hysteresis of sub micron indicators

not sure if this is the same as repeatability. what i mean is this: you have a surface rising by lets say 5 mu. you come in from the low side and take a reading. then you come from the high side to the same point and compare? what would be the expeced deviation of an analogue instrument graduated in 0.5 microns? or another example: you have a parallel and place a 5 mu shim under one end. you take a reading somewhere along the top and then remove the shim. to how much movement would you expect the indicator to react?

dian said:

not sure if this is the same as repeatability. what i mean is this: you have a surface rising by lets say 5 mu. you come in from the low side and take a reading. then you come from the high side to the same point and compare? what would be the expeced deviation of an analogue instrument graduated in 0.5 microns? or another example: you have a parallel and place a 5 mu shim under one end. you take a reading somewhere along the top and then remove the shim. to how much movement would you expect the indicator to react?

Click to expand...

Dunno what 5 microns is. School was a long time ago. You'll be more adept at converting than I.

But.. for a 4 AGD dial Hamilton all-mechanical, graduated @ 10 millionths (of an INCH) per division, the paper with it sez repeatabilty can be trusted only to TWO millionths. Dorsey Gage =>/ Dorsey Metrology owned that tooling and still made the DI last time I looked. They, too, posted the specs on their website at the time.

Would we class that as 20% allowable error? +/- ten percent?

20 millionths Mahr "Millimess" I don't have data handy, but it is electronic and "enough newer" the specs are probably online?

So back to your DI.

Does the maker not publish the info you seek, just as Hamilton did "back in the day"?

i have a 0.5 mu hikator, a 0.5 mu mikrokator and the 0.00002" millimess you mention. i dont know how old they are and besides i wouldnt know what to look for. accuracy, repeatability, hysteresis what?

dian said:

i have a 0.5 mu hikator, a 0.5 mu mikrokator and the 0.00002" millimess you mention. i dont know how old they are and besides i wouldnt know what to look for. accuracy, repeatability, hysteresis what?

Click to expand...

I didn't know what to look for, either. "Everything you can find" works. Even if you have to Google the definitions and value of what they publish separately.

FWIW-not-much:

Add Cadillac Gage Pla-Check, B&S Height-i-cator, Starrett/Weber Gage Digi-Chek.. and a lot more.

Nearly ALL of it under THIS roof is for the fun of research and learning, gets little or no actual USE.

So yah hit the internet, find the documents, try to 'sess out what matters off the input from what those who built and sold them - and their competitors, of course - considered important enough to put into the tables of specifications as they competed with each other and for Purchase Orders from customers. The useful parts of the reality behind the brag, lie, or hard facts, IOW.

That info is very nearly all I have.

OTOH, it has been instructive "enough" for any practical USE. I am only using the goods, not re-inventing them.

PM - where EXPERTS who made their living USING goods I mostly just covet and fondle these late days - fills in any blanks.

So you/we are in the right venue by being on PM.

The Metrology Forum just might not be the most actively watched place to pose a question and get rapid answers.

dian said:

i have a 0.5 mu hikator, a 0.5 mu mikrokator and the 0.00002" millimess you mention. i dont know how old they are and besides i wouldnt know what to look for. accuracy, repeatability, hysteresis what?

Click to expand...

I don't know about the hikator or millimess, but the mikrokator should have zero backlash and zero hysteresis. The mechanism is based on flexing metal and has no rubbing or rolling parts to create friction.

Johansson Mikrokator - Wikipedia
YouTube

jmkasunich said:

I don't know about the hikator or millimess, but the mikrokator should have zero backlash and zero hysteresis. The mechanism is based on flexing metal and has no rubbing or rolling parts to create friction.

Click to expand...

No it isn't 100% free of hysteresis, Yes it does have friction. Even challenges to repeatability.

They have simply been "banished" to the eye of less-critical observers by pushing them down to the lattice and molecular structure level INSIDE the metal.

Which never QUITE returns to a pre-disturbance state and position after BEING disturbed.

yes, the mikrokator is extremely sensitive. but the needle always trembling doesnt really help. well, maybe im trying to do something that cant be done with my equipment. thats why i asked. e.g., i put two parallels on the stand together and when i come in from one side one of them is higher, when i come in from the other side the other one is higher. in both instances the plunger has to come up from the gap to the surface, so im clueless. its clearly not a random error. and i have rotated the parallels to exclude any irregularities of the plate.

dian said:

yes, the mikrokator is extremely sensitive. but the needle always trembling doesnt really help. well, maybe im trying to do something that cant be done with my equipment. thats why i asked. e.g., i put two parallels on the stand together and when i come in from one side one of them is higher, when i come in from the other side the other one is higher. in both instances the plunger has to come up from the gap to the surface, so im clueless. its clearly not a random error. and i have rotated the parallels to exclude any irregularities of the plate.

Click to expand...

On the experience, you are looking in the wrong place. What is CARRYING your indicator so as to "come in from" a choice of more than one direction of approach to "static" work? Same equation and same question if the DI is "fixed" to a spindle or head or bridge and it IS the work, on a slide table, doing the moving to "come in from".

If a lathe or mill traverse? THOSE - and/or their lubes - just "do not" slide exactly the same way, both directions. See also surface grinders and "surfing".

At least - not down at the uber-fine measure you seek to utilize.

If it is the height gadgetry being slid atop a surface plate on its base-buttons, it is STILL not "immune" to variances in handling or the environment.

Have a "clean room"? Dust particles you cannot even SEE or feel may as well be a graveled road at these units of measure.

Working with BARE HANDS? The human body outer layer is DEAD skin cells, constantly renewed from new ones growing under it.

Good system. Keeps us from bleeding out or having to slither about all wet - like pond scum. It ALSO dumps an outrageous aount of dead skin cell debris, all the time, and WORSE just after a shower or bath.

Try your approach at a dead-slow crippled snail's rate of movement. With pauses to settle. You may see a change? Pick the one it is most predictable and repeatable with. Use that one, going forward. Else correct from a log of consistent errors, by chosen direction.

"And then...."

In general, a(ny) machine-tool's variances are GREATER, and not my just a little bit, than any hysteresis in the metrology gear of that fine-grain a class as you are applying.

SIP and Moore addressed such things. Tier One grinder makers as well. Few others played in the same league, then or now, that ordinary mortals could afford to purchase.

Message there is that if/as/when you get THIS part "dead nuts repeatable", you still have to go and do something USEFUL with it in the less-perfect arena.

In Dian's example of coming into his two parallels from opposite directions and getting different results. I would experiment with vibration to see where the indicator came to rest. In a similar but not identical testings setup, the place I used to work used dead weight testers to calibrate pressure gauges. It's just a carefully machined and measured piston and cylinder with calibrated weights on top of the piston. So in principle, you know the area of the cylinder and the weights installed so you get pounds per square inch or equivalent metric. In actual use, the operator would spin the weight and piston assembly while doing measurements to eliminate stiction errors.

The equivalent for a test indicator would be a small vibration to the test setup to see where the measurement settled. This doesn't resolve inherent repeatability or accuracy but ought to sort out your direction discrepancy.

In our business of fluid power test stands there was a somewhat similar problem of establishing measurement. An operator was directed to adjust a valve to achieve a particular pressure on a mechanical pressure gauge. In the real world with a a pulsating hydraulic pump and vibration of the whole setup, the operator would observe the needle of the gauge vibrating back and forth between two or more numbers. How does he discern what exactly 2500 psi is? Welcome computers. With electronic pressure gauges and a computer program that can compute a ten or twenty count running average, it can make a determination of when 2500 psi is achieved far better than human eyes and hands.

Long term, for your gauging application you could think about computerizing it. Let the computer control movement back and forth on the surfaces under test, take continuous measurements, bring in the calibration specs and characteristics of the gauge head and associated electronics and establish a reliable measurement. It starts to sound a lot like a modern CMM but that may be what you need to actually achieve what you're asking.

i come from the side on different points because this seems to give more consistent readings than riding the surface. you go in one direction and get other readings that in the other. of course i agitate the instrument, tilt it, use different tips like ruby or large radius (40 mm), move the part in a small circular motion etc. what repeats very well is lifting the plunger with the cable (or whatever you would call it), but only the digital mitutoyo has this feature and you only measure one point. i also tried putting and additional spring on the plunger.

in the end my measuring resolution is around 3 mu with all these contraptions. yes, i should go electronic, but i find electonics a big hassle.

This problem is always there, when the measuring equipment becomes more sensitive and has higher resolution, in an environment that is not set up for extreme-sensitivity measurements. I had a Heidenhain electronic gage system some years back, with resolution of 10 nanometers at its most sensitive setting, and it was a motorized measuring spindle mounted on a vertical stand, all made by Heidenhain as an integrated system. It was useless, at that sensitivity, in any setting other than on a vibration-isolation table in a very quiet lab, and even then, not great for stability.

The larger the parts are, the more uncertainty contribution you get from the support surface shape and the interaction of the part with that surface. Thermal contributions also occur from handling the parts during measurements.

Regarding the use of flexures in metrology applications, I can say with some certainty that a well-designed and well-manufactured flexure component is likely to be extremely repeatable in positioning over repetitive measurements. Flexures are often the basis for the motion required to make nanometer-level measurements in interferometric applications, and they have demonstrated very good performance as such.