Metric threads on an imperial lathe - disengaging half nuts!

[RobCox]

I have this threading dial that came with my recently(ish) acquired Harrison M300. It's a bit bigger than a standard Harrison dial and doesn't fit on the lathe using the standard mounting hole for the threading dial. It was attached to a shop made adaptor plate on the side of the Ainjest attachment:

I've taken it apart, cleaned it, painted it to match the lathe, then decided I fancied a standard looking dial so machined up a replica of a standard M300 one. I thought I'd list this one for sale as I didn't need it. But first, to make it more attractive to sell, it would need the 20T pinion listed on the data plate. I assumed that this would interface to a 6mm pitch leadscrew, so I made the blank 42.1mm OD, by my calculations the correct OD for a 20T gear with a 6mm circular pitch. Its wrong on several counts.

This morning I thought I'd write a spreadsheet to verify that 6mm was indeed the pitch that this box of tricks was designed for. For the uninitiated, me included before I sat down to work all this out this morning, if the leadscrew would move the carriage a whole number of "to be cut" thread pitches between numbers on the dial, you can disengage the half nuts and re-engage on that number (or numbers for some threads). Anyhow, the spreadsheet calculated that the dial wouldn't work as per the data plate for a 6mm pitch leadscrew, nor 5mm, nor 4mm. OK, for a laugh, what about 6.35mm (1/4")!
Brain explosion time. Yes, it is designed for a 4TPI leadscrew!! Cutting metric threads with an imperial leadscrew, disengaging the half nuts and winding the carriage back to the start!! No spindle reversing needed!

If you're still reading, your eyes haven't glazed over and you want to know how, here comes the maths that went into the spreadsheet. Rest assured, once I've sorted out the pinions, I will be trying this and will report back whichever way!

The pinion that engages with the leadscrew has Z teeth. The leadscrew has a pitch of P. So, per revolution of the dial, the "lead" or travel of the leadscrew is Z.P, so 16 x 0.25" = 4 inches for the standard 16T pinion. 4 inches is an integer number of any whole number of TPI thread, eg 13, 16, 20TPI and also a whole number of half or quarter TPI threads, such as 5 1/2TPI or 2 1/4TPI, both of which the M300 can do.
Half a turn works for whole TPI and half TPI threads and a quarter works for whole TPI threads.

But they're not metric! True, but if you put a 20T pinion on it, one turn of the pinion is 5 inches travel, or 127mm. So one number works for 0.5mm and 1mm pitch threads.

Still not very exciting, however the lever on the side of the dial has 3 positions. Position A is direct drive, albeit through a gear train, which gives us the answers above. Position B however switches in a gear reduction between the pinion and dial of 5:1 and position C a reduction of 6:1. Now the lead between one number coming around is 635 and 762mm respectively. This means the maths works for thread pitches of 0.5, 0.8, 1.0, 1.25, 1.50, 2.0, 3.0, 4.0, 5.0 and 6.0mm. Add the option of a 21T pinion and engaging on one number it works for 0.7 and 1.75mm pitches too:

Here's a table to summarise:

So I'm going to be making a couple of pinions and funnily enough, I'm not going to try and sell it now!

 

[Smouser]

My brain hurts!

I should have just looked at the table and accepted that all sorts of malarky inside the dial makes it work.

 

[Martin-Sy]

That's very interesting.

My brain hurts!

Mine too.

Do you have any details or pictures of what's inside?
If I understand this correctly ... the ratio, selected by the lever, of the gearbox inside the "Metradial" converts the leadscrew rotation to dial rotation such that the metric pitches can be cut with the half nuts just as they would for an imperial thread.

 

[smashweld]

the gearing is to save having a much larger pinion gear, so for the B setting (5:1) with the 20t gear is the equivelant of having a 100t gear meshing directly with the leadscrew, so 100 leadscrew threads would be one rotation of the dial
for the C setting (6:1) it would be like having a 120t gear meshing with the leadscrew, 120 leadscrew rotations for one turn of the dial

all imperial threads line up every 1, 2, 4, 8 or 16 leadscrew threads which fit evenly in to the standard 16t gear

metric threads aren't all the same so need the different gears/settings for different pitches, and with a 127 gear on an imperial lathe the time it takes for the dial to come round to the correct spot can be quite long, especially for the coarser pitches

 

[Brad93]

I have this threading dial that came with my recently(ish) acquired Harrison M300. It's a bit bigger than a standard Harrison dial and doesn't fit on the lathe using the standard mounting hole for the threading dial. It was attached to a shop made adaptor plate on the side of the Ainjest attachment:

View attachment 499759

I've taken it apart, cleaned it, painted it to match the lathe, then decided I fancied a standard looking dial so machined up a replica of a standard M300 one. I thought I'd list this one for sale as I didn't need it. But first, to make it more attractive to sell, it would need the 20T pinion listed on the data plate. I assumed that this would interface to a 6mm pitch leadscrew, so I made the blank 42.1mm OD, by my calculations the correct OD for a 20T gear with a 6mm circular pitch. Its wrong on several counts.

This morning I thought I'd write a spreadsheet to verify that 6mm was indeed the pitch that this box of tricks was designed for. For the uninitiated, me included before I sat down to work all this out this morning, if the leadscrew would move the carriage a whole number of "to be cut" thread pitches between numbers on the dial, you can disengage the half nuts and re-engage on that number (or numbers for some threads). Anyhow, the spreadsheet calculated that the dial wouldn't work as per the data plate for a 6mm pitch leadscrew, nor 5mm, nor 4mm. OK, for a laugh, what about 6.35mm (1/4")!
Brain explosion time. Yes, it is designed for a 4TPI leadscrew!! Cutting metric threads with an imperial leadscrew, disengaging the half nuts and winding the carriage back to the start!! No spindle reversing needed!

If you're still reading, your eyes haven't glazed over and you want to know how, here comes the maths that went into the spreadsheet. Rest assured, once I've sorted out the pinions, I will be trying this and will report back whichever way!

The pinion that engages with the leadscrew has Z teeth. The leadscrew has a pitch of P. So, per revolution of the dial, the "lead" or travel of the leadscrew is Z.P, so 16 x 0.25" = 4 inches for the standard 16T pinion. 4 inches is an integer number of any whole number of TPI thread, eg 13, 16, 20TPI and also a whole number of half or quarter TPI threads, such as 5 1/2TPI or 2 1/4TPI, both of which the M300 can do.
Half a turn works for whole TPI and half TPI threads and a quarter works for whole TPI threads.

But they're not metric! True, but if you put a 20T pinion on it, one turn of the pinion is 5 inches travel, or 127mm. So one number works for 0.5mm and 1mm pitch threads.

Still not very exciting, however the lever on the side of the dial has 3 positions. Position A is direct drive, albeit through a gear train, which gives us the answers above. Position B however switches in a gear reduction between the pinion and dial of 5:1 and position C a reduction of 6:1. Now the lead between one number coming around is 635 and 762mm respectively. This means the maths works for thread pitches of 0.5, 0.8, 1.0, 1.25, 1.50, 2.0, 3.0, 4.0, 5.0 and 6.0mm. Add the option of a 21T pinion and engaging on one number it works for 0.7 and 1.75mm pitches too:

Here's a table to summarise:

View attachment 499761

So I'm going to be making a couple of pinions and funnily enough, I'm not going to try and sell it now!

That's seriously smart

 

[Martin-Sy]

Sounds to me like it would feasible to make an electronic version of it.
Instead of the gears inside the 'Dial' a shaft encoder, some firmware and an LED to indicate when to engage.
Obviously the encoder needs to be on the apron and be driven by a gear meshing with the leadscrew.
It would also need a method of setting the TPI/Pitch being cut.
Not an ELS as Clough42 but an addition to a conventional manual lathe.
Just need to add it to the already long 'project' list

 

[Smouser]

Sounds to me like it would feasible to make an electronic version of it.
Instead of the gears inside the 'Dial' a shaft encoder, some firmware and an LED to indicate when to engage.
Obviously the encoder needs to be on the apron and be driven by a gear meshing with the leadscrew.
It would also need a method of setting the TPI/Pitch being cut.
Not an ELS as Clough42 but an addition to a conventional manual lathe.
Just need to add it to the already long 'project' list

Hurry up then.

I will be your first customer!

 

[minimutly]

I promise I will read this through and try to get my head around it, sounds like it might be useful.
Last metric thread I cut on the old Triumph I had to cut a smaller diameter thread to the right of the "required" thread, used this to pick up using the top slide (I've never set up the forward/reverse on the single phase motor). I found it was quick and repeatable, but a bit of a faff setting the doc every time, given the 30 plus thou of backlash in my cross slide.

 

[RobCox]

I now have the gears I need to try this thing out. I used some unknown flavour of bronze that came with my arbor press as "tooling" as raw stock:

This got turned into blanks for the gears:

3 blanks for a 20T, 21T and the earlier cock-up that's been turned down to make a 14T. The last one enables you to do 0.7mm threads (M4) without gear reduction. So, gash the blanks on the mill (makes the shaping operation easier on the machine):

then off to the shaper to put the involute profile on:

The end result plus the original 16T which I think is cast iron:

Do you have any details or pictures of what's inside?

I popped the back of the dial off and took some pics. While I was at it I counted the teeth:

Looks like it would be easy enough to make. Certainly preferable to this!

Just got a reply from DGR and a new metradial for my 4tpi M300 is £450 plus postage!

 

[RobCox]

I got the threading dial mounted on the side of the apron. The combination of a short length of bar stock and the pivot on the dial means regardless of the gear size installed on the dial it can be adjusted correctly. I'm making use of the standard threading dial mounting hole on the apron:

I had a go with this dial yesterday. Mixed results. First the good, M8 x 1.25, half nuts disconnected and the carriage wound back. This uses the 20T gear, lever position B, engage on any numbered line:

Now the bad and ugly. This was down to the operator, not the dial. M12 x 1.75, lever position C engage on one number. The first 3 passes were going well, until I stopped looking intently at the dial and watched the tool. I was running at 540rpm, the tool moved fast!

I missed engaging at the line so it double threaded. Now for the reasons why. Using position B or C slows down the dial as expected. It's just such a painfully long wait that it makes reversing the lathe quicker. Thats why I was running at 540rpm, the tool moving at 15mm/second! Sharp reflexes were needed to stop it in the runout groove. Secondly, using a 21 tooth gear means there are 21 positions round the dial that the half nuts will engage. Add in the 6:1 reduction factor and there are now 126 positions per rev!
To get an idea of what that looks like, this is the ideal engagement point:

and this is the one immediately before that I hit by accident:

Conclusion - it works fine but the advantages are somewhat limited. It really needs a means of physically reducing the number of engagement points. That'll be the Ainjest then. Work is ongoing to get all parts of that ready to go.

 

[Smouser]

Thanks for sharing!

I can see how you could very easily engage the halfnuts a little bit too soon or too late with the positions so close to each other.

Still a very cool bit of kit!