Mark,

I agree. The rationale is as follows. To make one full turn of the rotary table takes 90 turns of the handle. To divide a single turn of the table into 53 equal parts we need to turn the handle by 90/53 each time. Since 53 is a prime number we cannot simplify the improper fraction. However we can express it as 1 and 37/53. If we have a 53 hole plate then it is simple! Here's a picture of a 63 and 69 hole plate I made for my dividing head, to cut 63 and 69 tooth gears, because I was too idle to set up the gears for differential indexing:

Remember that any errors in the dividing plate will be divided by 90, which is good.

Regards,

Andrew

Hello Mark,

Your calculations are correct, 53 hole plate, 1 full turn and 37 holes, will give you 53 divisions on a 90-1 rotary table.

Here is a picture of a 95 tooth 6DP gear I cut last week, on a rotary table.

Richard.

Edited By RICHARD GREEN 2 on 14/11/2012 09:44:28

Sorry to sound thick but if you need to make 53 tooth gear and can cut 53 hole plate well whats the problem?.

but if you mean 37 hole plate ok but how is that easier?

sorry but not completely familiar with RT and dividing heads

90/53 ok get that

90/53 = 1 + 37/53 — ok here too

how does a 37 hole plate help..?

If you only want one gear it is hardly worth making a dividing plate. As Jason has pointed out the angular rotation between teeth is 6.792 degrees = 6 degrees 47 minutes 31 seconds. Just make a table of the angular rotation required for each tooth ie 6.792, 13.584, 20.376…….etc until you hav done a complete revolution. Convert this table into degrees, minutes and seconds. Then just increment the rotary table to the required angle for each tooth. It requires a little concentration but is perfectly doable.

Mike

You can't do that with a rotary table as you need to take into account the 90:1 reduction, going one hole at a time would give a 4770tooth gear.

If you were using teh plate for direct indexing on say the lathe spindle then it would work

The OP says that his rotary table has provision for dividing plates, so why not use them? Also, we do not know the diameter of the gear to be cut; the bigger it is the more accurate the division need to be.

I will be cutting my traction engine final drive gears on a rotary table, as per Richard's setup, as at 14.8" diameter they will not quite fit under the milling machine arbor. I am budgeting for an error of 2 thou at that diameter and will compensate for it by cutting the teeth slightly deeper. That's accuracy of course, not resolution, which is quite a different thing.

Regards,

Andrew

…. oh, I don't think so. 5 x 700 nm ?

You can make an approximate plate for reuse to make a better one using any existing plate, or a change wheel by the multi-turn method.

Say you have adapted a 60 tooth change wheel to fit the rt which gives 60 x 90 = 5400 steps all round the circle.. You want 53 so you need to count 5400/53 steps = 101.887 steps. But you can't step a fraction. So miss the first tooth (or hole) and see if you can do the second ie 2 * 101.887 = 203.774.

Still no good but keep going and you get to No 9 = 916.981. Ok so do 917 steps and you have only a slight error. Repeat the process and you eventually fill in all the teeth with a bit of an error by going round multiple times. the error is spread all over the target not built up progressively. But when you use that as the index device the error gets reduced.

This is what the stepper motor dividing gizmos do. Some combinations work better than others but basicly any existing plate can make any new plate.

You're right. I had about 450nm in mind as a typical 'wavelength of light', so perhaps I should not have guessed the colour.

Even so, it always surprises me that the difference between the small you can relate to (i.e. working toa a thou or less) and the small you can't quite grasp (wavelength of light, size of animal cells or even bacteria) is not really very much at all.

Neil

From my old Machinery's Handbook tooth-tooth tolerance barely increase as gears get bigger, but gets bigger more or less in proportion to tooth size.

It will upset the clockmakers, but the recommended gear quality for clocks is in the lowest category with milking machines and kitchen blenders, below hair and hedge clippers, with gun data computers and gyroscopes at the top.

Neil

.

So will there be a difference if you cut the gear in daylight, at night with a florescent light or under under the bed covers with a torch.

Wouldn't want to get it wrong you know.

John S.

HI Andrew,

I didn't measure it, I made the rash assumption my home-cut worm and wheel are accurate, and then calculated the theoretical error for being half of one division out.

All the gears I have made to date using this method of divising with Ivan Law's book beside me seem to work fine.

13T pinion with splined insert

My gear cutting setup, using home made rotary table/dividing head:

A pair of bevel gears made using the above equipment:

A 63-tooth metric changewheel for my mini lathe:

Regards

Neil