Colchester Bantam metric scewcutting on imperial only machine

Having bought a MK1 Bantam for the garage I am ready to screw-cut metric threads.

I have hit a brick wall when setting up the changewheels - I can get within 1% of true pitch but cannot mate outer threads with tap cut inner threads without making the thread sloppy.

All my calculations are way out and am ready to ask for help........

I have all the change wheels including the 127 tooth gear but cannot get the right combination - can anyone help with links to where I can get the information? Or have done it themselves and can post a photo of their setup?

Trevor

Hi Trevor

Try doing a search on the internet for a program called "nthreadp". You put in the list of change wheels that you have, the tpi of the leadscrew and the thread that your after with the error that you'll accept and it spits out a list of possibilities. Program is free and will run with most versions of Windows.

Keith

but cannot mate outer threads with tap cut inner threads without making the thread sloppy

That may be your answer

There can be a LOT of headroom in a metric nut for instance

My 10mm cut was snug, because it was actually around 10.3mm in dia

I as a layman, had always assumed 10mm WAS 10mm

It's not that simple, unfortunately

More knowledgeable folk than me may fill you in on the details... which can be a bit of a surprise

Edited By Ady1 on 10/07/2012 23:39:09

The best thing to do is use delrin or aluminium for practicing the thread until it pops right on because these materials are fast and simple

Then you need to do the real deal in steel etc and if you want a snug fit with an off the shelf metric thread you will need to oversize or undersize it slightly depending on the job

Edited By Ady1 on 10/07/2012 23:47:09

Thanks Keith, programme sitting on my desktop as I type, enough combinations to keep me busy tomorrow night.................

Trevor

Hi Guys,

It is not the OD that is important but the true depth/width of the valley.

The only way for the average home engineer to measure threads is with the three wire method as it measures the thread where it matters.

chriStephens

Hi Trevor

Have you got the manual for the Bantam? It has all the setup info for cutting metric on an imperial machine. I don't know if your problem is basically a lack of a manual or whether you are really needing to cut a deadly accurate pitch. If it's the former, I can help - let me know if you'd like me to send you the section on setting up the Bantam for metric cutting.

I did this myself 2 weeks ago and the recommended gears were 35 (headstock), 30 (gearbox), 120 (meshing with gearbox) / 127 (meshing with headstock) to cut a 3mm pitch. However, you also need to set the 2 gearbox knobs to positions 5 and B for this example. I don't know what pitch you are trying to cut.

I don't know what the theoretical error is but doubtess HH's article will tell you. However, as it was a internal/external threaded chuck adaptor I actually wanted a degree of slop so that the chuck and adaptor would engage with the register - my kind of thread!

You also have to get the thread profile right. If the peaks of your thread are too pointy or your valleys are too wide, they may bottom out in your "nut".

Murray

Just for the record, if there is a 127 tooth wheel in the train and an inch leadscrew, the theoretical error can be zero as 1 inch is defined as 25.40 mm. 127 is 254 divided by 2, so you can set up the exact ratio needed for a metric thread defined as a pitch that is a rational number in mm, given enough changewheels.

There you are then - thanks John. So if you use the right gears and gearbox settings, you should theroretically achieve zero error. PM me if you want me to send you the Bantam manual in PDF with the settup information.

Murray

(1) Whilst a 127 tooth gear gives the ideal conversion ratio there are other ways of achieving an accuracy of pitch which far exceeds that of any lathe . The best and most common use 63T and 21T gears .

(a) I've not yet worked out the details but an easy way to set metric threads on an imperial lathe on a regular basis must be to arrange a mini self contained gear box which will fit the imperial banjo in place of any pair of ordinary gears and have a ratio of 1.27:1 .

(b) A variation of (a) and a more permanent fitting would use something like a tumbler reverse set up on the final drive gear - direct drive for 1:1 ratio Imperial - engage tumbler for 1.27:1 metric .

Effectively you could switch between (for example) an 8tpi imperial lead screw and a 3mm pitch metric leadscrew .

With a fixed 1.27:1 ratio available the other gears needed for metric screwcutting are just set to easy to calculate simple ratios . The concept of doing all the metric conversion in one place and then using simple trains for the rest of the gear train is infinitely better than using multiple approximation methods and weird trains .

(2) If by doing it in any way at all a lathe user can achieve a pitch accuracy of better than half of one thou per inch in a home workshop he is doing remarkably well . Calculations of trains to give higher accuracies are meaningless .

Edited By MICHAEL WILLIAMS on 25/07/2012 15:02:53

Edit to above :

Effectively you could switch between (for example) an 8tpi imperial lead screw and a 2,5 mm pitch metric leadscrew .

New :

Gearbox/tumbler ratio can be adjusted so as to make any imperial pitch lead screw effectively a metric one and allow use of simple gear trains .

Michael

You do realise that the smallest number of teeth on the gears to give you ACCURATELY your 1.27:1 ratio is a 100/127 gear pair?

I think the problem with making a mini gear box to do what you want will be getting gears small enough and strong enough. Otherwise just use the 127 change wheel.

Martin Cleeve in his "Screwcutting" book shows effectively your idea when he used a 50(or 40)/127 gear pair but of 30dp on a Myford in place of the normal 20dp gears, on the grounds of it being physically smaller and easier to fit, but he wasn't inclined to go to much finer pitch gears I don't think.

Keith

If you use a 21 or 63 tooth gear on the INPUT side (instead of output with a 127 tooth gear) you will get more compact gear arrangements.

Do NOT use 63 teeth as an approximation of 127/2.

For 8TPI leadscrew 1 revolution = 25.4/8 = 3.175

If you have a 127 tooth gear on the output side your equation will include divide this by 127

3.175/127= 0.025. Multiply by 40 to get 1, so 40:127 gears will give exactly 1mm pitch.

If you have a 63 tooth gear in the input side your equation will include multiplying 3.175 by 63

63*3.175 = 200.025 so 63:200 gears (more practically two pairs such as 63:50 20:80 which give the same ratio) will give you a pitch of 1.000125 ior an error of just over 1 part in 10,000, well within the tolerance for your leadscrew, and hence indistinguisable in the result from using a 127 gear.

Neil

P.S. My 63-tooth gear, made with a handshaped cutter using the 60-tooth gear as a cutter filing template and a dividing head. metric threads made using this have all worked fine.

Murray, you are correct in that I do not have a manual, but me being me, I like to figure things out for myself.

Anyway, most of the above posts are reccomending the gears that I found did actually work. Here is the table I created;

Pitch... Lever/Pos..... Changewheel sequence...... %age error

3 ...........C6................. 35--127/100--57 ...................0.2%

2.5 ........C6................. 30--127/100--57 ...................3%

2.25 ......C5 .................30--127/100--57................... 0.2%

2 ...........C4 .................30--127/100--57 ...................0.2%

1.75 ......C1 .................35--127/100--57 ...................0.2%

1.5 ........C1 .................30--127/100--57 ...................0.2%

1.25 ......B6................. 30--127/100--57 ...................3%

1 ...........B4 .................30--127/100--57 ...................0.2%

0.75 ......B1 .................30--127/100--57 ...................0.2%

0.5 ........A4 .................30--127/100--57 ...................0.2%

Apart from the 2.5 and 1.25 pitches I would say that all the settings are as near perfect as you are likely to get.

Trevor

Edited By Trevor Wright on 25/07/2012 21:33:13

Hmmm table was spaced out when I wrote it, still, it is just about understandable.

Trevor

As has been stated the only exact way to cut metric threads using an imperial leadscrew is to include a 127 tooth gear in the drive train. The number has to be 127 since this is a prime factor of 254. Not only is 127 a prime number it's a Mersenne prime, useless but true! It is usual to combine the 127 tooth wheel with a 100 tooth to give the ratio 1.27, the usefulness of which has been explained by Michael.

A gear with 63 teeth can be used as an approximation, used in conjunction with an 80 tooth gear; 80/63 = 1.269841, an error of about -0.012%. I don't know what gear is normally used in conjunction with a 21 tooth gear.

However, both the 127/100 and 80/63 combinations result in relatively large gears. A ratio of 47/37 uses smaller gears, and gives 1.27027 recurring. This is an error of 0.021% or about 2 tenths per inch.

Regards,

Andrew

Well done Stub mandrel - you got it in one .

When I mentioned the plug in gear box I had (20/21).(20/15) in mind = 1.26984 which is good enough for almost any normal purpose .

Many Q/c screw cutting gear boxes with metric and imperial capability use the 63 T driver as basis of the imperial - metric converter .

Of the two things I mentioned - plug in gearbox or tumbler on final drive gear - the tumbler is much easier to implement . Properly designed and permanently fitted it could potentially solve almost all the normal 'metric conversion' problems on an imperial changewheel lathe .

Tumbler seems to be neatest but other ways possible - could be a sideways shift arrangement .

Purely as a matter of interest if another plug in gear box or another leg on the tumbler implemented a Pi ratio then DP pitches could also be cut with simple ratio gear trains . See if anyone can work that one out .

Regards ,

Michael Williams .

The biggest problems were that there is another gearbox between the change wheels and the leadscrew which gives another gear combination set - and is different for each selection lever. Also the recommended output gear of 30 teeth cannot be driven from the 120 tooth gear because the 127 gear hits the output shaft.

It was only when I recognised patterns of multiple tpi that it was possible to select a ratio to give a coarse pitch, work out the change wheels, and use the gearbox selections to acheive the other pitches.

Hope that makes sense.......I agree with the calculations above, unfortunately, the practical applications of the calculations are not physically available.

Trevor

You use a collar / spacer to clear the 127t gear. If you don't have one, it's easy to make one up. Same width as the gears. As I said, I did this exact setup myself just a few weeks ago. Took me very little time and effort. The machine is designed to take the 120/127t combination otherwise they wouldn't supply them....

The manual would cut out a lot of trial and error

So about 2 thou per inch error; doesn't sound quite 'perfect' to me? I would have thought that an order of magnitude better, ie, 0.02%, would be acceptable. Of course if you have the 127 tooth gear then the error could be 0%..............

Regards,

Andrew