[lawrencex2]

I know this has been talked about many times but I just wanted to share my design. Please excuse the design picture becuase i did this at work without any cad.

My advanced yet semi simple design to fix backlash. I have seen a few types of backlash compensation.

The two that are very common are:

1. Spring between nuts. (this will auto adjust backlash as it happens but the problem is the backlash only as good as the spring used.. Also with high spring compression comes excessive ware on the nuts.
2. Screw to adjust backlash. (this can be very rigid fix but problem is as wear happens you have to re adjust the screw/s to compensate.

advances not increases...


This design has both of the good features of the two designs. It automatically adjusts for backlash and is rigid because the screws are holding the nuts apart not a spring. This also reduces wear because only enough force is applied to the screws to turn them.

[rowbare]

Clever idea but I am not sure how well it would work in practice. The problem is that your adjustment is unidirectional i.e. your nuts can move apart automatically but cannot move back together.

Take this scenario with the initial condition that your screw is 10 tpi and two nuts are exactly one thread (.100) apart where they are threaded onto the screw. Now let's assume that the screw is a bit inconsistent so as you go down the screw the distance between the thread crests varies some. So there are sections where the threads are perfect, sections where they are slightly too far apart (.105) and sections where the threads are too close together (.095). This can be due to the way the screw is made, wear, or both. Well too close together can't be because of wear but you get what I mean. If this weren't the case, this entire discussion would be unnecessary.

Now as your nut goes down the screw, it hits a section where the threads are a bit too far apart and the gap between the nuts expands and the screw takes up the difference and locks in the new gap. So far so good. However at some point, you have to hit a section where the threads are closer together. But your nuts are set for the greatest gap and the nature of your adjusters is that they can separate the nuts but can't bring them closer together. Therefore your screw will jam.

This is why the spring works so well: it can dynamically adjust the gap between the nuts but at the cost of some wear as you have pointed out. As well there is almost nothing simpler and more reliable than a belleville washer.

bob

[lawrencex2]

Good point.. But if this is made from impregnated derlin the wear on the screw would be nearly non existant.

It also depends on the accuracy of the screw used. I am not 100% sure, how screws accuracies are measured. Let’s take a screw that has an error of 0.0003 in/in if im correct then the most it could be is +-.0003 and that is much less pressure then having a very large load with a spring.

[Bubba]

Yes, but the accuracy on some rolled screws is as high a .009"/ft and that is magnitudes of order beyond your indicated spec. As for pressure, when I did my ballscrews, I seem to remember the pressure of preload was on the order of 275# for my 5/8" ball screws.

I set them to that and have had no real problems from spring pressure in 10 years.

[lawrencex2]

I have never used ball screws. I have only used Nook leads for my projects. If i was to put 275# preload on them the nuts would have a very difficult time moving.

Plus most commercial anti backlash acme nuts have a maximum of ~10# preload. With my design, the preload is very small and will still maintain rigidity. Even if it adjusts a small amount, I do not think it will be too much loading between the nuts.

Even at .00075 in/in or .009 in/ft I do not think that it can produce enough force it to stop.

I would like to try to calculate the pressure that would be introduced when it adjust for errors in the lead. But, unfortunately I have no clue on how that could be done. I could probably test the toque needed to turn the lead with a prototype and 3’ of lead using some sort of gauge. Need to do some thinking about that setup.

[rowbare]

If you are using Delrin for the nut, then you don't need all of the fancy stuff. Mold the nut to the screw and let the compliance of the Delrin take care of the small pitch errors.

Check out this thread on HSM: Making Acetal leadscrew nuts the easy way - The Home Shop Machinist & Machinist's Workshop BBS

bob

[lawrencex2]

If you are using Delrin for the nut, then you don't need all of the fancy stuff. Mold the nut to the screw and let the compliance of the Delrin take care of the small pitch errors.

Check out this thread on HSM: Making Acetal leadscrew nuts the easy way - The Home Shop Machinist & Machinist's Workshop BBS

bob

Thank you rowbare for this gem!!! As soon as I got home from work yesterday, I tried this out.

It does work with impregnated plastic but probably not as good as the regular duron(black stuff).
- Even though the two halves fused, I noticed allot of problems with the threads not fusing properly.
- The plastic seemed to be more brittle after this process.

Other than those problems, it worked beautifully! Tried it out last night and my Y axis has less than .0003 backlash which is very acceptable for me at the moment.

I think I’ll still do the design just to test proof of concept.. Plus I like to tinker and see if I can get my ideas to work.. Most do not, but when it does what a feeling of accomplishment.

Once again thank you for the great link.

[rowbare]

That might be the nature of the plastic you are using. Some get brittle after melting. Other possibilities might be heating the screw too aggressively and overheating the plastic. Or the screw acts as a heat sink and cools the thin cross section of the threads too quickly.

If you have an extra length of screw a couple of inches longer per side than the length of your nut, you might try pre-heating the screw thoroughly to just under the melting point of the plastic, then clamping it and continue heating it gently from both ends till it melts and fuses.

Let us know how your proof of concept works out.
bob

[mike_Kilroy]

a slight variation on what u r doing: use a spring but NOT in series between the 2 nuts pushing them apart, but rather in series with them; so as u move, it ratchets the two nuts against each other and readjusts for wear as it happens, keeping both nits against opposite sides of backlash with very low spring tension?

[lawrencex2]

a slight variation on what u r doing: use a spring but NOT in series between the 2 nuts pushing them apart, but rather in series with them; so as u move, it ratchets the two nuts against each other and readjusts for wear as it happens, keeping both nits against opposite sides of backlash with very low spring tension?

Unfortunately that wouldn't work... it would be the same as the spring pushing nuts apart.. it would need heavy spring load to support the loads.

if you do a drawing you will see as load goes from left to right the only thing keeping the backlash is the spring.

[mike_Kilroy]

Actually it does work. So well it is patented by BSA (http://www.thomsonlinear.com/website...ews/xc_adv.php)

Unfortunately I did a poor job describing it. But for a home brew machine it may be a good make-it-urself solution? Actually we sell the XC nut for around $25-100each depending on size too. Let me know if you want a quote

[lawrencex2]

OK i think i understand now.. so basicly it would be something like this NUT NUT (Spring to press against nut) as the rod spins it will cause the second nut to turn against the first nut and not away because of the pressure of the spring.

Do I have this correct?

That brings me to another idea.. will post a pic asap..

[mike_Kilroy]

that was the idea - to give u more ideas

sorta but.... what I forgot to mention is yes, there is a spring between the two nuts BUT the KEY is that metal piece between them: the spring is under it of course but it RATCHETS the metal piece with tapered ends against the 2 nuts to LOCK them in place - no overcoming the spring by your load! This metal tube holds the nuts RIGIDLY apart at a solid distance..... it only ratchets further when wear causes backlash so the spring then moves it a tad more and pushes them apart again. self regulating... cool huh?

[lawrencex2]

Ok this one is much different and can be done pretty easy.. This one will have two nuts with a ratcheting surface on the inside. It would use a cheap spring with some bent tabs that would contact the ratcheting surfaces. It may only need a few of these 3-5.

To use you would put first nut on rod but not fully then insert spring then spin the second nut on top of first to preload. After preloading fully engage both nuts onto the rod.. This would cause the nuts to spin into each other as backlash happens..



*both nuts have the same racheting serface.

hmm now that i think about this design.. im not sure if it will work 100% of the time..
First nut is mounted to machine.. 2nd would be free.. It would be fine going one way but I’m not sure when it reverses the direction.. I know the spring will keep the 2nd nut in place but how well is the question. This will need some experimentation to see what happens. And its a prety simple design so ill give it a test some time soon.. Unfortunatly my experments will have to wait becuse i have a new job for my machines.. and money is more important then tinkering right now.

[lawrencex2]

that was the idea - to give u more ideas

sorta but.... what I forgot to mention is yes, there is a spring between the two nuts BUT the KEY is that metal piece between them: the spring is under it of course but it RATCHETS the metal piece with tapered ends against the 2 nuts to LOCK them in place - no overcoming the spring by your load! This metal tube holds the nuts RIGIDLY apart at a solid distance..... it only ratchets further when wear causes backlash so the spring then moves it a tad more and pushes them apart again. self regulating... cool huh?

Ok so something like this [/ /_\ \] ...

NUT tube NUT both nuts have a slight taper and the tube has also a slight taper and spring pushes up on the tube to keep the nuts apart. The casing will support the spring and also keep the 2nd nut from spining.

[rowbare]

Here is a link to the patent for the nut Mike mentioned: Backlash compensating assembly - Google Patent Search

It lists a lot of prior art so I am sure you will find tons of ideas there.

bob

[mike_Kilroy]

sent u PM.

[lawrencex2]

Actually it does work. So well it is patented by BSA (http://www.thomsonlinear.com/website...ews/xc_adv.php)

Unfortunately I did a poor job describing it. But for a home brew machine it may be a good make-it-urself solution? Actually we sell the XC nut for around $25-100each depending on size too. Let me know if you want a quote

I realy liked there idea and design.. but the 3/8 is 0.82" and that is too tall I think.. Right now i have to make sure I dont go over 0.7" and I think max is ~0.75.

[rbocaz]

I know this has been talked about many times but I just wanted to share my design. Please excuse the design picture becuase i did this at work without any cad.

My advanced yet semi simple design to fix backlash. I have seen a few types of backlash compensation.

The two that are very common are:

1. Spring between nuts. (this will auto adjust backlash as it happens but the problem is the backlash only as good as the spring used.. Also with high spring compression comes excessive ware on the nuts.
2. Screw to adjust backlash. (this can be very rigid fix but problem is as wear happens you have to re adjust the screw/s to compensate.

advances not increases...


This design has both of the good features of the two designs. It automatically adjusts for backlash and is rigid because the screws are holding the nuts apart not a spring. This also reduces wear because only enough force is applied to the screws to turn them.

Something like this:

[handlewanker]

One thing you are overlooking.....the screw HAS to be perfect in it's pitch, not only at the beginning but all along the thread.

Any variation in the pitch due to accelerated wear from the pressure of spring loaded compensating nuts and you have a variation in the pitch randomly.

Starting off with a hardened and precision ground screw thread, (Acme form?), you then add a pair of bronze nuts with a spring or any other form of anti backlash design, which would be the most logical approach, but precision hardened and ground Acme thread costs the same if not more than a regular ball screw, and even if'n it were marginally cheaper will not work better than a ball screw.

You are up against a sparsely lubricated sliding wear face in the bronze nut system and without pressurised oil to keep the metal to metal bronze nut/steel screw faces apart the wear will take place at a continuous and progressive rate.

Eventually the middle or most used part of the screw will wear more than the ends, and then the spring loaded compensator will work overtime adjusting for the pitch variation, but as it's putting a pressure against the screw flanks it also accelerates the wear exponentially.

Spring loading creates a drag, while a compensated fixed system locks up on the unworn part of the thread when the nut has taken up the slack.

You won't find commercially off the shelf hardened and ground Acme thread, so when you are left with a relatively soft steel screw, the wear factor becomes even more pronounced when you thrust a bronze compensator against the reverse screw flank and pre-load it.

The other problem is that unless the spring loaded compensator is powerfull enough to resist the thrust of the screw in reverse drive, it will give way until the spring resists the force applied against it from the reverse side of the nut, and on a soft steel screw this interprets to increased and uneven wear, usually in the middle section of the screw.

What is being proposed is the perfect world situation where the screw remains in pristine condition all along it's length, (no pitch variation whatsoever), and the bronze nut as it wears, progressively adjust to take up the backlash as it occurs, whilst maintaining a minimal running clearance.....all this without pressurised lubrication.

The roller nut design has been described and experimentd with in another thread on this form to overcome this problem of wear and backlash.
Ian.