We have a pair of stamping parts with 4 holes each. The drawing calls out that the hole locations are critical. We use a digital probe that can read hole location in x and y direction(y perpendicular to x) to measure the hole location.

I hope somebodies with similar experience can shed some light on how I shall structure the control charts so the operators on the floor can easily and quickly chart the readings?

Thanks for the help!:thanx:

Is there a geometric tolerancing call-out for the hole locations,
or do they have standard +/- tolerancing ?
Ed

it is sort of standard GD&T call-out. I have difficulty use this keyboard to type in the symbols.

basically, like this: Dia. 1.0 MMC [A, B(MMC), C(MMC)]

the simpliest way I think is to chart x and y separately. However, some information has got lost.

I am wondering if someone in the cove have met the similar situcation before.:thanx:

So use the deviation from nominal for the X & Y,
X(squared) + y(squared) = ______ (take the square root) (times 2)
that will give you the number you need. Hopefully it will under 1.0.

You probably don't want the people on the floor using the maxium material call out, tends to complicate things.

Ed

it is sort of standard GD&T call-out. I have difficulty use this keyboard to type in the symbols.

basically, like this: Dia. 1.0 MMC [A, B(MMC), C(MMC)]

Hi Roland,

I think, I have an idea what this GD&T callout looks like, but is it possible to scan (a part of this) drawing and attach it to your post?

- Does the callout apply to a four-hole pattern?
- Or, are the holes controlled independently?
- The datums and basic dimensions would be of help too.

It is very difficult to help you with just a description of the callout.

So use the deviation from nominal for the X & Y,
X(squared) + y(squared) = ______ (take the square root) (times 2)
that will give you the number you need. Hopefully it will under 1.0.

You probably don't want the people on the floor using the maxium material call out, tends to complicate things.

Ed

By all means, allow the use if the MMC principle.

Not allowing the use of the MMC principle would make the part unnecessarily expensive.

So use the deviation from nominal for the X & Y,
X(squared) + y(squared) = ______ (take the square root) (times 2)
that will give you the number you need. Hopefully it will under 1.0.

You probably don't want the people on the floor using the maxium material call out, tends to complicate things.

Ed


If I do that, I will have some misleading information. For instance, the difference between two same readings 0.3 and 0.3 may be 0.6 or 0 or other number between them.

Do you know what I mean?

to make live easier, I am going for X-MR chart:

X=(x^2+y^2)^1/2
MR=((dx)^2+(dy)^2)^1/2

dx, dy stand for MR-x and MR-y

But is it to much the calculation for the operators? any suggestions?

Thank you for the help!

Not allowing the use of the MMC principle would make the part unnecessarily expensive.Exactly. First of all, it would mean throwing tolerance away without reason, but even more important is the real beauty of MMC: That it allows you to use a function gauge. I think the operators would like that?

If I do that, I will have some misleading information. For instance, the difference between two same readings 0.3 and 0.3 may be 0.6 or 0 or other number between them.Not at all, you can do it graphically: All you have to do is to plot the deviation from the nominal positions in x and Y axis into a coordinate system. From then on I am uncertain as we need to see the drawing, but the MMC may provide you with even more tolerance width to play with.

We really need to see the drawing.

/Claes

By all means, allow the use if the MMC principle.

Not allowing the use of the MMC principle would make the part unnecessarily expensive.

MMC should be used for acceptance but not process control. the hole position and 'max material' are independent (mostly)

to make live easier, I am going for X-MR chart:

X=(x^2+y^2)^1/2
MR=((dx)^2+(dy)^2)^1/2

dx, dy stand for MR-x and MR-y

But is it to much the calculation for the operators? any suggestions?

Thank you for the help!

the Moving Range should simply be teh moving range of the vectors you calculate in equation 1. without seeing the results and having the raw data we can't propose anything different.

additionally - don't confuse product acceptance with process control. MMC goes to acceptance and what you proposed goes to process control - tehy are two separate things

MMC should be used for acceptance but not process control. the hole position and 'max material' are independent (mostly)If the holes are made close to the max permissible diameter, you will be able to accept a greater deviation from the intended position. The mating part would still fit, and this is why I suggested a function gauge. Not using the extra tolerance provided by the MMC could mean that perfectly usable parts are scrapped, even though they are within tolerance.

/Claes

If the holes are made close to the max permissible diameter, you will be able to accept a greater deviation from the intended position. The mating part would still fit, and this is why I suggested a function gauge. Not using the extra tolerance provided by the MMC could mean that perfectly usable parts are scrapped, even though they are within tolerance.

/Claes

yes - which is why MMC is used: for acceptance of parts that are still functional. but the OP was asking about control charts not acceptance sampling. parts are not accepted or rejected with control charts - the process is. the factors that control the hole location are not the same factors that control sizes (there is a small possible overlap - but in the main they are different).

because of this and because process control is not product acceptance I recommend not including MMC directly in a control chart for location. The OP may opt to also chart size. The OP could plot the vector (sqrt of x^2 + y^2) or could plot x and y separately: the nature of hte process and the actual results are required for me to recommend which approach is best.

I agree that functional gages can and should be used for product acceptance to take advantage of any MMC allowance.

of course I don't believe we've discussed the affect of the actual callout and the use of a circle or a square for the tolerance zone...but that also goes to acceptance and not control.

but the OP was asking about control charts not acceptance sampling. So he did...

/Claes

Generally, location callouts utilizing the MMC callout on a print are screaming for hard gage go/no go checks - and are used to simplify gaging, not to complicate it with geometric and trigonometric misery. If one was to follow that philosophy, then an attribute chart would work for that requirement. After all, to really control the process, you have to control a combination of things: such as location, hole diameter (if round), roundness, etc. As you inferred, there is a cost to that - a cost no one signed up for. Besides, before applying any control chart - ever - one should do the capability study. What is the result of that? Do the absolute locations vary? Do the hole sizes vary? Does the roundness vary? Those may be what need variable control charts to control the actual callout. The attribute chart would be a report card chart for the sake of the customer's curiosity.

Thank goodness you are doing stamping. Can you imagine adding cylindricity to the puzzle?

One last question: the control chart on the floor should be providing information to the operators. What controls and adjustments can they make to the process? Your chart should lead them to how much adjustment they can make, to what feature and when. Does it?

Thank you all for the valuable inputs!

I will not be confused with MMC and process capability which I am taliking about targeting and variation.

After talking with people on the floor, I will give a try for charting the Individual-Moving Range based on the x and y reading. we will use table to calculate each step. we will see how it goes.

I am very much tempted to use scatter chart. I can calculate the mean and sigma based on the capability studies, use them to construct 1, 2 and 3 sigma circles, use the zone analysis and quadrant analysis to identify out-of-control conditions. The problem might be I would lose the run information (or not)? or any other flaws you may think of ? I will give a try also when I collect enough data.

Any comments? Jim, Steven and Steve, would you care to jump in and develop the idea?

:thanx:

Thank you all for the valuable inputs!

I will not be confused with MMC and process capability which I am taliking about targeting and variation.

After talking with people on the floor, I will give a try for charting the Individual-Moving Range based on the x and y reading. we will use table to calculate each step. we will see how it goes.

I am very much tempted to use scatter chart. I can calculate the mean and sigma based on the capability studies, use them to construct 1, 2 and 3 sigma circles, use the zone analysis and quadrant analysis to identify out-of-control conditions. The problem might be I would lose the run information (or not)? or any other flaws you may think of ? I will give a try also when I collect enough data.

Any comments? Jim, Steven and Steve, would you care to jump in and develop the idea?

:thanx:

I think Claes and Bob are right; you need an attributes gage. Seems like you're going to an awful lot of trouble for a stamping. The likelihood of the relationships between the holes changing is (or should be) practically nil with a reasonably maintained die.

I am still curious what adjustments to the process the operators have. Then, how the proposed charting methodology will guide them on how much to adjust and when. Otherwise, the charting is only academic.

I am still curious what adjustments to the process the operators have. Then, how the proposed charting methodology will guide them on how much to adjust and when. Otherwise, the charting is only academic.

well, there's always calling in maintenance, supervisor or engineering, bu one would think that the change with a stamping would be either slow (so frequency of plotting could be as little as first piece last piece) or catastrophic (in which case a control chart probably won't help you avoid making bad parts). If it's smaller house they may not have automated SPC to notify offline support and so hand charts by the operator are the only option.

we really would need to see the capability study data - and understand how the study was performed - to truly provide good advice beyond simply answering the OP's question: how to construct a control chart. (he didn't ask if it would be useful, nor did he state if he already does use a pass/fail gaging system for lot acceptance - unless I missed it somewhere)

well, there's always calling in maintenance, supervisor or engineering...

True - but one would think a simpler charting methodology would be all that is necessary to determine if you need to flag help rather than to direct adjustment - especially if their only option is to change out the die.

Also, you are correct- it is difficult to suggest correct charting methodology without an solid understanding of what the expected variation is. Is it tool wear? If so, you would think size change would make more of an effect than location change. Location change would be likely tool verification (if the hole punches are all on one tool) or program verification (if it is one tool moving to make several holes). Hard to guess with the info so far.

I am all for elegant charting, if it gives you elegant information.

True - but one would think a simpler charting methodology would be all that is necessary to determine if you need to flag help rather than to direct adjustment - especially if their only option is to change out the die.

Also, you are correct- it is difficult to suggest correct charting methodology without an solid understanding of what the expected variation is. Is it tool wear? If so, you would think size change would make more of an effect than location change. Location change would be likely tool verification (if the hole punches are all on one tool) or program verification (if it is one tool moving to make several holes). Hard to guess with the info so far.

I am all for elegant charting, if it gives you elegant information.


Good Morning,

Excellent discussion. It's a topic (GD&T) near and dear to my heart...but without more information (earlier I asked for more specifics), it's going to be difficult to come up with a more definite answer.

Great points were made regarding process control and product control. Yes, if you are only concerend about functionality/acceptibility, a functional (attribute) gage may be the answer depending on the quantity of parts.

If the concern is controlling the process, in terms of monitoring the hole locations, then it becomes more challenging. I believe that this was the intent of the initial post. There are a few articles written about process capability on hole locations. A Google search has turned up a few. I am hestitant to attach them to this post because of copyright concerns.

I'm sure we'll continue the dialogue.

Stijloor.

If the concern is controlling the process, in terms of monitoring the hole locations, then it becomes more challenging. I believe that this was the intent of the initial post.

I agree, and the point I am emphasizing is if it the hole location that is the key significant variation (not just because location is called out on the print), then these techniques may be truly useful. But, the use of MMC makes me think it is not just location, but fit that is key. MMC is very meaningful for this callout! Hole size could very well be the key significant variation (that is, negatively affects the characteristic acceptance quicker than location), therefore more of a meaningful characteristic to track than location. That is why I keep trying to determine what decisions is the chart going to assist the operator (or as Bev states, maintenance, etc.) in making? Make sure the chart answers that question to be useful.

But, again, we just are not sure at this point. :cool: