Precision Measuring

"dave sanderson" wrote in message news: snipped-for-privacy@p77g2000hsh.googlegroups.com... Evening all, a bit of background, my current measuring tools are: A Dial Gauge, which whilst cheap (draper 0.001") has done the job til now. A very cosseted Moore and Wright 0-1" 0.001" Mike, which was my fathers, A secondhand Mitutoyo 0-1" 0.0001" Mike, which seems to agree with the M&W on the few things Ive checked on (I have no gauge blocks)

I am designing a spindle and ancillaries, and will have to do some pretty close tolerance boring, turning, dovetail aligning, and repositioning of parts, for instance according to the bearing manual the bearing bores/shaft need to be -0 to +0.0003" for the external races, and +0 to -0.0001 for the shaft, and Im pretty sure I wont be able to turn all things in a single setting, so resetting to a tenth or less might be important (impossible?). Apart from the fact Im not sure that Im up to the job the only tool I have that measures tenths is the mitutoyo 0-1" mike. Not the most useful tool for internal bores 30mm in diameter... What measuring tools should I look at getting (not all at once, as the budget is limited, and I understand that you get what you pay for)? I can see a DTI in 0.0001 resolution would be useful for this and other jobs and would be approx £50, so that I can justify, especially given the use my Dial Gauge currently gets. I suspect a bore mike would be needed? these look expensive. Any others? Is there any point in getting a 0.00005" resolution DTI?

Tolerances down to a tenth or two require such careful control of temperature that the only way you're going to achieve them repeatably is in a metrology lab with precision calibrated equipment. A 1" steel shaft will expand by just over a tenth for every 10 degrees C. The heat you generate during machining can alter the size by several times this amount. You need copious coolant and then some settling time before you can even measure things to this degree of accuracy.

One of my regular tasks is polishing new bronze valve guides down to the correct size to fit in the valve guide bore in the cylinder head. New guides come with about 1 thou extra on the diameter to cope with worn bores but most heads don't have worn bores. Stick a guide on the lathe and start polishing it down with 80 grit and it expands with the heat generated at about the same rate as you remove material. You polish away for a bit, measure, get down to about target size, wander away for a smoke and by the time you come back the bloody thing's cooled down and shrunk half a thou and is now too small. So what you actually do is get the first half thou off, let it cool for an hour, measure again, take another gnat's c*ck off and proceed in every decreasing circles until you get close to what you want. Even with hand polishing it's hard to work to a tenth. 3 tenths maybe is not so big a deal.

Same thing happens when honing out cylinder bores. The heat generated expands the block so you think you've hit target size and then next day when you measure again it's still half a thou small. The difference in size between summer and winter, say 25C, equates to over a thou on an 85mm bore in a cast iron block. You can double that for aluminium.

So there's really no point in buying measuring equipment calibrated to finer levels than you can actually work to or that the job needs. I can't believe any shaft/bearing needs making to a tenth or it'll stop working. A mike reading in thous can be interpolated to a couple of tenths by eye and one with a tenth of a thou vernier just ain't going to repeat to that every time anyway. Hold the mike in your hand for too long and it'll expand and read wrong by a tenth. Hold the workpiece for a while and the same thing will happen.

Yes you need a decent bore gauge if you're going to try and makes bores to some reasonable level of accuracy but you aren't going to be able to work to a tenth at the best of times. If you can make anything to three tenths you've done very well.

Either you have really tiny bearings or you shifted the numbers one decimal place to the right. The ballbark is about 0.01mm for 10mm.

Nick

?? 0.01 mm = ~ 0.0004"

I meant the 0.0001" (0.00254 mm) And that number is really frightening. Even in metric ;-)

Nick

"Chemical press-fit"[tm]

Nick

I fully support Dave's excellent post on illusory accuracy. A few extra comments may be useful.

With a decent mic, outside measurement to within a few tenths is not difficult.

Inside measurements are more of a problem. As Dave points out a decent bore gauge is the answer but this is fairly specialised and expensive piece of equipment. For general use we are mostly limited to inside calipers, a digital vernier and telescopic inside gauges.

Unless you are VERY skilled internal calipers are not safe to much better than whole thous and digital verniers are no better. Although the digital vernier has high resolution readability it is very difficult to position the knife edge inside jaws both accurately square to the workpiece and also across an exact diameter.

The old design inch verniers with their 0.1" radiused inside jaws were much easier to use. A modern alternative is the jaw type inside mic. This has similarly shaped jaws but since the micrometer scale reads in reverse, beware of catastrophic accidental wrong readings!!

A set of spring loaded telescopic inside gauges is pretty good and fairly easy to wiggle into the square+true diameter position. Most use cylindrical plungers but Starret do an up market version which expands two half balls which is potentially a slightly more accurate system.

Dial gauges are almost indispensable for accurate work and you need at least one, plunger or lever type, that can read to tenths, The plunger type is more robust and will read over a longer range but the lever type is more versatile and can sometimes reach positions inaccessible to the plunger version.

Jim

BBs need a tolerance field of quality 6. For 10mm that would be 0.011mm, for

30mm it is 0.013mm (or 0.016). If it is a precision thing, quality 5 is suggested (with little loads). Then you get 0.008mm and 0.011mm. Doable with a micrometer on the outside. On the inside ... well, you have read the other comments. I'd use a bore gage like this: and a good dial indicator. You can use the micrometer as gage. But not made in the crap-country that starts with a "C".

Nick

Measuring bores is indeed a PITA compared to measuring O/Ds. A few thoughts on the same.

1) Dial bore gauges. I have three of them. The largest, a very nice Mitutoyo, is for engine blocks and reads from 2" to 6" with a range of anvils. The dial reads to a tenth and probably repeats to 2 tenths. It's the fastest and most accurate way to examine a bore all the way up and down for taper, ovality and size. The two small ones cover the range down to 3/4" and are old and a bit less accurate but handy to have from time to time.

One problem with dial bore gauges is measuring a shallow blind hole because you need a certain depth of hole to get the anvils in and also rock the thing over centre a decent amount. I can't use them for measuring 8mm deep valve seat insert recesses on cylinder heads because of this and resort to the telescopic gauges instead.

2) Telescopic gauges. Definitely an art to using these and I wouldn't rate them to better than half a thou if that. Mind you, mine aren't top quality ones and are a bit sticky. Nip them up too loose and they move, nip them up a tad more and they're locked almost solid. You have to get them just right so they can set themselves to the bore size as you rock them. It's generally reckoned you measure a tad large with TGs. One day I must get an accurately calibrated bore and see just how well I do in a series of measurements. 3) Go-no go gauges. Probably as good as anything for really accurate testing but only of use for one particular size. However they tell you little about taper or ovality. If you're boring lots of holes to a size it's not a big job to turn a couple of these from scrap bar on the lathe. 4) Digital verniers. A waste of time for anything more accurate than about a thou. Hard to get them exactly across a diameter and as you can't zero the inside jaws or make any adjustment for wear or damage they can't be relied on. 5) Dial gauge plus test plug. For blind holes you can use a quite neat method of measuring as follows. Turn a plug a good bit longer than the hole is deep and a bit smaller in size. Measure it accurately. Sit it in the hole with a dial gauge zeroed on it and move it from side to side. The plug diameter plus the movement on the dial gauge is your hole size. You can work to a few tenths like this quite easily. 6) The valve/valve guide rock method. This is an old way of evaluating the wear in engine valve guides. Stick a valve of known stem diameter in the hole and measure the rock at the end of the stem as far from the end of the guide as possible with a dial gauge. You need to apply an equation to work out the clearance from the amount of rock.

G= the valve guide length (or hole bore length) V = the valve stem length from the bottom of the guide to the far end of the stem.

Let's say you have a hole 25mm long and a stem 150mm long which therefore protrudes 125mm out of the hole. The stem can rock 0.2mm from side to side at its far end.

Clearance = Rock / ((2V-G)/G)

So in this case if the stem is accurately measured at 8mm the hole diameter is..

8mm + 0.2/((2x150-25)/25) = 8.018mm

Not something you tend to use often but it's handy to know about.

Yes, a nice thing to own, but Im not sure the cost will get past swmbo, however Ill put it on the someday list...

Are these the sort you set then measure with a mike? if so I need a bigger mike... half a thou for the bore would do admirably for this job, and almost certainly for most things in the future. at about =A320 each then a 19 - 32 sort of sized one would be a good start, and I could add to the set as I need them.

Could make one of these, but again I need a bigger mike to measure it accurately...

I dont have a digital verynear, I have a traditional one, and I use it for verynear measurements only....

This seems like a good method without an accurate bore measurer, I can see how it works and everything :) and a good DTI (0.0001") is 'affordable' in a way a really good bore gauge isnt.

This sounds useful,though not for this. Not heard of this one before. Ill file it away for future use :)

Dave

To some extend, you can make longer "sticks" (opposite to the anvil). As you are only making comparative measurement, they don't have to have a precise length.

Nick

I'm nearly finished with my Quorn-ish grinding spindle*) (will be a toolpost grinder) and I ruined a bearing. But it was a cheap one. They will be replaced with the real ones and these really cost!

*) More or less just waiting for the V-belt ...

Nick

That is the kind of thing Im wanting to avoid. Ive only done the design, and not cut any metal yet, just building up to it. not sure what size you need, but for a good lead on bearings check out arceurotrade's website, very reasonable for precision angular bearings. (no business connection, just where i got mine from)

Dave

The telescopic gauges are quite cheap:-

number XC253 £10.21 for a set of 6

My set are not quite as good as the Mitutoyo of M&W ones that I also have, but they are perfectly acceptable.

Sets of micrometers can be had, quite reasonably via the Shows (Harrogate, this weekend) and Ebay

Mark Rand RTFM

I bought a cheapy bore gauge set from J&L some years ago, it was a complete pain in the a**e to use. I don't know if they still offer the same model, but it might be as well to be cautious about cheap unbranded gauges from them (or anyone?). Luckily I sold it on ebay, almost unused, for at least as much as I paid for it. I've now got some proper Mercer gauges, bought secondhand, but be a bit cautious about secondhand stuff like this if you can't see it & play with it before paying (the voice of experience again :-( ).

Cheers Tim

One point that may not be obvious is that the spherical radius of curvature on the outside ends of the "sticks" is important.

If this radius is equal to half the ID being measured it then forms part of a true sphere and the measurement accuracy becomes independent of misalignment in the bore.

A telescopic gauge for a single diameter is not very practical but a radius chosen to be right for the smallest setting is a good compromise and useful improvement over the casually rounded end found in cheap gauges. The Starret half ball type makes good use of this optimum radius principle

This is also the reason why Dave Bakers dial gauge plus test plug method works so well. Location on the blind hole eliminates squareness errors and, because the test plug is very nearly the matching diameter, it delivers a true diametrical reading.

Jim

The all do have the same small radius. Like the feeler tips of dial indicators.

Wiggling the bore gage until you get the right (not the wanted) reading is part of the skill required. The (spring loaded) anvil (or alignment help; U-shaped) helps in aligning in one plane. If you are using longer "sticks", that anvil works less good and you have to pay more attention. Tilt axial until you get the maximum reading, tilt lengthwise for minimum reading.

Nick

Eeek, did I really say that? My bore gauge will never forgive me. It's a Mercer not a Mitutoyo.

BTW, I got it, along with the other bore gauges, a couple of mikes and a dial gauge from Metrology and Quality Services in St Albans many years ago. Late 80s or just into the 90s from memory.

They calibrate and repair just about any measuring equipment and always have second hand stuff available that they've just rebuilt. I went there for the Mercer which was mint and pretty rare being the tenth of a thou one but got distracted by all the other shiny toys they kept showing me. In the end I came away with the Mercer, the other two smaller bore gauges, a mint Mitutoyo 2" to 3" mike, a Moore & Wright 3" to 4" one, both in their original boxes and with setting gauges and a 1" travel 0.0005" increment dial gauge for £200. I'm not sure what that lot would have been new but I think I did OK.

The Mercer's pretty terrifying to put down an engine bore though. It shows up the tiniest of faults and makes them look huge. I'll be honing away at a block I've just bored, stick the Mercer in and think "bollocks, it's now got a huge taper at the bottom end" and then realise it's only 2 tenths not 2 thou. Out of interest I'll mention that every now and then I'll get the chance to measure a brand new OE engine block from Ford or whoever and it's about par for the course to see 1 thou taper and ovality. Much more than you'd think they'd work to but it seems to make little difference to how they go. Piston rings are very forgiving bits of kit.

The other place y'all ought to know about is Sert Tools in Farnham Common near Slough. A treasure trove of second hand workshop equipment of all types which they buy up from closed down businesses and auctions. Racks upon racks of mikes and dial gauges, bins of taps, dies, reamers and drills for 50p to a quid a pop. You have to tip them out on the floor and spend a couple of hours sorting through it all but you can find brand new quality named stuff for pennies. Lots of clapped out s**te too of course. Put an old pair of jeans on cos their carpet ain't so clean and have a dig through. I always come away with a bag full for a few quid. In the 20 years I've been going there I don't think many of the mikes have moved an inch so they ought to be amenable to an offer rather than the sticker price which does tend to be a bit random considering what you can buy decent Polish stuff for nowadays from J&L and the like.

Hmm, now I think I need a crash course on what order the prefered precsion tool manufacturers are, or are there several that are pretty much the same (M&W, Mitutoyo, Starrett) and some better ones (Mercer)?

Thanks for the tips on where to go, these are the sort of things you dont find in books.

Dave

With Mitutoyo, you certainly don't make something wrong. Starrett has a good name (but not here in Germany). Others coming to my mind are Mahr, Fowler, Sylvac (they went completely digital + $$$), Kaefer (for dial indicators).

They are all playing in the same league. Read their specs (accuracy & resolution & repeatability). You can trust them. If some manufacturer doesn't display those values, forget him.

Nick