PFMEA - CNC machining

JuliaDee

Registered
Hi all,
I am in need of some help to get started and nudged in the right direction.
I work at a production plant that in broad terms consists of HPDC, machining and assembly.

When it comes to the FMEAs, historically we have been listing EVERY dimension from the part drawing as a separate quality criterion.
For a part with 12 sheets of dimensions, every dimension generated a row in the FMEA and a SOD assessment. You can only imagine how “easy” it has been to engage a group of busy people for hours and hours of FMEA work!

But now the decision has been made and we are going to focus on our processes with process FMEA (AIAG&VDA) and I am going to be the coordinator.
Me and a couple of co-workers are registered for a training course, but not until the end of August due to summer vacations and so on.
Meanwhile, I have read the handbook and discussions of how to approach our machining process have emerged.

For example, one structure/process flow step is a Robot Cell “RC123”:
This is a robot handled production cell with an input line, CNC machining, washing, tightness testing, assembly of thread inserts and an output line for manual inspection and packaging.

My question is more specifically regarding the CNC machining operation.
The process operation/function is machining a cast part to specification. (The machining is quite complex and there are a lot of cutting tools involved).
After the fixture step, should/could I divide the operation by the type of cutting tool?
I.e., face milling, profile milling, drilling, thread tap etc.

For example:
Process: Threaded holes / Function: Drill and thread holes to the right specification
Failure modes: Wrong dimension, bore depth, thread depth, position, thread (pitch), no thread.
And then the associated failure causes due to the elements Man, Machine, Material etc.

Or is there a better way of managing the CNC operation?

And how should special characteristics specified on the drawing be handled in the PFMEA? For example, holes with a position tolerance Ø0.03 to ref ABC.

Any advice would be very much appreciated.

/Julia
 

AndrewK

Involved In Discussions
Hello!
If you have different specifications and requierements for face milling, profile milling, drilling, thread tap etc. then i would split them. but basically you should start from process flow- which steps there mentioned as separate processes(or operations). Each step has own requierements,
basic logic in handbook is - failure is opposite of requierement. So the more detailed function and specs analysis you will do is better.
About special characteristics specified on the drawing - is product requierements. You must analyze process variation impact on them and consider which relevant process parameters needed to be controlled more precisely than other to achieve product requierement in the end.
 

Miner

Forum Moderator
Leader
Admin
I would recommend that you split them out. Each of the sub-operation will have its own unique failure modes and risks. You are less likely to miss some of these when you break them out.
 

Enghabashy

Quite Involved in Discussions
the competence is part of FMEA , MSA is part of reference , the relevant control plan ; relevant error proof , CSR & internal specific requirements which relevant of customer & standard regulation / mandatory specs.---- should be referenced with FMEA series , the Design FMEA should be part of process FMEA; design member could be presented in PFMEA team also as well
 

MOester

Starting to get Involved
I am very well versed in your process and have reviewed at least 500 PPAPs that are exactly that: HPDC+Machining+Assembly in some form.
Some folks put EVERY dimension in the FMEA & Control Plan, some do not.

My preference is to NOT put every dimension in as a failure mode. Here are the reasons:
1) If you put every dimension in, your are burning too much time making sure every dimension is in there. And not enough time focusing on the actual failure modes.
2) You will be constantly redoing your PPAPs every time any dimension changes. Even if the dimension doesn't change, but you find statistically that at some point after SOP you change your guard bands, well, your control plan changes and you have to PPAP again.
3) You will hamstring yourself with executing proper read across / lessons learned. What I mean is if you learn through a corrective action that you had a PFMEA problem, you will have a mountain of work to change paperwork where a more generic PFMEA is easier to change. I'll honestly take a shop with a robust lessons learned program over a shop with fully detailed PFMEAs any day of the week as a customer.

I could go on, but those should be enough.

Take your machining center, any machining center, and the failure modes are generally:
1) Problems caused by clamping (deforming the part)
2) Problems caused by tool wear
3) Problems caused by tool breakage
4) Problems caused by misloading
5) Problems caused by debris on the fixture
6) Temperature problems (coolant temperature changing, changing the temp of the part)
7) Uncontrolled changes to program
8) Casting problem: bent / missing clamping points or datums
9) Casting problem: heat check / flash on the datums cocking the part
10) Casting problem: missing cast material resulting in non-cleanup
11) Improper tool loading
12) Improper setup

These are the failure modes. Not "this ID is out of spec" then "that ID is out of spec." If you sent me a part with "this ID" out of spec and your root cause was "I made the ID out of spec," I would reject your 8D. If your root cause was "The chiller went down for a shift and the coolant got to warm." THAT's a root cause.

Come up with your own list of how the machining operation fails, not the dimension, and PFMEA how you will check for and control THIS.
Will you need a job specific dimensional sheet with dimensions, limits and gaurd bands? Yes. Your PFMEA should refer to this being a thing that exists, but that's your process specific sheet which may be changed as you run your process. This is in order to protect MY print. Folks tend to forget that you CAN adjust your process to keep the dimensions centered and in control. Don't tie your hands to do this by specifically putting all this in the PFMEA.

On to the Key Characteristics: THESE should be specifically called out in your PFMEA because I have told you they are critical to my unit functioning.

The smart strategy is:
1) Make a generic master PFMEA and Control Plan "block" for your HPDC, your machining and your assembly.
2) Now you make a process for my part. Let's say it takes 3 machining operations. You point to your casting block once, then your machining block 3 times. (After all, your list of machining failure modes should be universal.)
3) You then add in where appropriate my Key Characteristics at the points where they are made AND checked.

Now, you make an error and in your corrective action you need to change your master FMEA. If it's a worthwhile change, you change your MASTER block, and it cascades down to all your other blocks.

A strategy like this will let you focus on what's different from the norm in any new job, instead of burning daylight doing repetitive paperwork on the things that are standard.
 

Enghabashy

Quite Involved in Discussions
* The System & subsystem could be useful for organizing the processes Sequence;
**Criteria of SOD / RPN should be stated with considering CSR if applicable
PFMEA - CNC machining
 

JuliaDee

Registered
I am very well versed in your process and have reviewed at least 500 PPAPs that are exactly that: HPDC+Machining+Assembly in some form.
Some folks put EVERY dimension in the FMEA & Control Plan, some do not.

My preference is to NOT put every dimension in as a failure mode. Here are the reasons:
1) If you put every dimension in, your are burning too much time making sure every dimension is in there. And not enough time focusing on the actual failure modes.
2) You will be constantly redoing your PPAPs every time any dimension changes. Even if the dimension doesn't change, but you find statistically that at some point after SOP you change your guard bands, well, your control plan changes and you have to PPAP again.
3) You will hamstring yourself with executing proper read across / lessons learned. What I mean is if you learn through a corrective action that you had a PFMEA problem, you will have a mountain of work to change paperwork where a more generic PFMEA is easier to change. I'll honestly take a shop with a robust lessons learned program over a shop with fully detailed PFMEAs any day of the week as a customer.

I could go on, but those should be enough.
Thank you! I could not agree more :)
My absolute goal is to create generic PFMEAs for our main processes.

Take your machining center, any machining center, and the failure modes are generally:
1) Problems caused by clamping (deforming the part)
2) Problems caused by tool wear
3) Problems caused by tool breakage
4) Problems caused by misloading
5) Problems caused by debris on the fixture
6) Temperature problems (coolant temperature changing, changing the temp of the part)
7) Uncontrolled changes to program
8) Casting problem: bent / missing clamping points or datums
9) Casting problem: heat check / flash on the datums cocking the part
10) Casting problem: missing cast material resulting in non-cleanup
11) Improper tool loading
12) Improper setup

These are the failure modes. Not "this ID is out of spec" then "that ID is out of spec." If you sent me a part with "this ID" out of spec and your root cause was "I made the ID out of spec," I would reject your 8D. If your root cause was "The chiller went down for a shift and the coolant got to warm." THAT's a root cause.

Come up with your own list of how the machining operation fails, not the dimension, and PFMEA how you will check for and control THIS.
Will you need a job specific dimensional sheet with dimensions, limits and gaurd bands? Yes. Your PFMEA should refer to this being a thing that exists, but that's your process specific sheet which may be changed as you run your process. This is in order to protect MY print. Folks tend to forget that you CAN adjust your process to keep the dimensions centered and in control. Don't tie your hands to do this by specifically putting all this in the PFMEA.

I just wrote down a very quick draft in excel (not in any way complete) just to get your opinion if we are on the right course?
Or is it too simple?

PFMEA - CNC machining

And of course we will also discuss and find the preventive and detective process controls and a SOD assessment.
 

AndrewK

Involved In Discussions
Thank you! I could not agree more :)
My absolute goal is to create generic PFMEAs for our main processes.



I just wrote down a very quick draft in excel (not in any way complete) just to get your opinion if we are on the right course?
Or is it too simple?

View attachment 29612

And of course we will also discuss and find the preventive and detective process controls and a SOD assessment.
In my opinion looks good! do you create draft acc. AIAG4 or better to try in AIAG-VDA? In this case you need to breakdown your processes into elements (4M) - like process of loading fixture (opposite for product requierements in this step is a failures) - ivolved operator (Man), Fixture itself (Machine) and material/ For each of 4m you will have their own function and reqquierements - opposite of those will be a causes/
 

MOester

Starting to get Involved
It's a good start, but be careful you do not get repetitive.
In your machining section, you have failure modes "wrong dimension" and "wrong depth" for example. Is not the depth a dimension? I can see surface finish being different as that is very sensitive to speed/feed.
But from the snippet, I like it. It's addressing what the PROCESS can do wrong and how to detect/react to it.
 

JuliaDee

Registered
Thank you all for your response and input. It really helps!

In my opinion looks good! do you create draft acc. AIAG4 or better to try in AIAG-VDA? In this case you need to breakdown your processes into elements (4M) - like process of loading fixture (opposite for product requierements in this step is a failures) - ivolved operator (Man), Fixture itself (Machine) and material/ For each of 4m you will have their own function and reqquierements - opposite of those will be a causes/

We will be using the AIAG-VDA take on PFMEA.

The FMEA software we are working with is somewhat adjusted to the new format of AIAG-VDA, but not fully and I think this confuses me sometimes.
In the software, the “Potential Causes” are split up in the 4Ms.

As for the Failure Mode in my draft the Failure Causes could be;
Man: Part incorrectly placed by the operator
Machine: The fixture’s guides/support surfaces are damaged
Material: Defect cast references
Environment: Bad lighting

Or do you mean a different breakdown? English is not my native language, so there is always a risk of misunderstanding
 
Top Bottom