Thread: 0.3 Module Gear Cutting Query

I have not actually read the thread as it is too long for my attention span, but I just thought up an idea for a quick indexer besides Pete's good idea of making one out of plastic...

Make up a suitable piece of round material... Put it on a mill with DRO... Drill the required number of holes... Put the round piece back on the lathe and turn down the OD until only half the holes remain.. deburr, then you have a "gear" with "teeth" to put a locking mechanism in..

Originally Posted by .RC.

I have not actually read the thread as it is too long for my attention span, but I just thought up an idea for a quick indexer besides Pete's good idea of making one out of plastic...

Make up a suitable piece of round material... Put it on a mill with DRO... Drill the required number of holes... Put the round piece back on the lathe and turn down the OD until only half the holes remain.. deburr, then you have a "gear" with "teeth" to put a locking mechanism in..

Yeah, but you still have to stand there and layout 107 holes IIRC Bob is counting turns to do so too!!!

Bob

You could make a simple dividing plate easily by turning a piece of MDF/Ali to a diameter of 170.227mm & then use a tape measure around the circumference to mark 5mm increments.

Regards
JohnQ

Originally Posted by Pete F

Oh BTW Michael, I don't think this will work in the way I think you're imagining. You can't put in 2 rows of 54 and just leave one out at the end to get 107.

I was only using 54 as an upper bound on the space/ length needed

Originally Posted by Stustoys

Except it wont work like that

I think you'll find the first and last holes are on one circle, one pitch apart.

It works - it's just not elegant.
BC.jpg

Originally Posted by Pete F

... However there is no requirement for all the holes to be on the same radius, only that they are the same angular distance apart.

Originally Posted by Pete F

... AND it's plastic Michael

Hmpf! It 'aint real machining if it doesn't make swarf!

Michael

Originally Posted by Michael G

It works - it's just not elegant.
BC.jpg


Michael

Michael, you can see that 54th hole is interfering. I guess in that regard it would work, but it's beyond not being elegant, and I certainly wouldn't do it. There is no real disadvantage to dropping that odd hole down to a third radius and getting him out the way.

Hmpf! It 'aint real machining if it doesn't make swarf!

Yes but if I'm printing PLA is smells like maple syrup on pancakes, so in that regard it's no contest

Originally Posted by Pete F

...you can see that 54th hole is interfering. I guess in that regard it would work, but it's beyond not being elegant, and I certainly wouldn't do it. There is no real disadvantage to dropping that odd hole down to a third radius and getting him out the way.

True. For a one off I'd probably risk it (maybe smaller holes, larger PCD to get some space). For a repeat job I'd probably do a third row - actually, if I had a third row in there I'd probably set the plate up with 35 or 36 holes per row. If you have to move the index pointer in another diameter, may as well make it useful.

Originally Posted by Pete F

Yes but if I'm printing PLA is smells like maple syrup on pancakes, so in that regard it's no contest

Don't tell my better half - she loves that stuff ever since a trip to the US. She'd want me to print everything just for the aroma

Michael

Originally Posted by JohnQ

Bob

You could make a simple dividing plate easily by turning a piece of MDF/Ali to a diameter of 170.227mm & then use a tape measure around the circumference to mark 5mm increments.

Regards
JohnQ

A probably insurmountable problem for me John would be the accurate marking of the increments and ensuring that the marks were uniform. The need for uniformity becomes apparent
when trying to centre the mark with a 20X microscope. The attached photo is of the finest line our printer at work can achieve. It is from the corner of the drawing I posted a photo of earlier. Fine to the naked eye......

DSC_4862 (Large).JPG

The "line" width is approximately 0.08mm or 0.003". Finer than anything I reckon I could achieve with any consistency 107 times.

Now some might question this need for accuracy. The exceptionally smooth* fine focus control features easily divisible 0.002mm graduations. The machining of a replacement gear needs to be accurate to maintain the smoothness. The scope's maximum magnification is 1000X with the present combination of oculars and objectives. A lack of smoothness at that magnification would be um.. undesirable.

Bob

* until it encounters the split.

Bob I always like the way you do things "right", and clearly this is no exception. However I can't help but wonder if you're not overly concerned about the accuracy of this part in regard the indexing. As an exercise you may want to calculate the angular difference in the gear teeth with an indexing error at the circumference of the indexing plate. I haven't done so, but will be extremely small.

The gear you are cutting won't be hobbed, therefore it won't be true involute and only an approximation. On the other hand it is nylon originally (and if smoothness is a concern, that's what I'd replace it with too), and is naturally somewhat pliable. It also naturally moves and changes shape slightly with differences in humidity. The bottom line is, in my opinion, the smoothness you seek won't be affected by indexing errors in your dividing plate of the magnitude we're talking about here.

Originally Posted by Anorak Bob

The "line" width is approximately 0.08mm or 0.003". Finer than anything I reckon I could achieve with any consistency 107 times.

Originally Posted by Anorak Bob

A 0.5mm error at a radius of 300mm equates to about 0.03mm at 18mm radius.

Given you shouldnt have to much trouble picking that line up within 0.04mm thats 0.0024mm!!

Though if you use your rotary table and take the 5 second error thats 0.00043mm

I really think plates are about speed and reducing errors than outright accuracy. Though at tooth 100 I think you'd be feeling the pressure

Hey,

Gotta admit some of the discussions over my head.

Have been wondering if the very fine tools used for clock and watchmaking might be useful for this fiddly small 107 tooth gear.

Ie this thing is a DIY version of an $8,000 dollar clockmakers gear cutter,

http://www.projectsinmetal.com/free-...r-clockmaking/

I guess it still comes back to the accuracy of the indexing wheels.

Bill

Woodworkers sometimes use jigs where a small distance between two things is used again and again to repeat a measurement.

https://www.google.com/search?q=fing...w=1440&bih=904

COuld a similar trick be done to extremely accurately make a section of the 107 tooth index wheel and then keep using it again and again ie only drill a few holes but do it as accurately as possible and use the same segment again and again instead of attempting to very accurately locate all 107 holes ;

ie see attached sketch

Bill

Hi Bill,

Originally Posted by steamingbill

Woodworkers sometimes use jigs where a small distance between two things is used again and again to repeat a measurement.

Thats fine until you get to the last hole, the distance from it to the first hole will be a collection of 107 pitch errors +/- any random pitch error from the jig. Even an error of 0.001mm leaves you with a pitch error of 0.1mm. Now if in use the gear doesnt do a full rev...................

Stuart

In theory yes, but in practice because of the application it is unlikely to be satisfactory, simply because of the cumulative error that is possible (107 holes with the pitch out by say 1% means that you could possibly lose a complete hole. If that is on a PCD of say 150mm, 1% of pitch is 0.04mm which is probably around the clearance you have between a bush and pin/ marking device).

The best thing to use for laying out these holes is an analogue computing device - also known in this application as a dividing head (or as you linked to, a gear cutting engine). A digital device (digital rotary table/ CNC mill, printer, ...) would be the next best thing but depending on the graduations in the encoder will have varying amounts of error.

Laying out by hand or using a drafting package and pasting a plot onto a blank would work but would rely greatly on accurate work, paper not stretching/ distorting and probably at the end of the process, a dollop of luck. 3D printing would reduce the errors from systematic to random so would be better but it's also one of those effort vs. reward things - how much resource do you spend making one off tooling? Spend too little and potentially all your efforts will be wasted. Spend too much and you will have the result you want but the cost (especially time) may be more than is warranted.

Michael

I've got some polycarbonate jobs to do on the mill tomorrow, I could knock out a 107 hole polycarbonate disk using PCD with 2mm hole size and 100 mm diameter.

100 * 3.142 = 314.2 314.2/107 = 2.936mm between holes centers on a single radius, so a 2mm drill should do the trick, If your indexing pin is 3mm then just make a tapered pin, say from 3 down to 1.8mm

I'd do it all in one setup, just drill the center hole at 2mm as well, and then you could center it on the lathe to bore out whatever center hole you want.

Ray