Measuring location of NPT thread on CMM

[mhh]

Hi.
We're having some trouble with the location of some NPT threads.
Our CMM programmer have been scratching his head in a way that is leading to baldness! In his head he could find the center of a NPT thread the exact same way as a normal thread. (probing in a pitch)
In my head that would lead to a error in location as he is probing different diameters when he is using the pitch.

It was quite hard to convince him that this was the case since the first we tried was probing a taper with a small angle, causing the center to have very little runout....
By little I mean .018mm. which in my mind is a lot considering I tried probing to find the runout all the way up 1mm at the time and getting about -/+0.001mm

Then I got a BT40 and asked him to do the exact same with that and then he believed me!

But how then? I have very little training in using CMM machines, almost everything I know is selftaught.
Do we measure the taper first and then make some kind of calculation?


Of course we could just turn some plugs and probe on them but it would be nice if the CMM could do it.


Any help, experince and thoughts are welcome.

Mads.

 

[Red James]

This is a perfect example of when NOT to use a cmm. There can be no reason to know the location of an NPT thread to that degree of accuracy. This is much like using a lazer interferometer to measure the length of a snake.

 

[mhh]

This is a perfect example of when NOT to use a cmm. There can be no reason to know the location of an NPT thread to that degree of accuracy. This is much like using a lazer interferometer to measure the length of a snake.

I am not going to disagree with you on that! What really is silly is that the reference points is from a cast surface, and one of them is where they pour the metal so it has also been ground with a anglegrinder!
This is not something that we want to do but the customer wants full CMM reports on some of the parts.
In my oppinion it would actually be enough to just look at it and see if it was located nicely.

But my question is not about how stupid it is and how people rely waaaay to much on CMM.

 

[The real Leigh]

Measure the radius of two adjacent threads (r1 and r2);

Calculate the average ( ( r1 + r2 ) / 2 ).

Compare that with the radius of the thread exactly opposite that calculated point.

- Leigh

 

[Red James]

Measure the radius of two adjacent threads (r1 and r2);

Calculate the average ( ( r1 + r2 ) / 2 ).

Compare that with the radius of the thread exactly opposite that calculated point.

- Leigh

None of the threads are radii.

 

[The real Leigh]

All circles have radii, by definition.

If the circle is on the surface of a taper, the radius will vary with the position along the axis.

That's the whole point of calculating the average. It creates a pseudo-radius at the point midway between the two crests.

That point is exactly opposite the point being measured on the other side, and should be exactly the same distance from the axis (i.e. radius).

- Leigh

 

[gregormarwick]

It's a half hour job to make a plug. Wouldn't that be easier?

Since the customer just wants a nice looking report with no basis in reality (angle ground datum plane?!) just use a tap...

 

[brian.pallas]

Do you have an NPT gage? You could just beep the handle of that if all you're worried about is a half-way decent looking report. Or buy a fitting from the hardware and use that if you don't have a gage and don't want to buy one.

 

[The real Leigh]

Or buy a fitting from the hardware and use that if you don't have a gage and don't want to buy one.

Can you identify a hardware store that sells plumbing fittings with NIST-traceable Certificates of Calibration?
Or in fact even products that include mechanical drawings with toleranced dimensions?

- Leigh

 

[CarbideBob]

It's just a helical grooved cone, what can be so hard?

Seriously, on a CMM you will need to define the start of the thread dead on to get good numbers.
Points taken to measure the part (which as stated above are not a circle or radius, the material you need for a circle at a fixed depth is gone) must be offset in rotation to get anything that makes sense.

I would make measuring plugs, even with a lot of holes in one part it will go faster on the CMM.
It can be done with a taped edge disc stylus but it will take a lot of probing time and calculations to define where in the helix you need to probe the final points.

Probing threaded holes is never a good idea as you need to worry a lot about the probe tip and cylinder/tapered tips are never 90 degrees to the part.
Bob

 

[dstryr]

Easy...pull one off without the tapped hole...inspect the true position of the start hole for the tap...

 

[Heavey Metal]

3 points in a helical patern

 

[gbent]

Don't worry about it being tapered. Just probe it as if it were a standard tapped hole. Make sure your upper points are approximately the same depth, and your lower points are also all at about the same depth. Take a couple of points on the surface around the hole, and create your intersection at that surface. If your probe is 1/4 the hole diameter or larger you should be ok on accuracy. Depending on your software you can also probe it as a tapered hole with little more difficulty. But one of the keys is establishing the dimensions at the mouth of the hole. The axis as probed may be sufficiently angled the dimensions where the hole axis intersects the datum will be out of tolerance as well as unrepeatable.

If the crudeness of this method bothers you, validate it before discounting. Use a tapped hole and pick it up several times using the method I described. Compare the locations. Now validate the runout of the shank of a plug gage for the hole. Install the plug gage and pick up the shank. Compare to the original centerlines. Now compare to the print tolerances. What percent of the print tolerances are used?

 

[upnorth5axis]

First part going to the CMM it is standard not to tap an NPT thread if there is any question about location.

 

[Eric M]

Do you have an NPT gage? You could just beep the handle of that if all you're worried about is a half-way decent looking report. Or buy a fitting from the hardware and use that if you don't have a gage and don't want to buy one.

If you have a taperlock-style thread plug gage, you could remove the gage end from the handle and use the gage end as you would a tapped hole location gage. Since the gage end (that fits into the handle) is tapered, you'll probably have to probe it as a cone to get the centerline of it.

 

[The real Leigh]

Points taken to measure the part (which as stated above are not a circle or radius, the material you need for a circle at a fixed depth is gone) must be offset in rotation to get anything that makes sense.

The procedure I gave back in post #4 establishes the reference point on the circle that intersects the taper at that point.

This is just basic geometry.

The intersection of a cylindrical surface (tapered or not) with a plane perpendicular to its axis defines a circle.

That's why it's called a "cylindrical surface".

- Leigh

 

[CarbideBob]

Yes... if the material you needed to probe existed.

It s helical, conical groove, a fast spiral tapered flute.
It is not just a tapered hole which would be falling off a log simple.
The points across a plane will not work.
No two points can be taken at the same depth with a standard probe tip.

It is not a tapered cylinder to a ball probe and treating it as such will drive you crazy trying to hold size and position.

I'm not sure how you have had good luck with this approach in the past on a CMM.
Very large dia stylus and manage to pick up the inside of the crest without too much error from the ball C/L drop off?
Even then most realize that you can find the loaction of a threaded faster from the crest or high points. It's not what the mating part locates off of.
Bob

 

[The real Leigh]

Bob,

You obviously did not read the technique I proposed way back in #4.

Criticizing something without reading is is minimally productive.

- Leigh

 

[Gordon B. Clarke]

Hi.
We're having some trouble with the location of some NPT threads.
Our CMM programmer have been scratching his head in a way that is leading to baldness! In his head he could find the center of a NPT thread the exact same way as a normal thread. (probing in a pitch)
In my head that would lead to a error in location as he is probing different diameters when he is using the pitch.

It was quite hard to convince him that this was the case since the first we tried was probing a taper with a small angle, causing the center to have very little runout....
By little I mean .018mm. which in my mind is a lot considering I tried probing to find the runout all the way up 1mm at the time and getting about -/+0.001mm

Then I got a BT40 and asked him to do the exact same with that and then he believed me!

But how then? I have very little training in using CMM machines, almost everything I know is selftaught.
Do we measure the taper first and then make some kind of calculation?


Of course we could just turn some plugs and probe on them but it would be nice if the CMM could do it.


Any help, experince and thoughts are welcome.

Mads.

Mads, I've sent you a PM. Og jeg taler dansk

 

[rsegriggy]

How tight a tolerance do you need to hold? Can you put a note on your CMM report that says "visually checked"?