Nearest Imperial Thread TPI for metric thread cutting on imperial machine

[Russ BParticipant @russb]

I just threw this table together to see how close I could get to metric with the standard set of change wheels on an imperial lathe and I thought I'd share it, as it could be useful to others?

And I've just deleted it……. I think I can do better, watch this space

Edited By Russ B on 30/04/2017 20:30:17

[Russ BParticipant @russb]

[Russ BParticipant @russb]

Basically the issue with the table I just posted, is that i wrote it from behind my desk – a dangerous (but popular) method of engineering solutions these days!

Depending on the pitch of your leadscrew and the change wheels on offer, you could get more accurate results. I'll put the table back in, but obviously taking onboard what I just said, using non interger values for the TPI can achieve better results…. I'll update this once I've worked this out, for my machine (a 7×12 chinese mini lathe)

TPI	PITCH	ACTUAL	ERROR %
64 (or 63, same error)	0.400	0.397	0.750
56 (or 57 same error)	0.450	0.454	-0.889
50	0.500	0.508	-1.600
36	0.700	0.706	-0.857
32	0.800	0.794	0.750
25 (uses #32)	1.000	1.016	-1.600
26	1.000	0.977	2.300
20	1.250	1.270	-1.600
16	1.500	1.588	-5.867
14	1.750	1.814	-3.657
13	2.000	1.954	2.300
10	2.500	2.540	-1.600

[Russ BParticipant @russb]

So, the 7×12 mini lathe has a gear (A) driven 1:1 by the spindle via a tumble reverse, and one that drives that leadscrew 1:1 (D) . It then has a single moveable layshaft on which a pair of gears (B, C) can be fitted to obtain a different ratio, B meshes with A, and C meshes with D.

Or a single gear can fit where B&C go, to drive directly from (A) the spindle to the (D) leadscrew in which case, any gear will do as its just "passing it on" so whatever fits comfortably (use a larger one, the small gears will wear out quicker and the 20T's are weak due to the keyway)

So, I believe the standard gears are as follow

2x 20T
2x 80T

and you get the follow as an extra for screw cutting (some/most? vendors include them now and I can't seem to find anyone selling them – so I'm 3D printing them which is much more fun)

1x 25T
1x 30T
2x 40T
1x 45T
1x 50T
1x 55T
1x 57T
1x 60T

So, here are the combinations that achieve more accurate metric threads, I've not tested these actually fit since some gears will just be too large (the centre to centre of C & D + the radius of the leadscrew must always exceed the radius of B, and if C is much larger than B, it will hit the washer that holds A, but it doesnt look too close to me. I have stared at it, and the possible combinations so hopefully, these are compatible. Some combinations might simply call for too many large gears, and there wont be room – we'll see I guess.

Desired Pitch (mm)	A	B	C	D	Actual	Error
0.400	20	1	1	80	0.396875	0.781
0.450	20	65	55	60	0.447756	0.499
0.500	20	57	45	50	0.501316	-0.263
0.700	45	55	35	65	0.699388	0.087
0.800	45	55	40	65	0.799301	0.087
1.000	45	55	50	65	0.999126	0.087
1.250	35	40	45	50	1.250156	-0.013
1.500	65	55	40	50	1.500909	-0.061
1.750	40	57	55	35	1.750627	-0.036
2.000	40	50	55	35	1.995714	0.214
2.500	65	55	60	45	2.501515	-0.061

[Nick HulmeParticipant @nickhulme30114]

nThreadP from Tony's site –

http://www.lathes.co.uk/latheparts/page14.html

It allows you to enter your set of changewheels, pitch and maximum allowable error then yields the usable combinations.

– Nick

[BazyleParticipant @bazyle]

Have you seen Neil's article on the Arceuro site about using a 63 tooth gear as a driver.
Also he started a thread a few weeks ago with links for a 63 tooth 3D print file I think.

The 57 seems to be an oddball figure. Is it used in the imperial thread charts? I would doubt it so it makes more sense to do a 63 instead.

If you like manipulating the figures try putting TWO off 63s as driven gears. It is one of the more accurate approximations.

[Author]

Posts