Drawing calls out for 96 DP Diamond Knurl.
Dia of shaft is .250
Dia of knurl is called out at .256 +.005/-0

When measuring the knurl I am getting variations in the readings depending on just where I hit the part. These variations are about .001 inch.

We are trying to perform a Cpk check on these parts as requested by our customer.

My question is what dim would I put down for this?
Do I go for the highest point, an average, or the lowest.

Any Ideas?

James

Drawing calls out for 96 DP Diamond Knurl.
Dia of shaft is .250
Dia of knurl is called out at .256 +.005/-0

When measuring the knurl I am getting variations in the readings depending on just where I hit the part. These variations are about .001 inch.

We are trying to perform a Cpk check on these parts as requested by our customer.

My question is what dim would I put down for this?
Do I go for the highest point, an average, or the lowest.

Any Ideas?

James

It would be best to measure the part at say, three different places, and report the range (highest and lowest). Averaging dimensional readings is never a good idea. As far as statistical analysis is concerned, you would be best advised to consult with the customer, discuss the inherent difficulties, and ask for guidance.

From the Hindsight is 20-20 Department: This sort of thing needs to be identified and settled at the contract review stage.

My background is precision engineering and I used to work for a company that made, amongst other things, rotors of varying geometry, they were for the National Cash Registers (hole in the wall) and had several knurled features on them.

I have one observation though, when measuring the knurl, I wouldn’t measure it on each knurled point as you will get massive range differences as you cannot guarantee you have a corresponding knurl point on the other face of the micrometer. The best way to measure knurls (both straight and diamond form), and the best way to measure diameters in general, is to gently spin the part whilst slowly winding the micrometer onto the diameter. You should come to a point whereby the knurl is quite tight in the micrometer but does not stop spinning. If you manage to stop the part from spinning then you have the micrometer too tight. That way you get a sound ‘feel’ for the diameter as apposed to single point measurements. It’s a much improved and accurate way to measurement, I found.

Regards

Lee01

My background is precision engineering and I used to work for a company that made, amongst other things, rotors of varying geometry, they were for the National Cash Registers (hole in the wall) and had several knurled features on them.

I have one observation though, when measuring the knurl, I wouldn’t measure it on each knurled point as you will get massive range differences as you cannot guarantee you have a corresponding knurl point on the other face of the micrometer. The best way to measure knurls (both straight and diamond form), and the best way to measure diameters in general, is to gently spin the part whilst slowly winding the micrometer onto the diameter. You should come to a point whereby the knurl is quite tight in the micrometer but does not stop spinning. If you manage to stop the part from spinning then you have the micrometer too tight. That way you get a sound ‘feel’ for the diameter as apposed to single point measurements. It’s a much improved and accurate way to measurement, I found.

Regards

Lee01

The "best way" to measure anything should always take into account (if the specifications aren't enough) the way the thing will be used and the way the customer is likely to measure it. This means: when in doubt, ask the customer.

What is the knurled feature used for? I have encountered something like this when a feature was tied down to the 'nth' degree, only to find out that the spec. didn't reflect its application! (The feature was a hole in a circuit board, tied to a +/- 0.005" - it was a 'finger hole' to access dip switches):lmao:

I'd check out why the customer is using such a close value on a feature which is formed in a relatively imprecise manner.......:yes:

This OP brought a smile to my face, because in school we all called it 'knurdling' - which drove the teacher nuts:lol:
(Jim, do I need to explain the 'British' terminology..........?:notme: )

Andy

A diamond knurl requires that kind of precision? All of our knurls are reference dimensions, we are far more concerned with the shaft diameter.

Having questioned that, it depends on the tool.
Using a caliper I would measure the diameter in three place and record the highest reading assuming you are concerned with fit or interference with a mating part.

Using an optical comparitor I would look for the widest point.

A diamond knurl requires that kind of precision? All of our knurls are reference dimensions, we are far more concerned with the shaft diameter.

Having questioned that, it depends on the tool.
Using a caliper I would measure the diameter in three place and record the highest reading assuming you are concerned with fit or interference with a mating part.

Using an optical comparitor I would look for the widest point.


I take the reference dimension comment back... just went and looked a few drawings and we don't do that for all. We call out "+- .010 over knurl" in some.

Thinking about this some more has brought the thought that the form of the knurl is more important than the overall diameter. The formation of a full diamond or straight knurl was preferable to the specific diameter. Even if spec'd with +/- 0.010.

Andy

Responding to Jim ’Textbook’ Wynnes comments regarding his ask the customer suggestion. In an ideal world yes, and as instructed in many textbooks (which I might add, tend to be written by scholars who haven’t worked in industry for many many years – and there will be a reason for that) you could do that, but practically, in most instances, hmmm.

To be honest if I had a supplier who asked me how to measure a knurl, I’d be worried no matter what the value of the product issued to the supplier was.

I have never been requested to measure features of a part in a certain way (except for GD&T callouts whereby features are measured against datum’s), but that’s not actual measurement in this sense.

In any way, there is no need to speak to the customer regarding basic measurement principles, I’m sorry to say that measurement of this types is at its basic state, I’m not sure where you’re being educated but to measure knurls and/ or diameters in three places is absolutely ridiculous, I was trying to offer advice in this matter, besides falling off my chair in amazement at this question.

Lee01

Responding to Jim ’Textbook’ Wynnes comments regarding his ask the customer suggestion. In an ideal world yes, and as instructed in many textbooks (which I might add, tend to be written by scholars who haven’t worked in industry for many many years – and there will be a reason for that) you could do that, but practically, in most instances, hmmm.

If the suggestion is that I lack practical experience, you're going to be disappointed. I have it on both sides of the customer/supplier fence, and a lot of it, from doing basic mechanical inspection to managing large quality departments. I have, on innumerable occasions, been the poor schmuck with a part and a drawing and no clue as to what the designer might have been thinking, or what was actually needed. I have no idea what you might think the big obstacle is in (gasp!) actually talking to customers about how best to make them happy. True enough, it's not always possible, but it is in most instances, and I'm speaking from personal experience.

To be honest if I had a supplier who asked me how to measure a knurl, I’d be worried no matter what the value of the product issued to the supplier was.

One would think you'd have better things to worry about. It's not a question of asking a customer how to measure a knurl--I have a suspicion that I was measuring knurls before you even knew what a knurl was--it's a matter of understanding the customer's needs, and making sure that when there's any ambiguity at all, you do whatever's necessary to know and not guess. Look at the OP's question. It's not just a matter of how to measure the knurl; there's also the small matter of a "Cpk check." It appears that the measurement over the knurl might be important.

And speaking as a person who, for the past several years, had dealt with supplier quality for a large OEM, I like it when suppliers call to discuss measurement methods. I don't assume that they lack basic skills just because they want to understand the requirements. The fact that you seem to think that suppliers who want to talk about important things with their customers are perhaps not very bright, I can understand why many suppliers are reluctant to call customers.

I have never been requested to measure features of a part in a certain way (except for GD&T callouts whereby features are measured against datum’s), but that’s not actual measurement in this sense.

We're not talking about being told how to measure something, we're talking about asking questions when ambiguity exists. But if you've never been caught in a position where there's conflict over measurement methods, then...

In any way, there is no need to speak to the customer regarding basic measurement principles, I’m sorry to say that measurement of this types is at its basic state, I’m not sure where you’re being educated but to measure knurls and/ or diameters in three places is absolutely ridiculous, I was trying to offer advice in this matter, besides falling off my chair in amazement at this question.
[/quote]

Everyone's advice is welcome.

I have to side with Jim on this one. I also have many years of experience on both sides of the customer/supplier fence and have seen issues created by the the use of different measurement techniques between customer and supplier (e.g., min/max/average diameter). While all methods may make sense to the supplier, only one makes sense to the actual application.

If the drawing does not make it clear, the only sure way is to ask the customer. I have done this, and the customer has always responded in a positive manner. I have had suppliers do this, and I always received it positively.

The knurl is used as a press fit hold of the steel pin through a hole in and aluminum part with half of the pin protruding from each side of the part.

The customer evidently has been having problems with the pins slipping and is trying to determine the root cause. Hence the request for Cpk data to assure the knurling process is in control.

The measurements as taken by me origionally showed most of the knurls .0005 to .001 inch under the low limit. As measured using Lee's method of spinning yields most parts in tolerance, but at the low limit. Obviously the supplier was aiming for the print dim which gave him no play on the low side. (I hate one sided tolerances!:mad: )

This is one of those times where we are having a disagreement over Best vs Down and Dirty. I have the answer I need and will proceed accordingly. The rest of the discussion is purely accademic.

James

This may be more difficult than just the diameter of the knurl.

Since you are indirectly trying to control a retention force, several factors impact this.
The actual interference fit, which is the min hole diameter and max pin diameter;
the interference area, which is the diameter and the width of knurled area;
and, the form of the knurl itself (i.e., is it fully formed?)

Since this is an interference issue, not a clearance issue, the average diameter is probably the best indicator since material is being formed around the irregularities and any lobing that may be present.

However, Jim's point is still valid. If you contact your customer and propose the method of measurement, but leave it open for discussion, it should be viewed positively by your customer.