Microns …

[John HaineParticipant @johnhaine32865]

Resolution, precision, accuracy…

From Michael's photo and the context, this is a controller for a microscope focuser. I can well imagine that at very high magnification it would be necessary to move the lens very precisely, for example to focus on various layers of the object. By having calibrated Z you could estimate the distance between layers. Absolute accuracy will not be crucial; small, repeatable, precise movements will be; good resolution critical. Somewhere it says this is a piezo positioner – make small movements by changing the volts across a ceramic element probably.

With regard to the decimal point delimiter, though the "continental" system still uses that in principle anything made for a world market will probably use a full-stop.

[GeorgineerParticipant @georgineer]

Personally, I only work in zeptoparsecs (zpc), which saves a lot of argument.

My son uses the Metric Beardsmore, which is the length of his boss's forearm.

George B.

[old martParticipant @oldmart]

I have micrometers capable of measuring microns, but it is extremely difficult to get repeatable results at this scale. When I was working in a firm making parts for aircraft, there was rarely any requirement for drawings to call up tolerances of less than +- 10 microns, that is less than +- half a thou.

Those people who like to boast and have deep wallets can buy themselves a Mitutoyo digital micrometer with a tenth micron scale.

[SillyOldDufferModerator @sillyoldduffer]

Posted by old mart on 05/07/2019 16:01:52:

….

When I was working in a firm making parts for aircraft, there was rarely any requirement for drawings to call up tolerances of less than +- 10 microns, that is less than +- half a thou.

Those people who like to boast and have deep wallets can buy themselves a Mitutoyo digital micrometer with a tenth micron scale.

 

A digital micrometer reading tenths of a micron would be worth boasting about! But I think Old Mart has his Microns confused. A micron is much, much smaller than a thou.

In this picture the horizontal line represents half a thou (0.0005" ), and on the same scale the circle on the end is 1000µm in diameter:

The horizontal line actually ends in a ±10µm vertical line, 20µm in total. It's too small to see. Zooming in on the circle reveals:

Putting it another way, if flying Brisbane to London Heathrow represented half a thou, then 10 microns would be within walking distance of the airport.

Dave

Edited By SillyOldDuffer on 05/07/2019 16:49:42

[Former MemberParticipant @formermember53456]

[This posting has been removed]

[old martParticipant @oldmart]

Read my post again. 10 microns is less than 0.0005". 0.0005" is equal to 12.7 microns exactly.

One micron is a millionth of a metre, or one thousandth of a millimetre.

One thousandth of an inch is 25.4 microns.

[Anonymous]

Oh dear, we seem to have a CAD catastrophe. If the line is half a thou long then the circle at the end cannot be 1000 microns, as that is one millimetre, which is definitely bigger than half a thou. The scaling factor is 25.4, so more, but not much, much more?

Andrew

[Nigel McBurney 1Participant @nigelmcburney1]

Microns !!! theres millions of them in an inch.

[Mike PooleParticipant @mikepoole82104]

Don’t forget to add a temperature controlled room when shopping for your micron micrometer.

Mike

[Kiwi BlokeParticipant @kiwibloke62605]

That non-temperature-controlled apparatus thinks it can resolve 1nm repeatably enough to justify the scale's resolution? Delusional, surely. It reminds me of a friend tickled pink by his small bench-top oriental CNC mill. He discovered that his CAD software could work to microns, and sincerely told me that he could now machine things 'accurate to a few microns'. Hmm. But then he also thinks that scraping has been made obsolete by modern CNC machinery. Hmm again.

[Neil LickfoldParticipant @neillickfold44316]

When fitting pistons for race engines where between good and worn out engine is 2um on diameter , being able to turn and fit a new piston without lapping it, shows a performance edge with the stop watch. The other option is to turn a lot of pistons at slightly difference sizes and then find the best one that fitted. In realty, making 4 different pistons of different sizes and then finding the best fitting one, takes the same amount of time, but you have 3 spares that may never be used.

If I could have something that allowed me to turn a piston to 0.1 um, and be round to that tolerance, I think would be really great.

Of course the real biggest issue is having the material that can be stable enough to be sized to 1um or better. Until powder metallurgy arrived, pistons for model engines could not be sized to 1um at all.

[old martParticipant @oldmart]

Pistons fitted to petrol and diesel engines are never turned round, and are also tapered and if anybody tried to measure one to a 2 micron tolerance they would get a different reading each time.

Edited By old mart on 07/07/2019 11:44:25

[John HaineParticipant @johnhaine32865]

Posted by Kiwi Bloke on 07/07/2019 09:41:54:

That non-temperature-controlled apparatus thinks it can resolve 1nm repeatably enough to justify the scale's resolution? Delusional, surely.

So do they make it that way as a joke? It may well resolve to a nm, If it could do it repeatedly it would be precise to a nm, and if it could be calibrated (e.g. w.r.t. wavelength) it might be accurate to a nm. But for its probably purpose, making comparisons, it's probably fine.

[SillyOldDufferModerator @sillyoldduffer]

Posted by old mart on 05/07/2019 16:55:13:

Read my post again. 10 microns is less than 0.0005". 0.0005" is equal to 12.7 microns exactly.

One micron is a millionth of a metre, or one thousandth of a millimetre.

One thousandth of an inch is 25.4 microns.

Posted by Andrew Johnston on 05/07/2019 16:56:07:

Oh dear, we seem to have a CAD catastrophe. If the line is half a thou long then the circle at the end cannot be 1000 microns, as that is one millimetre, which is definitely bigger than half a thou. The scaling factor is 25.4, so more, but not much, much more?

Gentlemen, of course you are right. 1 thou = 25.4microns. How do these mistakes happen? Perhaps Nurse will explain after my afternoon nap…

Dave

[Michael GilliganParticipant @michaelgilligan61133]

Posted by John Haine on 07/07/2019 12:06:07:

Posted by Kiwi Bloke on 07/07/2019 09:41:54:

That non-temperature-controlled apparatus thinks it can resolve 1nm repeatably enough to justify the scale's resolution? Delusional, surely.

So do they make it that way as a joke? It may well resolve to a nm, If it could do it repeatedly it would be precise to a nm, and if it could be calibrated (e.g. w.r.t. wavelength) it might be accurate to a nm. But for its probably purpose, making comparisons, it's probably fine.

.

The specification sheet [*] makes clear what it can do, in terms of resolution and repeatability.

@ most contributors to this thread: Please bear in mind that the overall travel is only 0.5mm … and temperature effects are proportional to the relevant dimensions.

Scale everything up to suit [say] a 500mm long milling table and you may appreciate that Marzhauser is far from delusional.

MichaelG.

.

[*] see my opening post

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Nigel McBurney 1 on 05/07/2019 17:19:07:

Microns !!! theres millions of them in an inch.

.

Would you care to recalculate that ^^^

[Robert Atkinson 2Participant @robertatkinson2]

As said earlier, nanometer resolution is completly different from nanometer accuracy. I used to be one of the designers of machines that put 100,000 spots (100 each of a 1000 different samples) of DNA on 25mm x 75mm microscope slides 200 slides at a time. The feedback for the linear motors had 1 micron resolution. Accuracy was checked with a Renishaw laser interferometer which had nanometer accuracy and included compensation for the effect of air pressure and temperature on the laser light.
The arraying tip (split tungsten pin) position was compensated for thermal expansion of the frame and drive system. A traditional positioner like a CMM was not fast enough, we had a peak axis speed in excess of 3m/s.

Robert G8RPI.

[Phil WhitleyParticipant @philwhitley94135]

So, how many angels will actually fit on the head of a pin…………………….

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[old martParticipant @oldmart]

Never mind about the microcockup, S O D, next time its my turn.

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[Neil LickfoldParticipant @neillickfold44316]

I am trimming my pistons on a 1972 Myford Super 7 with 3phase motor VFD controlled with segmented belt drive drive, few would believe me. My smallest reliable cut is 1um on diameter, but have taken 1/2 um cut on diameter when fitting a piston for Glen Lewis. Rob is always impressed by what I am doing off a Myford S7. They had a new one back in 73 or 74 and it was not good enough for their liners or pistons back then. I have however refitted the spindle to bearing assembly, reset the bed clearances, and added a constant gravity oil supply, and refitted the compound slide as well. But nothing that anyone else could not do with time and care. Oh, and made a M8X 0.5mm pitch cross slide screw, but it is a normal RH thread, not the normal lefthand thread that is commonly used. The 0.5mm pitch with the hundred thou dial, makes it read just like a metric micrometer. My lathe is in room at around 20c temp to help with things. So do get there, I have the compound slide on a slight angle, so that 4 thou on the compound is 1um of X axis movement. The original compound as supplied was not capable of this fine enough and repeatable movement, so was reworked to be able to do so. I have a nice 1um 0-30 Tesa micrometer for measuring the pistons with in Combination with small Vee block. It gives a finer reading than just using the micrometer on its own. Another advantage is it allows you to see if the part is round as well. With the liners they are tapered so make for a really good gauge. If the liner is out of round you have to fit tighter to allow for wear etc to get a really good seal. Some piston alloys, even though they are from sintered stock, just will not stay round. So avoid those when possible. Some materials wont keep round to 8 microns or so. Typical taper is 2.5 micron per mm, but some have less taper like 1.2 micron per mm, while other earlier liners had lots of taper like 3.8 micron per mm. The only tool I have found to be able to take the very fine cuts on pistons is the very nice Summitomo PCD insert NF-DCMT11T304N-LD DA1000

Here is a picture of a piston being trimmed before I replaced the rear bearing set and reset the front taper and spindle again. and another just recently. Looking through the exhaust port , you can see the new bearing and refitted spindle is making a much nicer surface finish on the piston. It is hard to see in the 1st pic, but a 1um cut makes very fine sand is the way I describe it. I hand feed at around 0.02mm /rev, it runs around 700 or so rpm, and takes around 1 minute per pass for the less than 15mm length of piston. So that is like around 4 seconds per mm or so or 1/4 of a mm per second how ever you want to work the time/distance out. I don't use power feed as the surface finish becomes junk in comparison and it effects the repeatability of the carriage. Having an old fashion clock with a moving second hand is really helpful in gauging feedrates etc I find. I used to trim with the pison on a fixture and then fitted them to the rod. But with the new press fit pin pistons, the assembly has to be done with the rod in place, so a modified version of the fixture shown with a holder that held the rod and tightened from the back end through the spindle was made. Some days I don't trim pistons, due to temp in the room conditions, or if I don't have the right mindset to do it.

 

2019

 

Edited By Neil Lickfold on 07/07/2019 17:09:08

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[old martParticipant @oldmart]

Neil, you may be interested in a variation of fine feed rates for lathes which I had no alternative but to employ.

We stripped out the leadscrew, apron and rack of the museums Smart & Brown model A for some serious spindle and bush refurbishment. the lathe was still needed for much of the work, so I used the tailstock to push the saddle with great success. The tailstock leadscrew is 8tpi and I had already made a collar for it calibrated 0-125, so a fine feed was actually easier than using the rack.

[Neil LickfoldParticipant @neillickfold44316]

That is a good idea as an alternative to the carriage wheel. Being centred will have advantages as well.

Thanks for the idea.

Neil

[Author]

Posts