19th Century English micrometer with thread compensation?

[rivett608]

Here is another micrometer by Elliot Bros. of London.... it is related to this one

http://www.practicalmachinist.com/vb/antique-machinery-history/english-micrometer-1870s-171345/

This is a very heavy by weight because it is solid nickel silver, it also has a 50 pitch thread and in this case the thread must not be perfect.... notice the angled line on the barrel to compensate for this.... at least that is what I think it is for? There was a patent for a angled line but I think this is not it however I do think it was a English patent.....

Here it is with the 1/2" one from the above linked thread and it's original slip case.... notice this case is different.... you slide the mic in from the top.... the top flap is missing so it needs some restoration..... one more project

Also these nickel silver/50 pitch thread mics I believe are related in some way to the American Micrometer by George Church.... the inventor of the ratchet.... here is all three....

http://www.practicalmachinist.com/v...mportant-micrometer-invention-ratchet-171306/

 

[AntiqueMac]

Rivett,

Really interesting. And very super nice examples. But, what the hay?????

So, the esteemed Elliot Bros shop could not make a thread accurately? Was it a lathe turned thread or was it from a die, like at Starrett?

And, can you imagine any company today covering an error in manufacturing by making a compensating "error"? (Maybe I don't want to know?)

Unreal! Neat!

And, I'm waiting for the "P"!!!

 

[rivett608]

I think a few things are at play here..... a very fine thread, a super accurate application and remember this comes from a shop that had been making all sorts of scientic instruments by hand, often one of a kind to order since at least the early 19th century...... The traditions of this shop would be of hand fitting rather than mass production with interchangeable parts... This very well may have been made in time between when Brown & Sharpe made their tiny sheet metal gauge and their 1" micrometer with the squared frame.

Here is a little history of them.... I have a nice drawing set signed Elliott & Sons plus some later stuff....

William Elliott was born in 1780 or 1781 in the parish of St Andrew, Holborn, London and in 1795 he was bound apprentice for seven years to William Backwell a compass and drawing instrument maker. Company histories claim that William Elliott established his own first business in 1800 although 1804 is the more likely date. He was now trading under his own name of W.Elliott; this was the beginning of a highly successful business.
William Elliott made drawing instruments and mathematical scales many of which survive. Later he described himself as a mathematical, optical, and philosophical instrument maker.

Around 1850 William Elliott took his two sons, from his third marriage, Frederick Henry and Charles Alfred into the business and began trading as William Elliott and Sons. At the Great Exhibition of 1851 the company was awarded a Bronze Medal.
William died in 1853 and in his will he left his share in the partnership to provide his wife, Emma, with an annuity. Frederick and Charles obtained patents for improvements in drawing boards, barometers, and telescopes. The company prospered, its success attributed to Frederick's business ability. Before long the Company was producing most of the standard optical, surveying, navigational and "philosophical" instruments for home and overseas customers.

By 1854 his two sons were trading as Elliott Bros and in 1857 they took over the firm of Watkins Hill. The company worked willingly with academics and others to realize their inventions; instruments for the new applications of electricity were produced and became an important part of the firm's activity. Customers included J. C. Maxwell, Charles Wheatstone, Lord Rayleigh and other leading scientists. Products from these years include galvanometers and batteries.
It would be convenient for the avionics link to record that the barometers used in early balloon flights were supplied by Elliott Bros but these instruments were a only a minor product of the company and a there is no evidence to support that they supplied any for this purpose.

In 1865 Charles Elliott retired and he died in 1877. Frederick continued to run the business as sole proprietor and on his death in 1873 he left the business to his wife Susan.

Here is where this came from if you want to read more...

History of Elliott Brothers

 

[rotarySMP]

Very high standard of engraving. Intersting to see the difference between your Palmer and this in just 30 years (and crossing the channel).

 

[Asquith]

Rivett,

Very interesting. Thanks for posting.

I like your theory, and I’m sure you’re right. There’s another one here, with the line going the other way:-

Micrometer By Elliot Bros - Gilai Collectibles

On yours, if I’m interpreting this correctly, it looks as though the pitch error would be about 2 thou per inch.

In the absence of pitch correcting lathes, I can see that it would be difficult in those days to get the screwcutting gear train to give exactly the right pitch. Not impossible, but expensive, and possibly easier to make the correction in the way shown on your micrometer, especially if you’re not making many micrometers.

I wonder if they calibrated and engraved each mic individually, or did they work on the basis that any thread cut with that set up would give the same error, and ‘pre-engrave’ the barrel accordingly?

 

[rivett608]

Asquith You are right.... they must have some sort of measuring fixture to check the threads or mics once assembled and then set it to engrave the lines....... Thanks for sharing the link to the other one. On my 1/2" the line is straight.... but now I'll have to take a close look to make sure.

I also wonder would a 40 pitch screw be that much easier to cut with less error and hence it became the standard?

Thanks for all you're comments

 

[Billtodd]

Are the basis lines 'straight' - in a non-euclidean way (i.e. two reference points) or to they curved (many reference points - like a thread correcting lathe screw plate)?

Bill

 

[fciron]

Asquith You are right.... they must have some sort of measuring fixture to check the threads or mics once assembled and then set it to engrave the lines....... Thanks for sharing the link to the other one. On my 1/2" the line is straight.... but now I'll have to take a close look to make sure.

Wouldn't a one inch gage block (or any maximum measurement) be sufficient for setting the mic?

The thimble would be divided in advance. Next the mic is assembled and the zero position marked. Use the gage block to mark the one inch position. Connect the two points and divide the line appropriately. (If you think the pitch varies, use more gage blocks to check intermediate measurements.)

I'm having trouble picturing a simpler method than letting the mic be it's own fixture.

This has been fascinating. Now I want to try my hand at micrometer making.

 

[rivett608]

Billtodd asked "Are the basis lines 'straight' - in a non-euclidean way "....... this way beyond my elementary geometry skills (or lack of) that I don't even understand the question.... sorry.

fciron I agree you would only need the end points...... but how did they get such a nice consistent line engraved? I can see having the mic in a vise and pricking the start and finish points, laying a flex rule between them and scribe away...... but this looks machine done?

 

[fciron]

I think BillTodd is being sesquipedalian : the line can't be straight in a Euclidean sense because it's on a curved surface. If the basis line is "straight" then the error is regular, if the error varied then the line would curve. (If we measured intermediate gage blocks and they did not fall on the line between 1 and zero then we would have to find a curve that passed through the intermediate points.)

What level of production are we talking about for these things?

I can imagine a sort of helical milling fixture that can be set to the right angle, but I can't get past assembling the mic as being the simplest method of checking the threads. The threads are pretty small to register a tool in and you sure as heck don't want to count them each time. Maybe they had a test micrometer that each spindle could be mounted in?

This would be the part where I throw out the disclaimer that I'm a blacksmith and there could very well be a thread measuring device that every one knows about but me.

 

[rivett608]

Elliot like a lot of this type of English companies most likely made them in small batches...... they also would make custom made instruments for their customers often...... a lot of the English engineers had different ideas and companies would create what they asked.

 

[Billtodd]

I think BillTodd is being sesquipedalian :

Euclidean is only 9 letters

Just to clear this up: The geometry we learned in primary school (kindergarten) e.g. that a straight line is the shortest line between two points, internal angles of a triangle add up to 180° etc. is essentially the work of the Greek mathematician Euclid (Euclid - Wikipedia, the free encyclopedia) hence 'Euclidean'. The geometry of a non-flat surface is thus called non-euclidean.

Sorry for the confusion

All I was trying to ask is: Does it look like it was calibrated by scribing a simple straight line from zero to the 1" mark, or is it a wavy line passing through several reference marks (at say 1/4" intervals) ?

Bill

[edit]

but how did they get such a nice consistent line engraved? I can see having the mic in a vise and pricking the start and finish points, laying a flex rule between them and scribe away...... but this looks machine done?

I'm betting there's an engraver out there with a machine that can do this easily, but I was puzzling about it:

The engraving machine would have to have an infinitely adjustable spiral function - something like a taper attachment for a lathe, but where the motion of the cross-slide instead rotates the work piece as the carriage (carrying the engraving tool) traverses left and right.

 

[Asquith]

I now have an Elliott Bros 0 - 0.6" micrometer, featured in post #15 here:-

http://www.practicalmachinist.com/vb/antique-machinery-history/english-micrometer-1870s-171345/

It did not have a pitch error needing correction by sloping the fiducial line.

In fact, its accuracy is very good.

There is a zero error, and I didn't bother to correct it, but went straight into checking the calibration, using eight slip gauges over the range. There’s no ratchet, so I tried to apply the same feel each time.

The readings proved to be very consistent. Using a magnifying glass, I estimated the reading to the nearest 0.0001". To my surprise, all the checks showed it to be accurate within 0.0001", i.e. no reading deviated from that expected by more than an estimated tenth of a thou.

This accuracy was especially surprising in view of the zero error. I soon realised that the anomaly was probably explained by the relatively poor state of the anvil and spindle faces. When the two faces were in contact with each other, this gave the anomalous measurement - all the other measurements were taken from the flat surfaces of slip gauges.

In fact, close scrutiny showed wear of the faces, and even slight mushrooming of the spindle end! The micrometer had evidently been much used.

The small diameter of the end of the spindle and anvil, and the apparently low hardness, are definite weak points.

In post #2 above, noting the pitch correction on Rivett’s larger micrometer, the late, lamented Karl Sanger (AntiqueMac) wrote 'So, the esteemed Elliot Bros shop could not make a thread accurately?'. Well, Karl would have been surprised to learn that an esteemed US maker also occasionally sloped the fiducial line to correct for a small pitch error!

 

[rivett608]

That is great that you found one! They are things of beautiful workmanship. I think just about everything I have from Elliot Bros. is exceptional..... I would love to know more about their shop culture, what was it like in their shop. I have a drawing set with 3 levels of goodies signed Elliot and Sons, a mark they only used for a few years in the 1850's

You might want to look real close at the knurled end, they didn't use a ratchet but use a Breguet type friction mechanism on some models.

I really miss Karl's whit and comments, one of these days I have to get back to posting more stuff.... Just got a micrometer that measures 1/1000 ths of MM......

 

[Asquith]

Rivett,

I have a couple of square-framed micrometers which have a friction drive thimble, but the Elliott is solid. I don't know what the Breguet type entails, but I took one of mine apart and found nothing more than a leather washer sandwiched between two brass surfaces, pressure applied by a screw.

I for one would be very interested to know more about your 1/1000 mm micrometer, but I can appreciate that you might be reluctant to go to the trouble of assembling a thread about it, given the limited audience (especially when mm and not inches are involved!).

 

[stephen thomas]

.....even slight mushrooming of the spindle end!

Yeah, well, sometimes you need just one more C-clamp. (G-cramp)

The engraving machine would have to have an infinitely adjustable spiral function - something like a taper attachment for a lathe, but where the motion of the cross-slide instead rotates the work piece as the carriage (carrying the engraving tool) traverses left and right.

Just out of curiosity, when did pantographic style engravers come into existence (for "precision" metal engraving applications)?

For small helical angles, I can do this on a Gorton merely by putting the line reference on a swivel plate. For a dedicated machine it would be trivial to make the swivel base long, like on a Tool & Cutter or Cylindrical grinder; to amplify the resolution. I suppose there is a very slight mathematical error introduced; the actual curve would be slightly more hyperbolic (or would it be parabolic? Conics was too long ago for me.) than a true helix. Another way is to use a sine bar; and a follower with the index, under a fixed spindle. Like a rifling machine, e.g. Again, slight error, but very close for small helix angles.

smt

 

[Asquith]

No-one took the bait on pitch error compensation on US micrometers!

However, I’ve been sitting on this for months, but lost heart. Now I’m keen to pursue it. It’s an investigation that anyone with an old micrometer can join in!

I have a suspicion - but no more than that - that it was routine practice in the 19th century (and perhaps beyond) to correct for minor pitch errors by 'sloping' the fiducial line.





This is a Starrett No. 3 micrometer. The fiducial line slopes. The slope is very subtle, but it is definitely present. It is very plain to see when viewing obliquely along the barrel.

I discussed it with ‘Honrick’ some time ago, and he told me that the micrometer was made some time between March 1891 and 1896.

The slope is discrete enough to readily go unnoticed (unlike the slope on Rivett's Elliott mic!) It is there, though. I showed the micrometer to my wife and asked if she could see anything odd about the line, without otherwise prompting her, and she spotted it straight away.

Honrick then checked his early Starretts, and found that on one of the five, the fiducial line is slightly angled.

I’ve checked all my old micrometers, and found sloping lines on just two more, one Starrett and one by Massachusetts Tool Co.

The second Starrett was a surprise, because I assume that it is of a later vintage. It’s a No. 2 model, 0-2" type.

Going back to the one in the photo, it’s the only one with a slope sufficiently large to try measuring.

I mounted the micrometer on a surface plate (holding the frame with a magnetic block), and I set the barrel truly horizontal, using a dial indicator.

I used a vernier height gauge with a knife edge scriber, and set the knife edge level with one end of the fiducial line.

I then moved to the other end of the line and measured the change in height required to get the knife edge level with the line.

The difference was 0.015". I could be a thou or two out, due to the difficulties in getting close to the line and knife edge with a magnifying glass.

Divisions on the thimble are 0.06" apart, so 0.015" equates to a pitch error of just 0.00025" per inch.

I had envisaged a way that the 'slope' could be applied without too much difficulty, but have no idea how it was actually done. It’s basically as already suggested, and in line with Stephen’s ‘rifling machine’ concept.

The engraving machine could be in the form of a basic lathe, with a microscope. The micrometer frame would be set between centres, and clamped to a sort of faceplate. The faceplate has a protruding lug, or a slot, at its rim. The 'carriage' - holding the tool for engraving the fiducial line -would have a bar protruding horizontally from the front. This would engage with the peg or slot on the faceplate. The bar can be set over to a slight angle, set precisely by the operator according to the micrometer's error (see below). Then, as the carriage (and graving tool) is moved towards the faceplate, the faceplate and the micrometer rotate very slightly. Effectively scribing a very fast screw thread! As in rifling.

To determine the amount of angular correction needed, I can think of two ways. First, scribe a very fine horizontal fiducial line on the barrel. Zero the micrometer. Check the error at the 1 inch position with a test block. Use this figure to set the 'correction bar' on the engraving machine.

Alternatively, no fiducial line, just mark a dot at the actual 0 and 1 inch positions. Set the correction bar on the machine to give the helix required to pass through the two points. Too fiddly, I think, for production work.

 

[rivett608]

Hummmmm.... Interesting. I am guessing you are assuming the slight angle of the line on your mic means that the screw is not perfect? A couple things come to mind, if the line were straight with that same screw how far off would it be? If you took a similar vintage mic of the same mic what would happen if you switched screws? Would it fit? Do any of the mics with non straight screws have them angling up(the other direction)?

For some reason in my mind I just can't see Starrett going to that much trouble to make them to this degree of being custom made. But the funny thing is I have no trouble imagining the craftsman at Elliot doing this. I am going to have to look at a bunch of my early mics and see if I can find this on any of them.

Also on the Bregeut mechanism, that is it, just some leather or felt washers and a friction disc or two.....

 

[Clive603]

For some reason in my mind I just can't see Starrett going to that much trouble to make them to this degree of being custom made. But the funny thing is I have no trouble imagining the craftsman at Elliot doing this. I am going to have to look at a bunch of my early mics and see if I can find this on any of them.

When it comes to an old company that is still extant and in the same line of business, like Starrett, its easy to forget how much the relative prices and manufacturing techniques have changed over the years. Today we tend to think of a respectable quality 1" micrometer as being easily affordable by pretty much anyone (in the Western world at least) who really wants one. Even at UK minimum wage barely a days pay. When these were made the real world relative price would have been enormously greater. Probably well beyond what the craftsmen making them could aspire to saving up for. More like years of wages rather than days.

In that context individual calibration and scale marking is perfectly reasonable and affordable, an extra craftsman hour compared to just fitting a standard straight line sleeve being neither here nor there in the price context. But the man paying very serious money for the micrometer wants it right, as right as can be. No gauge blocks then so the micrometer would have been the final arbiter. Shades of Maudslays Lord Chancellor.

Clive

 

[Asquith]

Rivett,

Yes, the implication is that the screw has a pitch error.

According to my measurements of the slope of the line, the pitch error to be corrected was about 0.00025" per inch.

So, if the line had been straight, with the same screw, then the reading at 1.0 would have been 1.00025".

I approached it from another angle. Suppose that the screw was perfect, but slope was accidental. Improbable, I know. Then the reading at 1.0 would have been 0.99975" if the screw was perfect.

So, I did a calibration check, setting aside the unfairness of doing this on a 120 year old instrument! If the line sloped the way it did in error, we would have expected the 1.0 reading to be a bit under 1 inch. In fact it was 0.4 thou over! So, this does not support an accidental error in the fiducial line.

To address another of your questions, the line on my 0-2 inch Starrett slopes in the opposite direction.

I haven’t got two comparable Starrett micrometers to check the effect of swapping spindles. However, I did have a go at swapping the spindles from the two old Starretts that I have (0 - 1" and 0 -2"). What follows may not be comprehensible!

The 1 - 2" spindle was too tight in the 0 - 1" frame, so no check possible.

The 0 - 1" spindle went nicely into the 1 - 2" frame, but it was too short to reach the anvil, and because of the different spindle lengths, the readings didn't mean anything in absolute terms!

Nevertheless, I checked the reading against 0.2" and 1.0" slip gauges. I found a progressive error of 0.0007" over the range checked (0.8"). Correcting to 1.0 inches, it would have read too high by about 0.0009" at 1 inch. No point trying to over-analyse this, but in short, the fact that the spindle gives different readings in two different frames does support the assumption that there’s a pitch error of several tenths of thous per inch.
...........................
Regarding the extra work involved in custom engraving the fiducial line, bear in mind we are talking about the marginal cost, as the frame had in any case to be carefully set up to accurately engrave a 'non-sloping' line. Also, the time and cost of this custom engraving has to be set against the difficulty in making - and testing the accuracy of - the screw, with the possibility of rejection.