APQP, Gage R&R, MSA HOW? Maintenance?

[Icy Mountain]

Preface: We make electronic hardware that converts 12 volts direct current (car battery) @ 100 amps into 120 volts alternating current (wall socket) @ 10 amps. I am attempting to finalize an APQP and a PPAP package for our customer. The control plan, gage R&R and the MSA for this have me stumped since this is not a caliper checked camshaft or casting, it's electronics! We check 3 areas of this product:
1) During PFMEA we determined that 80% of our reliability issues are related to loose hardware, specifically the nuts and bolts that hold down heavy transformers and large busbar connections that carry high amperage current. We have included torque specifications in our assembly instructions for each critical nut and bolt. These specs are ~25% greater than "tight" and ~25% of the torque required to break the hardware. So, while we want to know we've "tightened" the nut, it is not critical torque as in an engine head application. We use a your basic $40 torque wrench, a small window shows the torque value as you twist the handle to set it.
2) The product has a 4 bolt mounting pattern. We plan to use a go-nogo platform to make sure the mounting holes line up. The holes are much wider than the width of the recommended mounting bolt and are U shaped. Once again, the acceptable variation is huge.
3) Finally, we functionally test all of the operational features of the product on every unit. This is required 100% because the units are "calibrated" at this point. This term is used loosely since we are setting a potentiometer to adjust the output to 120VAC at a specific load. The output of inverters varies in relation to the amount of load (current demand) applied. The output spec is very wide for this reason and the kind of equipment that gets plugged in will operate from 110-130 VAC so the acceptable variation is huge here as well. We measure the inputs, outputs, and loads with HP and Fluke equipment that is +/-0.05% accurate.
Questions:
1) Units are tested 100% to specifications in the test procedure that are based on the customer's requirements and all of the functions are checked/set. Is that the control plan?
2) What exactly do I need to show in gage R&R? A measurement systems analysis seems to be overkill. All of this equipment is overkill for what we are measuring. Unless you are blind, it is impossible to misread the equipment enough to make any significant impact. Is that a MSA?
3) We have three identical HP/Fluke test sets, calibrated annually. The wrenches are off-the-shelf. Unless the whole plant disappears, the line can't shut down for equipment failure. So what do I say about maintenance?

 

[Sam]

Preface: We make electronic hardware that converts 12 volts direct current (car battery) @ 100 amps into 120 volts alternating current (wall socket) @ 10 amps. I am attempting to finalize an APQP and a PPAP package for our customer. The control plan, gage R&R and the MSA for this have me stumped since this is not a caliper checked camshaft or casting, it's electronics! We check 3 areas of this product:
1) During PFMEA we determined that 80% of our reliability issues are related to loose hardware, specifically the nuts and bolts that hold down heavy transformers and large busbar connections that carry high amperage current. We have included torque specifications in our assembly instructions for each critical nut and bolt. These specs are ~25% greater than "tight" and ~25% of the torque required to break the hardware. So, while we want to know we've "tightened" the nut, it is not critical torque as in an engine head application. We use a your basic $40 torque wrench, a small window shows the torque value as you twist the handle to set it.
2) The product has a 4 bolt mounting pattern. We plan to use a go-nogo platform to make sure the mounting holes line up. The holes are much wider than the width of the recommended mounting bolt and are U shaped. Once again, the acceptable variation is huge.
3) Finally, we functionally test all of the operational features of the product on every unit. This is required 100% because the units are "calibrated" at this point. This term is used loosely since we are setting a potentiometer to adjust the output to 120VAC at a specific load. The output of inverters varies in relation to the amount of load (current demand) applied. The output spec is very wide for this reason and the kind of equipment that gets plugged in will operate from 110-130 VAC so the acceptable variation is huge here as well. We measure the inputs, outputs, and loads with HP and Fluke equipment that is +/-0.05% accurate.
Questions:
1) Units are tested 100% to specifications in the test procedure that are based on the customer's requirements and all of the functions are checked/set. Is that the control plan?
2) What exactly do I need to show in gage R&R? A measurement systems analysis seems to be overkill. All of this equipment is overkill for what we are measuring. Unless you are blind, it is impossible to misread the equipment enough to make any significant impact. Is that a MSA?
3) We have three identical HP/Fluke test sets, calibrated annually. The wrenches are off-the-shelf. Unless the whole plant disappears, the line can't shut down for equipment failure. So what do I say about maintenance?

#1 - It is not a control plan until it is written down and documented. The CP starts with receiving and ends with shipping. You can have one big CP or one for each process.

#2 - An MSA not only measures operator variation but also gage variation. I think you will be required to do an MSA. I've never been able to get out of doing one, even for a tape measure.

#3 - The torque wrenches require calibration. Your line may not be shut down, but an auditor could require you to notify your customer that you have shipped product tested/assembled with non-calibrated equipment.

 

[Icy Mountain]

Good Feedback!

Thanks Sam, I was beginning to think no one saw this.
#1) We have documented processes that are standard for getting parts from the back door to the actual assembly line and from the end of the line to the shipping truck. In conjunction with the specific assembly and test process instructions, this is the whole of the control plan.(?)
#2) I suppose I'll have to get the MSA manual and train myself
#3) Just like the test equipment, the wrenches will get annual cal. Again, covered by SOP. So I shouldn't have to say anything special. (?)
This is our first shot at a customer specific APQP

 

[p_tww]

Thanks Sam, I was beginning to think no one saw this.
#1) We have documented processes that are standard for getting parts from the back door to the actual assembly line and from the end of the line to the shipping truck. In conjunction with the specific assembly and test process instructions, this is the whole of the control plan.(?)
#2) I suppose I'll have to get the MSA manual and train myself
#3) Just like the test equipment, the wrenches will get annual cal. Again, covered by SOP. So I shouldn't have to say anything special. (?)
This is our first shot at a customer specific APQP

Hi,

#1) did you check with the annex A of Standard of ISO/TS16949? you may find the content must be shown in control plan.

I am not sure whether you did have control plan, but it seemed you had working instruction only, you know control plan was a heading document which was linked with PFMEA. you work instruction was just for deploying of control plan.
#2) You got it, you need understand the purpose of MSA.
#3) I assume the maintenance you mean is against measuring devices. although you had advanced HP testing devices, but you still need some daily maintenance for electronic testing device as common sense, cleaning, dust proof, voltage shocking etc.
hope help

 

[Icy Mountain]

thanks p_tww

#1) I re-read Annex A. Thanks for the reminder. Very, very good advice. I have a document to write.
#2) MSA same thing.
#3) We do have standard procedures for care and feeding of our equipment as part of our section 7.6 documentation. I'll reference those.

 

[Icy Mountain]

I need to go fishin' w/ energy!

This is why MSA has me staring at the walls:
I want to measure 120VAC output. The specification limit is 110-130VAC. The control limit is 114-126 VAC. So I create my "reference", it is an HP power supply, calibrated and certified traceable to NIST. I set it to 120.00VAC/60.000Hz. I measure with my Fluke DMM. Fluke reads 120.01VAC/60.000Hz. It reads that at turn-on, an hour after, morning, noon, night, full moon, new moon, August, February, this year, last year, ASEE operator, BSEE operator, orangutan operator, 1 reading, 10^6 readings, DOESN'T EVER CHANGE!!!

Isn't that stable? OK, the bias is 0.01 out of 120, so what? It does the same at 12VAC/6HZ and at 1200VAC/600HZ, it was designed to (It's linear)!!! The digital readout has the same numbers no matter who (or what) is looking at it (Gage R&R)!! There is no range, there is no run, there is no variance. IT'S ALWAYS THE SAME VALUE!!!!!!! <- Is that a measurement system analysis?

 

[Atul Khandekar]

I don't think you should do an R&R study (or only R study if there is no operator influence) with a "standard". You need different parts spanning the spec limit or process variation. If measurement is not affected by operator, you could just do a P/T ratio.

 

[Icy Mountain]

If measurement is not affected by operator, you could just do a P/T ratio.

The readings I used as an example above are from a Digital Multi Meter (DMM, sorry for the acronym) . There is NO operator variation or device to device variation on the reading. Please, explain P/T ratio. Using multiple parts (we only produce 100/month on a high volume line) won't net me much. A (very)wide range of variation would be 6 units at 120.00, 2 at 120.01, 2 at 119.99 for a standard daily run of 10 units.

 

[Atul Khandekar]

It reads that at turn-on, an hour after, morning, noon, night, full moon, new moon, August, February, this year, last year, ASEE operator, BSEE operator, orangutan operator, 1 reading, 10^6 readings, DOESN'T EVER CHANGE!!!

Just trying to understand...Would that really happen with the 'unit' that you manufacture or only with the reference power supply?

 

[Icy Mountain]

Good question!

Atul,
The unit is designed to (and will) run an extremely stable voltage, given stable conditions (i.e, adequate and stable DC voltage and current input). For example, at 120VAC with a 5 amp load (600 Watts) it would measure 120.00VAC output on the DMM (different operator, different DMM, different day, same result). It will remain there (or within 0.01 VAC) unless some other external condition changes (like reducing the DC input voltage or current, raising the ambient temp 25 degrees C, etc.). Assuming all other conditions the same, increasing the load to 8.33 amps (1000 Watts) might drop the voltage to 119.87 VAC (this Output Voltage to load current curve is a known factor, part of the design testing). The reading will then remain steady (again +/- 0.01) just about forever.

Short answer, same Measurement System results from either our unit or the calibrated reference power supply.