic 5&6 the pinions are one approach for security and self adjust of the pressure

[BillTodd]

Lucien,

I'm pleased you have found the problem

However, are you sure that mis-aligning the nut/screw is not simply pre-loading the thrust bearings? (see attached pictures)

[rokag3]

http://www.cnczone.com/forums/attach...8&d=1230893471[IMG][/IMG]
hello bill,
The point is that if we do not turn the screw then even if we push the table back and forth there is no backslash.
but if we turn the screw then the screw find a way to go out a little bit this make the springy effect.
the way we have made our roller does not allow any backslash the pre load of the bearing is really correct , i really think that we do not master one parameter in the design of the profile of the roller something stupid as the V angle of the rail of the train.
We are going to check your theory anyway