worm drive backlash quantification 2

[Slice]

Hello all,

GOAL:
I am looking into using a worm drive to set the angular position of a turntable. The desired angular postion accuracy is about 0.1 degrees.

BACKGROUND:
The vertical shaft of the table can be as small as 0.5 inch diam (small components), the system will have a hand driven knob attached to the worm shaft (forces are not a problem), high friction is fine because this sytem is not transmitting power...

QUESTION:
My question is, how do I quantify(estimate)the backlash in a worm drive so that I can design the system to the accuracy requirement?


any tips would be much appreciated,
Cheers
Scott

 

[diamondjim]

use an adjustable backlash
design to allow you to
end up with the minimum
backlash to keep it operating
and to allow for wear.

 

[EnglishMuffin]

Expanding on diamondjim's suggestion, here are some of the methods of varying the backlash that I have seen used, or have used myself :
1. Simple radial adjustment of worm shaft center distance.
Depending on the type of worm and amount of adjustment, this can have problems in some cases.
2. Eccentrically mounted worm shaft (rotating the eccentric mounting changes the center distance). Depending on the type of worm and amount of adjustment, this can also have problems in some cases.
3. Two piece worm. The pieces can be spring loaded against one another or fixed-adjustable. Worms of this type can sometimes be purchased ready made from certain manufacturers. In the case of spring loading, this design can provide a zero backlash preloaded worm drive.
4. Variable pitch worm. This allows backlash elimination by axially adjusting the worm. It won't work for double enveloping worms.

 

[EnglishMuffin]

Since my post seems to have piqued your interest, here is some more information :
There are at least two manufacturers of split worm gears that come to mind. One is Cone-Textron - which is a double enveloping type. The other is the Ott (German), obtainable through Advanced in Rockford Il. Both of these may only be available in sizes which are a little big for what you need. The Ott is very high precision, and expensive. Another approach might be to consider using a spiroid gear set, obtainable from ITW. These can be adjusted for minimum backlash without compromising the geometry, which is not the case with radial adjustment of single enveloping worms. Although they look somewhat worm-like, Spiroids are not really worms - they fit somewhere between bevel gears and worms, but are very good for positioning drives. Both members can be made of steel, which is not the case with wormwheels, which are usually bronze, and axial movement of the tapered "pinion" enables backlash to be adjusted.

 

[JRCD]

Single and double enveloping worm gearing must run at the design center distance. To fix the backlash I suggest sliced and adjustable worm wheel or two axial adjusted worm gears, usually positioned 90 degrees apart (as used in table drive of some gears hobbing machines).

Small backlash demand high-precision gears. Backlash-compensating mechanisms are used only to counterbalance the wear.

 

[carburized]

Slice

You can do this easily by variable pitch worm. Variable pitch worm has thicker tooth at one end and keeps getting skinnier to the other end. So, when you need to adjust the backlash, you just move the worm in or out as the case may be. By turning gear comes in contact with thicker or thinner tooth and adjusts the backlash. We have been making adjutable backlash worm gear drives and they have been operating in the field successfully. look at our website

http://www.winsmith.com

Ravi

 

[EnglishMuffin]

JRCD : Your last statement is not quite correct. Both Cone and Ott zero backlash split worms can be arranged to eliminate all backlash due to worm wheel pitch errors and can be preloaded. This can be achieved either with springs or hydraulically, but it does considerebly reduce the efficiency. Your first statement is correct, but an adjustable center distance can be a cheap way of achieving the desired bore center accuracy in some relatively crude cases.
Variable pitch worm works fine - only drawback in some cases is that it cannot be double enveloping and cannot reduce the backlash below the pitch errors in the wormwheel. For the .1 degree you are talking about, any of the methods discussed can be made to work initially, but some are more forgiving than others over time. Since worm wheels are normally bronze and cannot usually back drive with low leads, one other thing you must bear in mind is that they are very susceptible to local crushing if you get an unexpectedly high reverse impact load (abuse). So if this is likely to happen, give greater consideration to designs which remove all the backlash due to pitch errors.

 

[involute]

we have used a split worm gear method of controlling backlash by splitting the worm gear in 2 and adjusting the tooth offset to control the back lash
true only half the tooth is in full contact with the worm
but it is good for indexing purposes .
an eccentric worm design to vary the center distance is good but is more difficult to make
Just so a tapered worm is also difficult to make with the
adjusting mechanism

 

[JRCD]

EnglishMuffin: Thank you for comments.
Slice: I come again to the backlash subject.

_ Why is backlash necessary?
Backlash is necessary to supersede:
a) the worst combination of gear meshing tolerances (including assembly center distance) and errors;
b) the operational strain deformations;
c) the thermal effect over dimensions;
d) the thickness of the lubricant film.

_How the above mentioned occurrences must be considered at the Slice case?
a’) The meshing pair total composite error test show how gear quality act over the backlash;
b’) In this case there are small forces, so are no significant deformations to be compensated;
c’) Temperature variations promote relative dimension changes in the gear box, worm and worm wheel, mainly if dissimilar materials are used. In this case dimensions are small and the heat generation by sliding is negligible, so that rest only the small environmental temperature variations. and for such case the necessary backlash is minimum;
d’) With manual worm drive (small speeds) and small forces gears can run with solid lubricant impregnation (molybdenum sulphide or graphite).

The above evaluation show that in the present case gear accuracy is the most important factor to be considered.

_Is small backlash necessary for accurate positioning?
Yes, but this asks for good quality gears. But in the present case precision positioning is as accurate as the meshing total composite (single flank) error.This can be attained even if there are larger backlash. To achieve this the first thing to do is to turn the worm in the opposite sense wanted for enough amount .to take off the backlash when regressing to the starting point, after this, go on to accomplish the wanted positioning motion without any torque relief.

_How many backlash are there in the gear meshing?
At power transmission the usual backlash is many times larger that the amount necessary for compensation of meshing total composite (single flank) error so this last factor is worthless. In this circumstances under practical view point is valid refer to only one backlash. In fact under influences of meshing total composite error the backlash is different for each meshing position, and the minimum, the average and the maximum backlash can be distinguished.

_Are backlash compensating mechanisms the solution for precision positioning?
Mechanical, pneumatic or hydraulics backlash compensating mechanisms can be used but even if correctly applied, they can take off the loose motion but never can diminish the meshing total composite (single flank) error and her influence on the positioning accuracy.

_At last.
May be there are commercial stock products that fulfill Slice requirements.

 

[EnglishMuffin]

Some general comments :
For precise positioning, very low or zero backlash is essential. But whether or not you need very precise gears depends on how you are measuring the position. If the position indication device is on the output gear, and you have a backlash elimination system, the pitch error does not necessarily matter. I have designed worm positioning systems which achieved less than six arc seconds using worm wheels which had nothing like that accuracy.

 

[JRCD]

Involute (Aug 25):

Really with splitting worm gear only half the tooth is in full contact. In my opinion the best method for backlash reduction is that using two worms, one at least axially adjusted.

Backlash reduction by center distance decrease is not good solution. Firstly because put in contact worn with not worn flank surfaces. Secondarily, throat-form radius surfaces, in single or double enveloping worm gear no more contact well. If the transmission shaft angle is not orthogonal other troubles will arise.

Eccentric worm adjustment gives the same kind of problem

When a gear set wears out all pitch remain unchanged but worm gear teeth become thin. Axially adjusted worm with increasing teeth thickness along the shaft direction can be used for backlash adjustment. I don’t recommend variable pitch worm for this purpose.

EnglishMuffin (Aug 26)

I agree with you “For precise positioning very low or zero backlash is essential”. I suggested manual backlash take-off for manual drive worm when there are no better solution.

Yes, if position indication device is on the output shaft the gear precision does not necessarily matter. Otherwise single flank meshing total composite error (not only the pitch error) must be so small as possible.

The 0,1 degree Slice asked accuracy has a size dependent difficulty and corresponds , per example, to:
0.0017 inch on two-inch diameter circle;
0.0105 inch on twelve-inch diameter circle.

 

[EnglishMuffin]

JRCD: Yes, you can use two worms of course, and at one time that was the only worm design that was used for totally zero backlash. But it's a very expensive solution. I don't know how familiar you are with split-worm drives, but I can assure you that they are currently being used very successfully on thousands of NC machine tool positioning tables, particularly German ones. The ones made by OTT have a special fine pitch tooth form which increases the number of teeth in contact. The split-Cone type made in the US were first used on the Sundstrand Omnimil machining center over thirty years ago, and in that version the split was not exactly in the center. The "number of teeth in contact" argument is less significant with double enveloping worms like the Cone, and of course you can always use a larger worm - it's still much more cost effective than two completely separate worms and drives. It is interesting to note that when first approached about making a split worm, Cone refused to do it, citing arguments not unlike the ones you are making about changing the center distance on single enveloping worms. They were eventually persuaded to do so, but would not guarantee it - and for many years they made the design only for Sundstrand. In recent years, however, they have started to market split worms as a standard product, with the split exactly in the center.
Your comments about adjusting the center distance are technically correct with single enveloping worms. However, if a worm is used in conjunction with a conventional gear, which can be done, they are not. Since Slice has not told us much about this application, it is not clear how important all the various factors are - he could be talking about a very lightly loaded drive or a power drive. You have to consider each individual case on its merits - there is no single "correct" solution.

 

[JRCD]

EnglishMuffin
Thanks for yours comments. I agree with you all-in.

 

[JRCD]

EnglishMuffin

I would like to send more comments.

For backlash reduction, split worm gears are as good as two worms axially adjusted. In fact split worm gears are more compact and cheaper, but the unique drawback is the load capacity reduction unless the design provides larger width worm gear but this asks for larger size worm diameter and so less gear transmission efficiency. If the worm turns only in one direction (clockwise or counterclockwise) it’s a smart design solution to adopt a larger worm gear split width for that part making the active flank contact. Probably this was the target of Sundstrand Omnimil design.

 

[EnglishMuffin]

Well, I don't quite know why the split was done that way on the Sundstrand Omnimil, which was definitely bi-directional - even the guy involved in the original patent could not remember when I once asked him. The split was 2/3 to 1/3. In fact, in that case there is so much preload across the split (well over a thousand pounds) that directionality was not really an issue. It's only practicable to have an off-center split with double enveloping worms of course, but one important advantage is that it improves the bending strength of the worm shaft - the two parts are splined together and because one piece goes inside the other it considerably weakens the shaft. I have seen heavily used bronze worm wheels of this type come back from the field with tooth thickness reduction visible to the naked eye - but still providing zero backlash in all positions. With this amount of wear, discussions about "tooth form" etc become rather academic of course.