Measuring Diameters - argument ensues.

[#2#]

I have a question of all you masters...

I know some, if not most of you feel that a CMM is not good at measuring inside diameters. I have seen/read the reasons several times and I tend to agree.

What never seems to come up is the measurement of OUTside diameters. Is the CMM equally unreliable at measuring these?

What about the tip qualification procedure. That is measuring an OD is it not?

Please debate.

[Matthew D. Hoedeman]

INSIDE or OUTSIDE, the same reason apply, especially if you are using a TP2, the lobbing is a HUGE issue. What I would do for an outside diameter measurement is calibrate 4 tips, lets say as 45,-180, 45,-90, 45,0 and 45,90.

Take 1 vector point hits on each of the 4 'sides' of the diameter, using the correct probe for the job. This will at least give you a true, good reading of the roundness, but maybe no of the diameter, unless you calibration routine for the tip was touching at the 45 degree point. You could also go to the A90, and then you would have a fairly accurate diameter result, since you will be touching on a part of the tip that does touch for the calibration AND that should not be affect by the lobbing of the TP2.

[Wes Cisco]

I feel ID or OD does not make a significant difference in the CMM's ability to measure. It is the D as in Diameter and the algorythems involved that are the limiting factor not whether it is internal or external. My Opinion Only.

[Jan d.]

CMM's CAN measure ID's and OD accurately. NO DOUBTS AT ALL.

HOWEVER....................

1) If you try to measure a small hole with large probe tip, small variations in your vectoring accuracy (tunneling) can cause great deviations... So when you have a 0.1600 hole that you need to measure to +/-0.0001 with a 3 mm stylus: you will find decent position and form on the CMM, but I would always check the diameter with a pin. My rule of thumb: if the feature hole diameter is smaller than twice your ball diameter, verify diameter using a pin. But ALWAYS look for form and position on the CMM though, because your pin will give you none of that.
2) How much of a cylinder do you cover with your CMM? If you measure cylinders (OD as well as ID) with just a 4 point circle in the middle, you're doomed! Your diameters will be awful. So try to measure every cylinder with at least 2 circles (I always use at least 8 points per circle). Lay one circle as close as you can to the bottom. One very close to the top. If you have time, put a few 4 point circles in between. The more circles, the better your end result. Construct a cylinder through ALL the circles. Guaranteed, the diameter results will be just fine...



Jan.

[BOB PARSONS]

I Agree W/ Wes And Matt, For Tight Tolerances We Use Other Methods For Diameters, Namely: Pins, Air Gages, Dial Bores, Or Swing Gages!

[bw_bob]

Here is my problem with the cmm. Its the way dmis calculates the diameter. You have best fit (average)...Max inscribe (largest stud that would fit in id)....min circsc (smallest circle that would fit over stud)

How about a function that actually measures the hole.

[#2#]

Ok guys, well interesting stuff here.

On my machine I am able to get repeatable diameter results on calibrated ring gages with certain tips. With other tips I get terrible results. I have not been able to figure out why and that is why I continue to use an alternate method.

[Wes Cisco]

CMM's CAN measure ID's and OD accurately. NO DOUBTS AT ALL.

HOWEVER....................

1) If you try to measure a small hole with large probe tip, small variations in your vectoring accuracy (tunneling) can cause great deviations... So when you have a 0.1600 hole that you need to measure to +/-0.0001 with a 3 mm stylus: you will find decent position and form on the CMM, but I would always check the diameter with a pin. My rule of thumb: if the feature hole diameter is smaller than twice your ball diameter, verify diameter using a pin. But ALWAYS look for form and position on the CMM though, because your pin will give you none of that.
2) How much of a cylinder do you cover with your CMM? If you measure cylinders (OD as well as ID) with just a 4 point circle in the middle, you're doomed! Your diameters will be awful. So try to measure every cylinder with at least 2 circles (I always use at least 8 points per circle). Lay one circle as close as you can to the bottom. One very close to the top. If you have time, put a few 4 point circles in between. The more circles, the better your end result. Construct a cylinder through ALL the circles. Guaranteed, the diameter results will be just fine...



Jan.

So does that mean if I use your method and measure the top and bottom of a dia. 1562 +.0005/-.0000" by .125" deep lee plug hole and it is belled, egged, or ringed o/s in an area I did not probe and the customer rejects the part, will you buy it? The CMM can only report based on the individual hits and formulas it uses to turn those hits into circles or cylinders. With a tolerance of .0001", that is 1/30th the thickness of piece of paper. A bore gage can tell you more, more accurately, and faster about the diameter than a CMM. The CMM is a great tool, but it has it's weaknesses too. It is not the onestop answer to all of your measuring needs. IMNSHFO

[BOB PARSONS]

Wws

[Peter Warcholyk]

Wes,

I concur. Too often I have various program managers, engineers, etc., that think that the cmm is a "one stop shopping centre".

Quite often, what they need measured can be accomplished quicker and more acurately by another method. They just can't be bothered, so lets just pass it off the the "cmm" programmer.

ZydecoPete

[Matthew D. Hoedeman]

Oh, I hear that! I had a guy come up one day and wanted me to measure the thickness of a block he just got done grinding on the surface grinder. I asked him what was srong with him micrometers and got that 'deer-in-the-headlights' look from him. Whyintheheel would he drag a block from the opposite end of the shop (literally, about 120 yards away) to me to have me check the thickness when there is a calibrated, certified set of micrometers closer to the grinders than to me in the crib? They have all gotten lazy must be, P.T.B when ever possible (pass the buck).

[Jan d.]

So does that mean if I use your method and measure the top and bottom of a dia. 1562 +.0005/-.0000" by .125" deep lee plug hole and it is belled, egged, or ringed o/s in an area I did not probe and the customer rejects the part, will you buy it? The CMM can only report based on the individual hits and formulas it uses to turn those hits into circles or cylinders. With a tolerance of .0001", that is 1/30th the thickness of piece of paper. A bore gage can tell you more, more accurately, and faster about the diameter than a CMM. The CMM is a great tool, but it has it's weaknesses too. It is not the onestop answer to all of your measuring needs. IMNSHFO

Ah, good point. I presume you mean .1562 . I hate those and they fall into my category 1. I try to stay away from probes smaller than 3 mm. So measure size with a pin. But still look at form and position from the CMM.

I guess I was not clear about what I wanted to argue: many people have come to the conclusion that the CMM can NOT measure bores. Therefore they stick a pin in and declare victory. I have seen people ACCEPT bores that were totally oval or slotted due to tool wander. Pin went in great! Form was louzy, "but CMM's can't measure bores, right?". So the part was accepted. Major problem.

So before going off and make the general statement that CMM's can't measure bores, let's be qualify that a little more. Yes, small bores have perticular problems. But with the right measuring strategy, larger ones can be measured accurately. Been doing it for years. And always look at the form of even small holes. It can tell you a lot about that particular hole.


Jan.

[BOB PARSONS]

Oh, I hear that! I had a guy come up one day and wanted me to measure the thickness of a block he just got done grinding on the surface grinder. I asked him what was srong with him micrometers and got that 'deer-in-the-headlights' look from him. Whyintheheel would he drag a block from the opposite end of the shop (literally, about 120 yards away) to me to have me check the thickness when there is a calibrated, certified set of micrometers closer to the grinders than to me in the crib? They have all gotten lazy must be, P.T.B when ever possible (pass the buck).

I FIND MYSELF ASKING THIS SAME QUESTION ALMOST EVERYDAY!

[Wes Cisco]

Jan,
Yes I meant .1562 not 1562. My fingers were too fast for the computer.

I would let you carry on, but I know there are many new folks who read this forum and they need to know that you CANNOT rely on even a Deltronic pin to determine close tolerance hole size. All you can determine with a pin is the size of the top of the hole. If your customers are happy only knowing how big the front or top of the hole is, then wear out those pins, but if they want real size information about the entire hole you need a bore gage. The lee plug hole I described is a great example. In some cases if the bottom .1" of a hole is up to .0005" o/s it can be approved and used because it does not affect the sealing area, or if you talking about a filter or checkvalve, the top may be accepted large if the critical portion of the bore is in tolerance. The same things apply to larger bores that will fit orings. With a bore gage you can give the decision makers the information they need to determine if the part will function as is or not. With a pin you can tell them how big the top of the hole is and nothing else.

[BOB PARSONS]

Fussy Dia.

In Other Words Good Metrology Is Still Good Metrology.