Is it possible to machine a lathe more accurate than the one you machine it on? If so, how?

[Henry BainbridgeParticipant @henrybainbridge40560]

Hi, just trying to get in to the classifieds, but thought I might as well ask, since its always intrigued me.

[Henry BainbridgeParticipant @henrybainbridge40560]

[Michael GilliganParticipant @michaelgilligan61133]

It depends how you define accuracy

The ‘first’ lathe need not have anything recognisably accurate about it … It could be the simple bodger’s lathe, or the watchmaker’s turns : both of which are little more than a couple of centres and a hand-rest.

Thanks for actually contributing something to the forum, instead of just making dummy posts

MichaelG.

[Rod RenshawParticipant @rodrenshaw28584]

I don't really know, but logic suggests that the earliest makers of lathes must have used some sort of dead centre lathe with hand held tools ( as Michael suggests), to make at least the turned parts of what we might recognise as the first type of lathe having any kind of similarity to what we would call a machine lathe today.

If the parts produced were not as good as wanted they may have been corrected by hand tool methods and then built into a new machine better than the one they were turned on.

I understand the very finest watchmakers' work is still done on primitive looking dead centre "turns"

I remember reading someone who had an Ornamental Turning lathe saying that all the screws holding his lathe together had centres in the ends showing that they must have been turned between centres, perhaps using a single point hand held "threading tool" to cut the threads, no wonder the screws were not interchangeable.

I should think that making an accurate lathe bed would have been very challenging until planers (or similar) were developed which begs the question "how did they make the planer bed?"

Rod

[John PParticipant @johnp77052]

"Is it possible to machine a lathe more accurate than the one you machine it on If so, how?"

I suppose the question here is posed incorrectly and perhaps should be.
Is it possible to produce a machine that will produce more accurate work
that the machine 's that it was made on.

If that is so then it is here is one

This cylindrical grinder was made in just that way ,the cylindrical square here
held between centres is within 1 um along the entire length and stands
square on a surface plate to within the thickness of 1 um clock gauge
indicator needle in any 4 places around the diameter.

As for the how, ME have had the construction article for the last 7+ years
or so if they haven't lost it again.

John

[JAParticipant @ja]

Henry Maudslay.

JA

 

Henry, sorry for the glib answer, just follow the name. [I don't usually post anything late at night].

Edited By JA on 28/10/2021 22:54:13

[BazyleParticipant @bazyle]

If you decide which part of the first lathe is letting it down you can use the lathe to make either a new part or a another machine that helps you make the new, better, part.
An easy example is the leadscrew. Lathes were made that averaged the errors of the leadscrew while making a new screw. Loop that back into the lathe and repeat.

[FarmboyParticipant @farmboy]

"Is it possible to machine a lathe more accurate than the one you machine it on?"

Surely the underlying principle behind all human development for a couple of million years . . .

[John OlsenParticipant @johnolsen79199]

Maybe worth mentioning a couple of techniques.

Accurate flats and straight edges can be made by making three at a time. if you hand scrape a single flat, you need another to compare it to. (using engineers blue.) If you scrape two, they end up matching but not necessarily flat. If you do three, and keep comparing A to B, B to C, and A to C, then eventually they all end up flat. This also applies to straight edges, so you can make an accurate straight bed with enough patience.

Then we will want a leadscrew. One technique is to draw some wire through a diamond die, which can give you a nice long length of wire of a consistent diameter. This is then wound around a mandrel which you have turned parallel. As implied by others above, this can be done by hand turning between centres. Then wind the wire around the long mandrel, and now you have a leadscrew. Coat it with some graphite and cast a lead nut around this, then use it to make a master leadscrew quick before you wear it out. As mentioned above, you can also correct errors in a leadscrew if you have the means to determine what they are.

If you can get hold of the Gingery books, they give quite a lot of information about how you can make a set of machine tools without having access to other machine tools. The problem with using other machine tools to make the new ones is of course that usually the new machine will be smaller than the one you have, when the usual problem we have is that we want a bigger one.

The other useful thing to look at is the Handmaiden series in ME by Tubal Cain, where he built a stationary steam engine using nothing but hand tools.

regards

John

[Bill PudneyParticipant @billpudney37759]

The concensus of the people that I used to work with, was that the accuracy achieved by a manual machine depended on the skill of the user, not so much on the machine. The accuracy achieved by an CNC machine was determined by the design of the machine, not so much on the user, or "button pusher" in the cruel parlance sometimes used by those who should know better.

Obviously as with all sweeping generalisations this is open to debate!!!!!

cheers

Bill

[Howard LewisParticipant @howardlewis46836]

The very first lathes were probably made using purely manual methods.

Having produced a crude lathe, it would have been used to make a second generation none, of greater accuracy based on the skill of the operator.

The first shaft would have been filed or chiseled as round as possible, The work produced from that machine would have been "rounder" and less tapered, so could then be used to turn ODs and bores more accurately than by hand.

Re iterating this, MANY times, results in the machines that we and industry use today. Obviously industrial machine are heavier, more rigid and more consistently accurate than our hobby machines. They are vastly more expensive!

That is not to say that hobby machines cannot produce accurate work. It may result in more scrap before the operator produces the desired degree of accuracy, but bit can be done if the operator has the skill and patience.

Howard

[Oily RagParticipant @oilyrag]

I endorse Michael G's post above – a very good question!

JA's post to look up the life and work of Henry Maudsley is a good starting point. The follow up to that would be to visit the Birmingham Centre Point which hopefully still has Joseph Whitworth's first lead screws, all three of them, (made from three screws hand chiselled and filed then each fitted to a lathe to remove error in the time old A-B, B-C, A-C. Don't bet on it though as some significant exhibits haven't seen the light of day for years!

A word here about what used to be the Birmingham Science Museum, now partially housed in the Centre Point building alongside City of Birmingham University (note: Not Birmingham University – that's in Edgbaston ). This facility is a shadow of the old Museum and some of that is down to its juxta-positioning to the C-o-B Media Studies and Performing Arts faculty next door. It has been dumbed down to now little more than a TV screen which alongside the exhibits does nothing to inform you of the object but rather berates the social injustices perpetrated on the 'victims' of capitalism forced to work in dingy work places!

As for more modern examples, when I was an apprentice at AH Ltd,. I worked on a Cincinnati vertical mill which had been fished out of the hold of the ship bringing it from the USA in 1941 (the ship had been sunk in Liverpool docks ) It had so much wear having been flogged to death by day and night shifts that the only way to accurately use it was by use of dial indicators and dead length bars (an early version of DRO, but actually ARO! ). This machine was used almost exclusively for 'fitters returns' where we 'fettled' parts which would not fit correctly due to some error. It often included easing a gearbox mating face by '0 thou, 3 thou, 8 thou, and 2 thou on the 4 corner mating edges. Another job was a lengthened leadscrew for a special machine which needed an extension shaft which had also to run in a housing at the point of the join – this meant the shaft had to be 'tongued and grooved' to an accuracy of 2 tenths of a thou runout between the two components.

I got a bar of chocolate off the foreman for that job.

The other machine I worked on was a No. 16 Universal which was part of an order in 1916 for Tsarist Russia, it had been rejected by the Russian on site Inspectors as being out of limits, so in Herbert tradition was put to work in making parts for there own machine tools! One component was the intricate drum cams for the cam auto machines Herbert's made. Backlash was eliminated by a wire rope and a sack of scrap iron hanging over a roller at the end of the table.

Martin

Edited By Oily Rag on 29/10/2021 07:51:48

[Michael GilliganParticipant @michaelgilligan61133]

If you want a particularly good example of ‘making something more accurate than the tools you are using’ … read about Jesse Ramsden’s creation of the Dividing Engine.

It is inspirational.

MichaelG.

[Martin KyteParticipant @martinkyte99762]

Yes.

As an aside George Thomas stated that accurate work can be produced on a worn lathe provided you know what the shortcoming are and act accordingly.

regards Martin

[larry phelan 1Participant @larryphelan1]

Oily Rag, I seem to remember that trick of the cord and the weight from somewhere else, dont remember where, but I did use it.

[Howard LewisParticipant @howardlewis46836]

A weight on a string wrapped around a suitable round component is a usual way of either taking up backlash or measuring torque to rotate, such as when preloading bearings in a Headstock or the differential for a vehicle.

At one company where I worked Torque Wrenches were checked by hanging weights on a hanger on a thin wire around a 24" diameter pulley, to give a direct reading.of lb ft.

Howard

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by Michael Gilligan on 28/10/2021 20:16:16:

Thanks for actually contributing something to the forum, instead of just making dummy posts

MichaelG.

I wasn't even expecting an answer, and I got more than I have ever had on a forum anywhere. Looks like its alive and kicking thanks everyone.

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by John P on 28/10/2021 21:25:18:

"Is it possible to machine a lathe more accurate than the one you machine it on If so, how?"

I suppose the question here is posed incorrectly and perhaps should be.
Is it possible to produce a machine that will produce more accurate work
that the machine 's that it was made on.

If that is so then it is here is one

This cylindrical grinder was made in just that way ,the cylindrical square here
held between centres is within 1 um along the entire length and stands
square on a surface plate to within the thickness of 1 um clock gauge
indicator needle in any 4 places around the diameter.

As for the how, ME have had the construction article for the last 7+ years
or so if they haven't lost it again.

John

I am not sure what you mean by incorrect. I posed a recursive question deliberately because AFAIK there are some parts of a lathe that can only be made by a lathe and I wondered how that can be. I take your point though, so for the sake of argument let's say the thought experiment is about how some tools are recursive in nature, and that being so, how accuracy comes about.

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by JA on 28/10/2021 22:51:23:

Henry Maudslay.

JA

Henry, sorry for the glib answer, just follow the name. [I don't usually post anything late at night].

Edited By JA on 28/10/2021 22:54:13

Googled Maudsley, wondered where you were going with that. Re-googled Maudslay with an A, faith restored. Thanks for the tip.

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by John Olsen on 29/10/2021 00:17:57:

Maybe worth mentioning a couple of techniques.

Accurate flats and straight edges can be made by making three at a time. if you hand scrape a single flat, you need another to compare it to. (using engineers blue.) If you scrape two, they end up matching but not necessarily flat. If you do three, and keep comparing A to B, B to C, and A to C, then eventually they all end up flat. This also applies to straight edges, so you can make an accurate straight bed with enough patience

OK that makes sense, so accuracy comes about by iteration, comparison and as Bazyle says, averaging.

Then we will want a leadscrew. One technique is to draw some wire through a diamond die, which can give you a nice long length of wire of a consistent diameter. This is then wound around a mandrel which you have turned parallel. As implied by others above, this can be done by hand turning between centres. Then wind the wire around the long mandrel, and now you have a leadscrew. Coat it with some graphite and cast a lead nut around this, then use it to make a master leadscrew quick before you wear it out. As mentioned above, you can also correct errors in a leadscrew if you have the means to determine what they are.

That reminds me of a ballscrew I have that seems to be a track wound around a shaft, rather than being cut into it like a thread. Seems pretty accurate!

If you can get hold of the Gingery books, they give quite a lot of information about how you can make a set of machine tools without having access to other machine tools. The problem with using other machine tools to make the new ones is of course that usually the new machine will be smaller than the one you have, when the usual problem we have is that we want a bigger one.

Thanks I will try and hunt them down.

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by Bill Pudney on 29/10/2021 00:28:31:

The concensus of the people that I used to work with, was that the accuracy achieved by a manual machine depended on the skill of the user, not so much on the machine.

This is encouraging

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by Oily Rag on 29/10/2021 07:43:42:

I endorse Michael G's post above – a very good question!

JA's post to look up the life and work of Henry Maudsley is a good starting point. The follow up to that would be to visit the Birmingham Centre Point which hopefully still has Joseph Whitworth's first lead screws, all three of them, (made from three screws hand chiselled and filed then each fitted to a lathe to remove error in the time old A-B, B-C, A-C. Don't bet on it though as some significant exhibits haven't seen the light of day for years!

A word here about what used to be the Birmingham Science Museum, now partially housed in the Centre Point building alongside City of Birmingham University (note: Not Birmingham University – that's in Edgbaston ). This facility is a shadow of the old Museum and some of that is down to its juxta-positioning to the C-o-B Media Studies and Performing Arts faculty next door. It has been dumbed down to now little more than a TV screen which alongside the exhibits does nothing to inform you of the object but rather berates the social injustices perpetrated on the 'victims' of capitalism forced to work in dingy work places!

As for more modern examples, when I was an apprentice at AH Ltd,. I worked on a Cincinnati vertical mill which had been fished out of the hold of the ship bringing it from the USA in 1941 (the ship had been sunk in Liverpool docks ) It had so much wear having been flogged to death by day and night shifts that the only way to accurately use it was by use of dial indicators and dead length bars (an early version of DRO, but actually ARO! ). This machine was used almost exclusively for 'fitters returns' where we 'fettled' parts which would not fit correctly due to some error. It often included easing a gearbox mating face by '0 thou, 3 thou, 8 thou, and 2 thou on the 4 corner mating edges. Another job was a lengthened leadscrew for a special machine which needed an extension shaft which had also to run in a housing at the point of the join – this meant the shaft had to be 'tongued and grooved' to an accuracy of 2 tenths of a thou runout between the two components.

I got a bar of chocolate off the foreman for that job.

The other machine I worked on was a No. 16 Universal which was part of an order in 1916 for Tsarist Russia, it had been rejected by the Russian on site Inspectors as being out of limits, so in Herbert tradition was put to work in making parts for there own machine tools! One component was the intricate drum cams for the cam auto machines Herbert's made. Backlash was eliminated by a wire rope and a sack of scrap iron hanging over a roller at the end of the table.

Martin

Edited By Oily Rag on 29/10/2021 07:51:48

OK I definitely need to find out more about this A-B, B-C, A-C idea. Seems like this thought experiment is throwing up some good general advice about how to do accurate work. Where there is a will there is a way!

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by Howard Lewis on 29/10/2021 06:53:15:

The very first lathes were probably made using purely manual methods.

Having produced a crude lathe, it would have been used to make a second generation none, of greater accuracy based on the skill of the operator.

The first shaft would have been filed or chiseled as round as possible, The work produced from that machine would have been "rounder" and less tapered, so could then be used to turn ODs and bores more accurately than by hand.

Re iterating this, MANY times, results in the machines that we and industry use today. Obviously industrial machine are heavier, more rigid and more consistently accurate than our hobby machines. They are vastly more expensive!

That is not to say that hobby machines cannot produce accurate work. It may result in more scrap before the operator produces the desired degree of accuracy, but bit can be done if the operator has the skill and patience.

Howard

Interesting to note that things tend to get more accurate the more times you repeat them. It seems that the key to making that happen is to keep making more tools, not just more things with the tools you have.

[Henry BainbridgeParticipant @henrybainbridge40560]

Right, now to check out the classifieds

[Henry BainbridgeParticipant @henrybainbridge40560]

Posted by Michael Gilligan on 29/10/2021 10:58:25:

If you want a particularly good example of ‘making something more accurate than the tools you are using’ … read about Jesse Ramsden’s creation of the Dividing Engine.

It is inspirational.

MichaelG.

That is really fascinating. The wikipedia page doesn't actually say how it works, but luckily I found a .pdf online of the original 18th-century description by the man himself. In order to get his share of the longitude prize, he had to share how it worked, so its good to see he is still true to his word

[Author]

Posts