Attribute Data and Capability Study of Holes in Parts

[jimmccafferty]

Hi everyone,

I have been asked to find out if the process of putting in the holes on one of our parts is capable. This part has 810 (that’s right 810) 0.030” +/-0.002” holes in it. These holes are put in with a laser and the customer want to know if it is a capable process. I’m being told that when this job started we sent 10 parts out and had the holes checked for diameter and location and all was good, but now no-one can find the reports.

We have the hole locations as “Program Controlled”, and we do random go/nogo checks for the diameter. Now the customer wants to know if the hole diameter is capable. What do I do?

We have decided to use the go/nogo gage and check 10 parts. 810 holes X 10 parts = 8100 holes. Ok now what. I’ve been reading a lot about np-charts but I’m not sure. Also, I have been reading about defect rates and simulated cpk values. What is the best approach to this?

Thanks,
Jim

 

[bobdoering]

Re: Attribute Data and Capability of Holes in Parts

Is the customer's question of capability on the location or the size? If the location - how is the hole location called out? Are the holes punched out? Do the care about the roundness (go/no go only checks MMC).

 

[jimmccafferty]

Re: Attribute Data and Capability of Holes in Parts

They are only interested in size. The holes are put in with a laser. roundness is not an issue. The qeustion from the customer is "Is the hole size capable?" The only way management wants to do this is with a go/nogo pin and a body! go figure! Thanks Jim

 

[Rand T]

Re: Attribute Data and Capability of Holes in Parts

Is there anything special about 1 hole that makes it likely to have more variation than any other?

I would think just performing the capability on a handful of holes on different areas of the part would be sufficient instead of doing every single one of the 810 holes.

 

[Bev D]

Re: Attribute Data and Capability of Holes in Parts

been there done that! my condolences...

however, this isn't all bad.

a few clarifying questions:

how are the individual holes created: a single laser and incremetning the part? a single laser, incrementing the laser? multiple lasers and incrementing the lasers or part?

how long does it take to create the holes in a single part?
how many parts do you make in a day?

are there other parts run on this machine - I'm poking at hwo many set-ups are done in a day or week...

how exactly does the go/no-go gage work?

can you use a set of pin gages to categorize the hole size? If so I think we can get your sample size down. (I would recommend random sampling of various holes across the part and over more than 10 parts)

 

[David DeLong]

Re: Attribute Data and Capability of Holes in Parts

Hi everyone,

I have been asked to find out if the process of putting in the holes on one of our parts is capable. This part has 810 (that’s right 810) 0.030” +/-0.002” holes in it. These holes are put in with a laser and the customer want to know if it is a capable process. I’m being told that when this job started we sent 10 parts out and had the holes checked for diameter and location and all was good, but now no-one can find the reports.

We have the hole locations as “Program Controlled”, and we do random go/nogo checks for the diameter. Now the customer wants to know if the hole diameter is capable. What do I do?

We have decided to use the go/nogo gage and check 10 parts. 810 holes X 10 parts = 8100 holes. Ok now what. I’ve been reading a lot about np-charts but I’m not sure. Also, I have been reading about defect rates and simulated cpk values. What is the best approach to this?

Thanks,
Jim

Here is a case where the customer does not seem to have common sense. On a hole diameter this small, a go/no go gauge is appropriate. If someone wanted capability of size, one would have to take 1 hole on at least 30 different parts and measure the diameter in the same orientation (direction) on each. It is just too small to perform so I would ask the customer how they would suggest performing hole size capability and then only perform reflecting the capability on 1 hole and not 810.

Location of holes - If the location on each of the 810 holes is shown in GD&T with a positional tolerance, the job become colossal & I would not perform it unless it was written into the contract. Just imagine, 810 holes a at least 30 parts = 24,300 measurements. I have a web site shown why capability of positional tolerances is not practical
https://www.qmsi.ca/capabilityindex.html

If co-ordinate tolerance are used, then we would have to double the number of measurement since we need 2 (length and width) for each hole giving us a result of 48,600 - We would need capability results on 1620 measurements. This is madness.

 

[Bev D]

Re: Attribute Data and Capability of Holes in Parts

I don't think its a question of 'madness'. It is a reasonable request for a customer to ask for the capability of the process to manufacture hole size to spec. 'madness' would be if they dictated how the capability was to be measured, calculated and reported if that approach involved a 'traditional' calculation of Cpk assuming a Normal distribtuion and they specified the sampling plan without regard/lnowledge of the process by which the parts are manufactured...

The OP didn't say the custoemr asked for a Cpk value, only capability. there are smart ways to do this - but we would need to understand the answers to the questions I posted as there are many ways and I would hate to randomly guess and throw out different scenarios, none of which may fit the OP's need.

 

[jimmccafferty]

Re: Attribute Data and Capability of Holes in Parts

been there done that! my condolences...

however, this isn't all bad.

a few clarifying questions:

how are the individual holes created: a single laser and incremetning the part? a single laser, incrementing the laser? multiple lasers and incrementing the lasers or part?

Answer: Laser fires in the center, then moves out the radius, cuts the hole, moves out and the table indexes 1.3333degs, repeat. Only one laser but it is a 5 axis laser head.

how long does it take to create the holes in a single part?

Answer: One hour

how many parts do you make in a day?

Answer: ten, then maybe none for a week (except other part numbers)

are there other parts run on this machine - I'm poking at hwo many set-ups are done in a day or week...

Answer: Many different parts are cut or drilled on this machine.

how exactly does the go/no-go gage work?

Answer: Operators sketch tells him to check a row of holes at 12, 3, 6, & 9 O' clock. No charting is done. Before the next operation a worker uses a pin to make sure that there are no blockages or burrs in each hole. This is only a stiff peice of wire

can you use a set of pin gages to categorize the hole size? If so I think we can get your sample size down. (I would recommend random sampling of various holes across the part and over more than 10 parts)

Yes we do have pin gages. but as stated above we make these ten at a time and then maybe ten more within two days or it could be two weeks, two months, it all depends on the orders from the customer JIT.

And yes you are correct, customer asked if the process is capable not for a Cpk.

Thanks everyone,
Jim

 

[bobdoering]

Re: Attribute Data and Capability of Holes in Parts

can you use a set of pin gages to categorize the hole size? If so I think we can get your sample size down. (I would recommend random sampling of various holes across the part and over more than 10 parts)

That would be my approach. Get a set of gage pines that divide your tolerance into at least 10 divisions. Poke the hole until one goes in and the consecutive one does not. Report the pin size that went in, it is variable data, and you should be able to calculate the Cpk from that.

 

[jimmccafferty]

Re: Attribute Data and Capability of Holes in Parts

Here is a case where the customer does not seem to have common sense. On a hole diameter this small, a go/no go gauge is appropriate. If someone wanted capability of size, one would have to take 1 hole on at least 30 different parts and measure the diameter in the same orientation (direction) on each. It is just too small to perform so I would ask the customer how they would suggest performing hole size capability and then only perform reflecting the capability on 1 hole and not 810.

The above may be the direction we take, unless you guys can help me make this less of a big PIMA.


Location of holes - If the location on each of the 810 holes is shown in GD&T with a positional tolerance, the job become colossal & I would not perform it unless it was written into the contract. Just imagine, 810 holes a at least 30 parts = 24,300 measurements. I have a web site shown why capability of positional tolerances is not practical
https://www.qmsi.ca/capabilityindex.html

Yes, thank heavens the person before me had the sence to do this part up front and then talk the customer into "Program Controlled" for the life of the part.


If co-ordinate tolerance are used, then we would have to double the number of measurement since we need 2 (length and width) for each hole giving us a result of 48,600 - We would need capability results on 1620 measurements. This is madness.

Thanks,

Jim