Accuracy in Machining Large Gears

Advert

New Replies

Report Bug

Help/FAQs

Accuracy in Machining Large Gears

This topic has 6 replies, 4 voices, and was last updated 7 August 2011 at 23:02 by Anonymous.

Viewing 7 posts - 1 through 7 (of 7 total)

Author

Posts
7 August 2011 at 12:31 #72984
Anonymous
At some point in the near future I’m going to have to face up to cutting the final drive gears for my 4″ scale traction engines. Details are; 5DP 20°PA, 72 teeth, OD 14.8″. This is just a bit too big to go under the cutter on my horizontal mill.

Having considered various alternatives I think that the best solution is to use a rotary table in horizontal mode (on the horizontal mill) and use the knee to apply the cut.

The question I have concerns the accuracy of the final gear. My dividing head manual quotes a maximum cyclic inaccuracy of 1°30′, and I’d be surprised if the rotary table was any better. Taken on a 14.8″ diameter circle this angle equates to a bit over 3 thou. In other words the tooth thickness could vary by 3 thou from tooth to tooth. I can think of two ways round this. One, just cut deeper than theoretical and accept the backlash. For 20°PA the formula seems to be to cut deeper by 0.73 times the required backlash. Two, use a plate with a 72 holes placed on a large diameter, say 12″, and use these to index the gear.

I think I’m probably going to go with option one, as it is simpler, and I don’t need ‘precision’ gears here. What are other peoples’ experiences?

By the way, I know that the original, full size, gears would most likely have been ‘as cast’ and that the tooth engagement varies according to the rear axle springing. So I don’t need any smartypants comments about OTT over-engineered twaddle.

Regards,

Andrew
Advert

7 August 2011 at 12:31 #2720
Anonymous
7 August 2011 at 12:45 #72985
JasonB
Moderator
@jasonb
Does the model have working springs or just dummies?

Either way I don’t think that a little backlash is going to be a big issue. You want a bit of clearance in the axle and shaft bearings otherwise as the heat affects the hornplates and shafts there is a risk of things tightening up on you. This clearance should help reduce tightspots on the mesh by allowing the gear centres to move slightly, a bit of coal dust that will get onto teh gears also helps smooth things out.

Is there a chance that you could add a “stub axle” to the machine table that the mating gear could be slipped onto, that way you could cut to calculated depth, slip on the small gear and see how things mesh by turning the rotary table. If you find things are a bit tight either cut the whole gear deeper or possibly skim the edge of any thick teeth.

J
7 August 2011 at 12:53 #72988
John Stevenson 1
Participant
@johnstevenson1
Andrew,

First of all it’s a proven method, I have never done it but I did turn a blank ring gear up for Tim Leech at Warrington to fit onto a narrow boat diesel engine.

The blank was 27″ and beyond his machining capabilities at the time.

it then went onto his horizontal mill and was driven by a stepper driven rotary table.

These are some pictures that Tim posted at the time.

John S.
7 August 2011 at 14:16 #72996
JasonB
Moderator
@jasonb
Just another thought can you use the dividing head in a vertical position to space the teeth and arrange a couple of substantial blocks either side of the cutter to clamp the gear blank to. That way you get the same accuracy of division that you had with the rest of the gears and the gear is supported rigidly as it would be on a large rotary table.

J
7 August 2011 at 19:26 #73006
MICHAEL WILLIAMS
Participant
@michaelwilliams41215
(1) I wouldn’t worry about a three thou error in a gear this big which is going to be used for a fairly rough and ready application . There are other errors in both mating gears and in the meshing distance any of which could be additive or subtractive . Near 100% probability your new gear will mesh perfectly if just made standard depth but worst case you’ll get a few high spots to be removed by hand lapping .

(2) If the accuracy of your rotary table ever really mattered to close limits you could do a one time calibration . Really high class makers of rotary tables – like Societie Genevoise -always provided their already highly accurate tables with a ‘ table of deviations’ which was a table of minute offsets which could be applied at the input handle dial to compensate for very small variable pitch errors in worm wheel .

.

Edited By MICHAEL WILLIAMS on 07/08/2011 19:36:54
7 August 2011 at 23:02 #73022
Anonymous
Thanks one and all for the advice.

The engines as designed don’t have real or dummy springs; just a rear axle that runs in a cast axle tube, that bolts to the hornplates. Being contrary I haven’t made the hornplates yet. I intend to made a test bar to measure the centre to centre distances for the gear sets and then adjust the holes in the hornplates accordingly, should they not be to the calculated numbers.

I had considered using the dividing head in vertical mode, but the gear would then be the best part of 12″ up in the air. That would need some pretty substantial supports to counteract the horizontal cutting forces. With the rotary table the gear will only be about 3″ above the milling table, and the rotary table itself will help resist the forces.

JohnS: Thanks, I have seen those pictures on my trawl through the web on this subject, but I didn’t know you were involved. I should have known better!

Michael: That’s a succinct summary of the issue, and you’re right I don’t need to worry!

I just need to get on and cut the gears now.

Regards,

Andrew
Author

Posts