Cylindricity

How do I go about checking Cylindricity of a part if I do not have a CMM? Parts are as follow:

Material - Nuclear Grade Graphite
OD - 4.200"
ID - 2.378" +-.001" with a Cylindricity tolerance of +-.001
Length - 10.620"


Thanks for your input.

Talyrond?

Or perhaps you are looking for a method using just standard measurement tools?

I would like to know if there is a way to check it using standard measuring tools.

Roll it in a V block. Locate the indicator at several angles to the V tangents to determine lobing and other errors. With a little math you can quantify the error.

Probably set it up on a bench center, put the work between the flanges of a shop-made end clamping mandrel, and dial indicate the OD is the best way. That way you can tabulate the radial run-out in terms of am angular reference.

cant measure cylindricity properly without precision spindle whether it is virtual (as in CMM) or real (roundness gage). V block and bench center method give idea but unless customer agrees to method, it is not cylindricity.

Why wouldn't Forrest's Bench center method be acurate? Flanges with centers on the outside, clamping the cylinder between them and turning on the dead centers of a bench center? Once you got both ends of the cylinder indicated true, you could graph the rest of the cylinder and there would be no spindle runout to deal with. Am I wrong?

Sorry Stewart. One geometrical definition of a cylinder is a rectangle rotated around one edge. A cylinder can be proven and quantified between precision prepared conical centers. A metrology spindle has to be proven before it's certified for use as a part of a calibration instrument. These usually feature conical bearings and a ball thrust bearing. Their precision derives from the large area of the tapered bearing integrating via an oil film. Microinch circularity is not difficult with this basic design but it still has to be calibrated before acceptance.

A test piece rotated between precision centers is one of those intrinsic standard, self checking things like a square, three flatness specimens, or a level. Provided metrological technique is observed, checking a specimen for roundness by rotating it between centers is about as close as it gets, particularly if you reverse it between the same centers, conduct the same checks, and secure identical results.

Gentlemen:

I believe the poster is asking about cylindricity of the ID?

Gene

easymike29 said:

Gentlemen:

I believe the poster is asking about cylindricity of the ID?

Gene

Click to expand...

This is correct, I am looking for the ID Cylindricity. Running it betwen centers on a shaft and checking the OD tells you nothing due to the "Cylindricity" being an isolated feature.

Justice25 said:

This is correct, I am looking for the ID Cylindricity. Running it betwen centers on a shaft and checking the OD tells you nothing due to the "Cylindricity" being an isolated feature.

Click to expand...

I think that you could first ascertain that the OD is cylindrical (within limits) by rolling it between 3 different V-blocks of 60, 90 and 120 degree included angles or even with crudely rough sawn notches in scrap metal. Anything to approximate the angles. This will give you a pretty good idea of its roundness.
After being satisfied of the integrity of the OD you can then indicate the ID using the same setup and applying the same reasoning. If you were trying to split a tenth I wouldn't recommend this test but for +/- .001 I think it would be accurate enough.
You will of course have to indicate both ends and as far in as you can with a long reach indicator.
When in doubt I always ask the customer what method they will use to inspect it.

Gene

Just as a point of reference you can kiss a CMM goodbye at this number over this length unless you have got close to a million bucks to spare.
Why do people trust CMMs so much .
If your total tol. is a thou. you had better be able to check down to at least 2 tenths.

Adcole, Tallyround, Mahar, and a few specialty gage houses can handle this but you will be better off with some purpose built in-house fixtures.
Two LVDTS can remove the error from the rotation axis and a third will tell you where the true cylinder is running. Some not so trivial software work involved here.

Cylindrical checking is not easy. Lots of data points. Making a good part is much simpler than verifying it. This stuff usually relies on real high accuracy rotary tables in the mega-bucks range (not to mention the Z axis you need).
Bob

A Moore measuring machine would easily be able to ascertain cylindricity to .001 These days Moore measuring machines although not that easily found, sell for almost nothing.

Their incredibly accurate spindles make concentricity checks easy. They do have the ability to move in the "Z" axis.

OK. Sorry I'm so late but it took a while to cook up a written description of something fairly complex to even draw. Before anyone starts implementing this method, he should read through this write-up so he fully understands and visualizes the various elements used and their relationship to one another.

What follows is not my briilliant thinking. It's an adaptation of a similar procedure used on a bit larger scale in a test fixture. The object was to determine the roundness of bearing bores in 10 to 100 HP super-quiet motors. At one point, we used LDVT's and data logging equipment for recording roundness data in the 10 millionths range.

Certifying internal roundness using home shop equipment and a knee mill for a test bed. It may be a little eqipment intensive but it can be done. It aint rocket science. You do not use the vertical spindle for anytning but the installation of a dead center. The spindle stays locked and stationary through out the test. This will take some prep.

You will need one (two would be better) 0.0001" graduated dial test indicators preferably of a compact design. Also provide the necessary indicator fittings -and an inspection mirror.

Make a round bar perhaps 18" long having center holes at each end. Cross drill the bar to accept indicator fittings at locations best suited to keep the set-up short. This will take some trial assemblies to see just where to drill and just how to mount the indicator. The vertical length of the bar means you'll be working with the knee all the way down on most knee mills. If your mill has been long in service, now would be a good time to clean off the crusty goo and service the lower ways on the column before you lower the knee over it.

Make two short dead centers by grinding 60 degree conical points on 1/4 dia x 5/8" dowel pins.

Clamp a 3/8" think block securely on the table. Drill and ream a hole for a snug fit with a dowel pin center. Secure the dowel pin from the side with a set screw. Interferance fit would be better. Note the X Y coordinates before you move the table or the saddle so you can return to them if you have to off-set for some reason.

Mount the other dowel pin center in a collet in the spindle.

Make a "bridge" from parallels on 1-2-3 blocks and lightly clamp them so they dont move. The object is to place the bar vertically between the parallels. The parallels are spaced so the work is supported with the end parallel to the table and there is manipulation room to access the indicator on the bottom end. The bar is free to turn with the mounted indicators within the bore of the work. The opening between the parallels is to be roughly centered on the block on the table. We now have an open bottom stage on which to place the work.

Insert the bar through the bore in the work and capture the bar with the centers so the work rests on the bridged parallels. The bar is free to turn on centers and the work may be gently tapped to and fro to center it. Install 0.0001" reading dial test indicators to contact the ID of the work.

Dial in the work by moving it on the parallels not by moving the table or the saddle. If the work is slightly out of round average it noting the amount and orientation of the ellipticity and any lobing present from chuck jaws, clamps, etc. This forms the first entries in the roundness readings you collect. Move the indicator to the opposite end of the bar and repeat the tests. Note the ellipticity and orientation ect.

You've tested the work for roundness in a free state with the axis vertical, Gravity dos not influence the reasings. Photo the set-up for the recorrd.

This test cannot be accomplished in the horizontal position as accurately or with the same confidence because of the apparatus sags an undeterminable amount from gravity.

Anyone want to try cooking up an image of this set-up?

I sort of 1/2 understand. Maybe.

The bar is vertical, running through the center of the workpiece. The workpiece is also vertical.

But I still can't see how a measurement gets made? Nor how to fit an indicator inside a workpiece less than 3" ID.

Brian, How was the race?

The term I used was "dial test indicator" the little guy that uses a lever contact point not the big hionkin 2"+ dia dial indicator. A dial test indicator with a little enginuity can be used in amazingly cramped places.

Dial test indicator and dial indicator tutorial: http://www.mini-lathe.com/Measurement/Dial_indicators/Dial_indicators.htm
(sorry its so basic but I wanted to show the pictures to clarify my meaning)

As I said in my earlier post it will take some thinking to get the right indicator geometry and clearance in the cramped space forced by the mill's limited vertical travel and other interferances. I also cautioned: "This will take some trial assemblies to see just where to drill and just how to mount the indicator."

You don't take an actual measurement using this set-up. You determine the actual diameter by conventional means. The method I describes in post #13 detects only departure from roundness by rotating a dial test indicator array inside the bore to be tested. The axis of rotation is a vertical shaft between centers, an almost definitive metrological evolution resulting in a very close axis of revolution. Dial in the bore, set a zero at max, rotate the shaft in fixed increments and record the dial readings. Be sure to check and report a repeat zero. What you do with one end of the bore you do to the other. Thus you have roundness readings expressed in radii and angular increments.

Forrest - race was great fun until the gearbox broke, which resulted in Great Slowness and a retirement from the race. Will try again at Rd America next Sunday.

As for the cylindricality - I can imagine one could insert a test indicator into the bore for the length of the "quill" plus part of the mechanism.
I guess some test indicators are very small (mine happen to be micron sensitive ones that are a little large.)

But the bore in question is 10.6" long. And of course it could be out of cylinder or out of round anywhere along there.

I guess there's a minimum ID this trick can work for - which is related to the minimum stiff fixture shaft OD + smallest indicator that can fit in the bore with it and have required precision, and be readable via mirror, etc.. (If the bore were, say, 6" diameter, it'd be obvious how this all worked - get a mirror...)

Bryan

They do make some very small DTIs, including some very sensitive Swiss made ones (unfortunately they're not small enough to survive being dropped by fumbling nitwits - DAMHIKT !)

which I can well imagine would work in the OP's situation with Forrest's ingenious solution.

(OK, that said, from here on I'm talking to the 'wider public'. Obligatory "soapbox alert" to those with sermon allergies !) :

It's easy to slip into a rut, and imagine the work being rotated rather than the indicators, making it hard to use the elegantly pure "dead centres" concept, as the pure, simple generative element it is. Which in my book makes it 'top shelf'

In the age of CMMs it's a reflex to reach for the high-tech whenever things get a bit hard, and guys like Forrest are living treasures for bringing us back to 'doing what you can, where you are, with what you have'

I wish I had a bit of spare time to knock up a sketch; apart from the value to others, it would be good to have on file...
Maybe tomorrow will be another day

The stiffness of the shaft is, methinks, not very material, given that it's vertical, and the applied force will be almost exactly constant and of very low value.

I can see a fabricated 'bar' made from two flat bars either side of the indicator body - or a slot milled in a solid bar everywhere an indicator is required. For stability, it would obviously be preferable to avoid welding !

When I fabricated a metrology rig a bit like this once, I just used plain brass rod for fasteners, countersunk the (dowel-like) holes, and rivetted it over, like the way some machinists squares were (are?) put together.

All the above described methods employ good thought processes, but at what cost and
reliability?
In the end, I think it would be more cost effective to find and pay someone in your area with a Federal Form Scan or similar Mahr, Mitutoyo rotary inspection platform to obtain the numbers for you. Your going to fork out $ one way or another.

Maybe because I have access to these units here in my local area, my thoughts are tainted?

alan

Forrest Addy said:

Make a round bar perhaps 18" long having center holes at each end.
...

Make two short dead centers by grinding 60 degree conical points on 1/4 dia x 5/8" dowel pins.

Click to expand...

Think I understand the concept, and I'm not necessarily disagreeing with you, but...
Even given that these dead centers are stationary, wouldn't using the combined conical points and center holes pick up (and magnify, perhaps) the out-of-roundness of whatever machine was used to make them; thus entering some other unkowns into the fray?

Lots of companies make electronic test indicators which don't require that you can see a dial down in a ten inch deep hole. Just the probe goes down the hole, and the meter is conveniently located where you don't have to do a headstand to see it. PLUS you can rotate your indicator 360 degrees, and watch the meter easily. I'm thinking Federal or Mahr, the same meter used on Moore measuring machines.

Three months ago I paid $2000.00 for a Moore 1 1/2 measuring machine, complete with several probes and the Federal gaging head. The machine is VERY NICE. Now that's an economical way to check every aspect of just about any part. For checking concentricity of an internal hole, it's really nice! MIRRORS-WHY? Just the probe goes in the hole.