PDA

View Full Version : "Differential Thread"... What is the benefit?



Your Old Dog
12-31-2009, 08:45 AM
Someone mentioned the differential thread on another post and I'd like to explore it more as I've heard the term, made one but don't see the advantage?

Seems to me that there will always be some slop between thread and mate. When you double the threads and mates in a mechanical application I would think you would double the slop but these are used for refined applications if I'm not mistaken?

http://www.homemetalshopclub.org/news/jul99/thread.jpg

J Tiers
12-31-2009, 08:55 AM
Slop does not destroy the usefulness of a lead screw, and it won't destroy the usefulness of a differential thread either.

The advantage is a fine-thread movement without needing an actual fine thread. All slop-eliminating methods would be used to mitigate the effect of slop, taking it up etc.

But it won't be doubled..... it will have the proportional slop of each size thread, in that thread..... slop larger than the pitch is obviously not possible, for example, so a finer thread has to have less possible slop..

John Stevenson
12-31-2009, 09:05 AM
YOD,
I needed some fine threads many years ago for steering laser mirrors.
The originals had 1/2" x 40 tpi on them but that was still a bit course, the main problem was these being in constant use soon wore out.

I made new threads using the differential method [ which at the time I thought I had invented :rolleyes: :D ] using 1/2" x 20 running with 1/2" x 26 which are both common British threads in the Cycle series.

This gave me a fine action but with deep threads to prevent wear and after trialling these on one machine they were all fitted with them and none were replaced up until they were scrapped out.

.

Evan
12-31-2009, 09:18 AM
It's used on tap handles to increase the clamping force.

Here are examples used for metrology including micrometer screws and laser mirror adjusters.

http://www.photonlines.com/optosigma/screw.html

DannyW
12-31-2009, 09:29 AM
Hi all,

If the goal is to advance less, than the second thread has to be left-handed, is it not?
Say we start with a 2 mm RH thread that acts on a 1.5 mm LH thread, the outcome would be an .5 mm advancement after one turn. Right?

Regards,

Danny

Evan
12-31-2009, 09:49 AM
Here is a demonstration model. It should make it clear.

http://ixian.ca/pics7/diffthread.jpg

Ed P
12-31-2009, 09:54 AM
Hi all,

If the goal is to advance less, than the second tread has to be left-handed, is it not?
Say we start with a 2 mm RH tread that acts on a 1.5 mm LH tread, the outcome would be an .5 mm advancement after one turn. Right?

Regards,

Danny

The threads are given at "16 pitch" and "18 pitch" which is most likely in error. The poster probably meant 16 threads per inch (tpi) and 18 tpi.
If both threads are right handed and the screw is turned clockwise then the screw will move to the left. However the traveling piece (which I'm assuming cannot rotate but can slide left or right) will move to the right.
The screw will move 1/16" for each rotation (left) and the traveling piece 1/18" (right). Subtract the two and get the net movement of the traveling piece to the left.
Now if the 18 tpi were left handed then the traveling piece will also move left. The net distance travelled by the traveling piece will be 1/18 plus 1/16 (left).

Ed P

Your Old Dog
12-31-2009, 10:25 AM
Thanks guys. I understand how the concept works but it was the slop I had trouble dealing with and JTiers pointed out something I had overlooked. And that is the slop is taken up when the threads all mate up. I suppose in actual usage, this differential screw might adjust one item against another but a lock down screw would be necessary to hold all in place as the slop inherent in both sets of threads would come into play.

The durability issue that John points out was another point that had escaped me. The example Evan post was one I had not seen. I can understand how it might be used where the more conventional one (such as the optical) might as you could apply more force to it to move larger objects.

Has anybody made a triple differential thread? :D

motorworks
12-31-2009, 10:30 AM
I did post this once before... nice idea
I was going to use it as a feed for a line boring bar, but went a different direction thanks to a UK buddy ;)

http://www.youtube.com/watch?v=iPkoGug-oC0

digger_doug
12-31-2009, 10:30 AM
My fellow designer pulled the "differential screw thread trick" out
of his pocket, and used it on a inspection fixture.

The large gear product weighing around 5 tons was needed
to be finely (maybe 30 - 40 pitch) adjusted for leveling or elevation
using a screw on a c.m.m. fixture.

The screws were around 2" in diameter (also support the
"v" blocks in alignment) but had a very fine adjustment.

Which the inspector (c.m.m. operator) appreciates daily

DannyW
12-31-2009, 10:32 AM
Got it Ed!

And Evan's model does what? Given both RH threads and turning clockwise?

The middle piece goes to the right?

Right?


Danny

dp
12-31-2009, 11:54 AM
Got it Ed!

And Evan's model does what? Given both RH threads and turning clockwise?

The middle piece goes to the right?

Right?


Danny

Yes - and the handle winds in at the pitch of the left most thread.

barts
12-31-2009, 12:15 PM
I used this trick on a steam engine valve linkage coupling nut once; fine thread rod eye on one side, coarse thread rod from the eccentric on the other and both w/ jam nuts. Works very well, and is really simple; you cab adjust the valve timing while the engine runs (slowly).

- Bart

Black_Moons
12-31-2009, 02:04 PM
Hmm, I could sware my tap/die handles have diffrential thread, but its 'reversed' (Left hand/right hand thread) in that it gives me extra travel for less turns, Not more torque for more turns.

Its definately a cool trick.
as far as the 'slop' is conserned, its not a big problem if you can insure its allways loaded from one direction (One way movement stops, lifting parts, clamping parts, etc etc etc) Or just have an application where a little slop isent a consern, or if you used slop reduction methods like nylok nuts, etc.

The Artful Bodger
12-31-2009, 02:31 PM
.....and on a different but somewhat related (and I trust interesting) topic I recall reading once that Maudsely et al would parallel two lead screws to cut a new lead screw which would have only the the average error of the pair.

Evan
12-31-2009, 03:03 PM
There are two main errors to worry about with a lead screw. One is the obvious error of actually having the correct number of threads per inch. That is fairly easy to deal with and not as important as the other error. The other is the always unavoidable periodic error. This shows up as a variation from ideal in the distance moved once per rotation of the lead screw. It can also be a longer period depending on the gearing that drives the leadscrew.

There are a couple of ways to reduce the periodic error. One is to use a very long lead screw nut that will average out multi turn periodic error. The other is to use a very small diameter master leadscrew as that reduces the size of the one turn error in proportion to the ratio of the size to the copy screw.

The two lead screw approach can be implemented in the same manner as differential threads with one lead screw turning a captive but rotating main lead screw nut on the carriage as it is driven by the other leadscrew.

Paul Alciatore
12-31-2009, 03:35 PM
The "Slop":

Black Moon said it, we all use techniques to take out the slop or backlash or lost motion when we use standard lead screws. Most of the time we aren't even aware of what we are doing. A lathe cuts with some resistance to the inward and forward motion of the tool so it keeps the lead screws under pressure and effectively eliminates slop. Likewise on the mill and other machine tools which use lead screws.

Of course, a good fit and adjustments in the nuts or half nuts will help to eliminate it. Springs or the weight of the element being moved can also help. Differential screws would make nice "fine" levelers for a heavy machine. Or a "lock down" screw with spring washers next to the differential screw would allow very fine adjustments. I am surprised that these techniques are not used more often as the cost is minimal.

All of these techniques can be used with differential screws. It is important to design the mechanism that uses them to incorporate at lease some of them. Without loading and the other techniques you will, indeed experience "slop".

Arcane
12-31-2009, 05:11 PM
Years ago when they built the synchrotron here in Saskatoon (the only facility of its kind in Canada and one of just a handful scattered around the world :) ) a local machine shop got the job of making the leveling blocks for the machine to align the light beam. I worked with the shop owners brother and he brought one block to work to show us...just a threaded bolt through a block of steel. It was no larger than about 2 1/2" by 2 1/2" by 1 1/2" with the screw part being about 1" diameter..and 180 threads per inch! I was amazed that they would use a single thread for adjustment instead of a differential screw but I guess they had their reasons. This got me wondering what the guys here who do commercial work in machine shops would have chosen for this application. Would you use a differential screw or a single threaded block? I'm not sure how much weight each block supports but on the one tour I had of the place, it's obvious it's not a trifling amount.

Jim2
01-01-2010, 10:34 AM
Hmm, I could sware my tap/die handles have diffrential thread, but its 'reversed' (Left hand/right hand thread) in that it gives me extra travel for less turns, Not more torque for more turns.

Yeah, I think mine is made that way, too. I've always thought it to be an example of a cheap copy where the imitator didn't really understand what they were looking at in the first place. My tap handle annoys me no end. It's always coming loose because it's impossible to get it tightened properly. I consider it to be cheap junk.

Jim

Black_Moons
01-01-2010, 10:39 AM
Intresting I did'nt know it was 'backwards'.. Iv bought some other larger tap wrenchs and they have the same config.


180TPI? woah that must be a fine thread... iv seen 40tpi and it starts to look like a bad surface finish. I think my lathe maxes out at like 56tpi and I never thought anyone would need finer!

Paul Alciatore
01-01-2010, 12:42 PM
Years ago when they built the synchrotron here in Saskatoon (the only facility of its kind in Canada and one of just a handful scattered around the world :) ) a local machine shop got the job of making the leveling blocks for the machine to align the light beam. I worked with the shop owners brother and he brought one block to work to show us...just a threaded bolt through a block of steel. It was no larger than about 2 1/2" by 2 1/2" by 1 1/2" with the screw part being about 1" diameter..and 180 threads per inch! I was amazed that they would use a single thread for adjustment instead of a differential screw but I guess they had their reasons. This got me wondering what the guys here who do commercial work in machine shops would have chosen for this application. Would you use a differential screw or a single threaded block? I'm not sure how much weight each block supports but on the one tour I had of the place, it's obvious it's not a trifling amount.

Hmmm! 180 TPI is 0.0055" pitch. With an 80% thread, it is only 0.0044" deep. Add some allowance between male and female and you are down to perhaps 0.0035" engagement on each side; 0.004 at most. Not much room for tolerances or temperature changes or whatever. I would bet there was a good market for replacements as they wore out. Or were cross threaded. I would suspect that the choice of alloy was dictated by the requirements for a good surface finish.

Astronowanabe
01-01-2010, 02:54 PM
yet another use for differential threading is constant length, variable angle connection. two rods with different pitch threads and machined flat ends
are screwed onto a nut that is threaded half way with each pitch
(and a gap in the middle) then the rods are screwed in tight so they
butt they have a constant length, but loosen and change the rotation
of the nut. and it will change the angle of the second rod with respect to the first when re-tightened.

juergenwt
01-01-2010, 03:21 PM
You can in some cases use a standard screw and drill and tap an inside thread. Let's say you use a M12x1.5 metric fine and drill and tap a M8 (1.5)
hole. that would give you an advance of 0.25mm per turn if you insert a standard M8 rod. We used to take out any play by adding a compression spring on one side. All kinds of combinations are possible.

Arcane
01-01-2010, 03:40 PM
Hmmm! 180 TPI is 0.0055" pitch. With an 80% thread, it is only 0.0044" deep. Add some allowance between male and female and you are down to perhaps 0.0035" engagement on each side; 0.004 at most. Not much room for tolerances or temperature changes or whatever. I would bet there was a good market for replacements as they wore out. Or were cross threaded. I would suspect that the choice of alloy was dictated by the requirements for a good surface finish.

Yes, the 180 TPI is the one thing that stuck in my mind! Definitely fine as all get out! Since these were leveling/support blocks, once in place they would never move unless the light beam needed to be adjusted so wearing out is not too likely. I know they made a few extras for whatever reason and cross threading probably happened to a few of them. Any time I get feeling smug about what a fine thread cutting job I am doing, I just have to remember those blocks...and I get grounded pretty fast! :D

danlb
01-01-2010, 03:41 PM
When I first read about the differential thread, it was in a dispay of Leonardo Da vinci's work.

The stated purpose of the thread was to provide the rapid movement of the larger thread with the greater pressure applied by the smaller thread. The examples were in wood.

That same display also had a replica of a 15th century screw cutting lathe. It's amazing how far back these things go.

" invention of this type of screw was traditionally attributed to Hunter (1781). It is used to apply extreme pressure in machines such as hand and hydraulic presses. "
http://brunelleschi.imss.fi.it/genscheda.asp?appl=LIR&xsl=modello&lingua=ENG&chiave=100805


Daniel

Richard Hanley
01-01-2010, 05:18 PM
When I Was A Kid My Older Brother , Tom , Needed A Special Turnbuckle For Some Thing He Was Building. His Problem Was He Didnít Have A Left Hand Tap. He Wound Up Making One Side National Fine And The Other National Course.
I Don'tí Think He Knew What A Differential Screw Was , But I Know He Was Very Good At Solving Mechanical Problems.
Richard

darryl
01-01-2010, 06:43 PM
There have been a few references now about increasing the clamping force, applying a greater pressure, etc. I can not see how this can come about. Sure, you have more turns of the rod to achieve less relative motion between the nuts, but it's not akin to a geared system where there's a torque multiplication. The pressure against either nut would be the same for a given loading as it would be for a single nut on a threaded rod. The torque required to turn the shaft doesn't change either, per nut. It does in fact increase when you have two nuts on a differential thread coming together to clamp something.

All you're achieving with a differential thread is a slower relative motion between the nuts per turns of the rod. You get finer control of position, yes. Whether the accuracy of positioning is better with regards to the degree of turning of the rod- could be because you are getting some integration of the periodic error in the threads by having a longer total length of engagement, plus whatever might be gained by having the possibility of the periodic errors in the different threads skewed from each other-

By the way, someone asked if you should have one thread left handed and the other right handed- this will get you the opposite effect- there's a quicker change in the relative position of the nuts per rotation of the rod this way. A usefulness of this is faster relative motion while not steepening the helix angle of the threads. The nuts retain their self-holding ability better than if the single thread had to be made steeper to suit the faster nut motion. The smaller the diameter of the rod, the more this effect comes into play.

Then there's this talk about 180 threads per inch and such a small diameter of engagement between rod and nut- somewhere at some time industry must have arrived at some optimums. I wonder what those are? Coarse thread with greater diameter of thread engagement, vs fine thread with lesser diameter of engagement, but more threads per inch- add in clearances to assure any nut will thread onto any bolt (in the same sizes of course) and there should probably emerge some optimums. From what I've seen, I'd have to suggest that it's the NF format, or metric rough equivalents, but I'm guessing now.

Black_Moons
01-01-2010, 11:10 PM
Richard Hanley: Please do not capitalize every single word, it really makes it hard to read.

winchman
01-02-2010, 02:59 AM
A thread is an inclined plane wrapped around a rod.

With a differential screw, the effective slope of the plane is much shallower than that for either of the threads. For example: 1/4-20 is 0.05 per rev and 1/4-28 is 0.036 per rev. On a differential thread, they give you 0.014 per rev.

Therefore the same clamping pressure can be achieved with less torque than would be needed for either of the threads.

Your Old Dog
01-02-2010, 07:50 AM
I just ran across this site and thought some of you might find in interesting.

http://kmoddl.library.cornell.edu/model.php?m=71

Thanks for the responces. I thought it was a good thread and I learned a lot.

Asquith
01-02-2010, 08:09 AM
The ancient Chinese used a very similar principle in the Differential Windlass. This could give any mechanical advantage you wanted, simply depending on the relative diameters of the two winding drums:-

http://en.wikipedia.org/wiki/Windlass

If in doubt, go into the shop, find a piece of stepped bar in the scrap bin, take a piece of string, and make a windlass.

Your Old Dog
01-02-2010, 09:25 AM
Asquith, I'll see your Chinese Windlass and raise you another question. !!

Here we go again.

So, with the Chinese Windlass as described in your Wiki find, if you stop cranking at any point along the way with a bucket full of water, will the bucket just stay there or fall back to the water? It looks to me like it might just stay where ever you park it.

http://upload.wikimedia.org/wikipedia/commons/a/a8/L-differentialwinde.png

Lew Hartswick
01-02-2010, 09:49 AM
Asquith, I'll see your Chinese Windlass and raise you another question. !!

Here we go again.

So, with the Chinese Windlass as described in your Wiki find, if you stop cranking at any point along the way with a bucket full of water, will the bucket just stay there or fall back to the water? It looks to me like it might just stay where ever you park it.

Well it depends on the friction :-)

Seriously no. the torque is different on the two drums so it will
go back down . same force at different radii.
...lew...

andywander
01-24-2014, 10:13 PM
could someone post a link or a photo for a tap/die holder that uses a differential thread?

I don't believe I have ever seen one of those, and I am having trouble visualizing it.

Jaakko Fagerlund
01-25-2014, 01:23 AM
could someone post a link or a photo for a tap/die holder that uses a differential thread?

I don't believe I have ever seen one of those, and I am having trouble visualizing it.

http://www.cdxetextbook.com/images/tapsDies_tapWrench.jpg

Baz
01-25-2014, 08:52 AM
The Wohlhaupter boring heads use a differential screw to hold the arbor to the head. However it is counter clockwise to tighten.

Peter.
01-25-2014, 09:22 AM
Asquith, I'll see your Chinese Windlass....

http://upload.wikimedia.org/wikipedia/commons/a/a8/L-differentialwinde.png

The only problem I see with that device is it will use a lot of rope to move a short distance, so when winding it 'down' there will soon come a time when the smaller hub will be loaded with rope to the diameter of the larger one.

Ray Sidell
01-25-2014, 09:28 AM
Re-post of an instrument I made which uses the differential thread principle:

http://bbs.homeshopmachinist.net/threads/49033-Mighty-Mag-amp-differential-clamping?p=683745#post683745

EddyCurr
01-25-2014, 10:14 AM
Thank you for the link to your Mighty-Mag & differential clamping thread.

Perhaps I didn't see it at the time or have forgotten about it. Enjoyed
how you employed the differential screw concept in the bracket arm.
A candidate for the Shop Made Tools thread at the top of the General
section.

Nicely done.

.

andywander
01-25-2014, 11:59 AM
http://www.cdxetextbook.com/images/tapsDies_tapWrench.jpg

Where is the differential thread on that one? I only see one thread.....

dp
01-25-2014, 12:03 PM
The moveable jaw is threaded.

dp
01-25-2014, 12:41 PM
Here's a couple of interesting applications of differential screws: http://mechanicaldatahelp.blogspot.com/2011/09/special-screw-arrangements.html

dian
01-25-2014, 12:42 PM
what is a mighty mag? is that a magnet you turn on with a knob? there are two red knobs. why?

EddyCurr
01-25-2014, 12:58 PM
Yes, the lever is at one end. The plastic-headed knobs serve
to capture extension bars in holes cast in the mag housing for
this purpose.

.

J Tiers
01-25-2014, 04:11 PM
Yes, the lever is at one end. The plastic-headed knobs serve
to capture extension bars in holes cast in the mag housing for
this purpose.

.

Not on mine, not on any others I have seen.

You are thinking of the more usable ones (cheaper, too) that do have them, from Enco and HF, etc..

EddyCurr
01-25-2014, 06:03 PM
P/N 400-3 on the Westhoff site: "Mighty Mag w/ Quick Release"

https://www.westhoffinc.com/measuring-and-inspection/mighty-mag/

.

J Tiers
01-25-2014, 07:12 PM
Well, if that is "improved", I'll pass. IME they don't stick well enough to need a lever :D

I wish they had put a couple more mounting points on, when they made mine. I think only the one with the "V" is useful, and the lever seems to block it now. The other allowed my indicator stem to wobble sufficiently to render it worthless.

I tried the differential thread method once for making a pull-in/push-out screw for collets and other spindle tooling. It was more hassle than it was worth. It didn't really work much better than a regular drawbar. And there was a lot of turning to get the thing released.

I concluded that it would take more thought than I had given it, to make a system that worked well.

garyhlucas
01-26-2014, 10:48 PM
When I was younger we had a differential chain hoist. It had two chain sprockets cast as one piece with one sprocket having one more pocket than the other. The hook had a sprocket that supported it and the chain was one continuous loop. You pulled on the loose loop and as the load was raised that loop got longer and longer. It was self locking, the load stayed wherever you stopped and it didn't drift. It worked quite well.

Guido
01-26-2014, 11:42 PM
Diff screws mechanically equal a hydraulic pressure intensifier?

J Tiers
01-26-2014, 11:46 PM
When I was younger we had a differential chain hoist. It had two chain sprockets cast as one piece with one sprocket having one more pocket than the other. The hook had a sprocket that supported it and the chain was one continuous loop. You pulled on the loose loop and as the load was raised that loop got longer and longer. It was self locking, the load stayed wherever you stopped and it didn't drift. It worked quite well.

Yep, I have 2 or 3 of them. They do work well, and they do hold the load. I like them better than a 2 chain hoist for general use within their ratings (mine are half ton or so).

boslab
01-27-2014, 05:28 AM
Im sure ive seen machinists jacks with diff screw arrangement, they were big buggers as the wothpeice on the mill was 14 tons of copper casting, i cant seem to remember a lot except when the guy was jacking the handle was mowing disproportionatly to the amount the work was.
Mark

J Tiers
01-27-2014, 08:33 AM
i cant seem to remember a lot except when the guy was jacking the handle was mowing disproportionatly to the amount the work was.
Mark

Well, that is the principle of mechanical advantage.... you trade a large force applied over a small distance, for a smaller force applied over a larger distance.

There is no free lunch, it's the same energy either way. The mechanical advantage ONLY lets you do the job even though your amount of "push" can't move the thing directly.

krutch
01-30-2014, 02:42 PM
It works just great in a boring head I have that adjusts in .0002" increments.