In-Process Production Test Stations in PFMEA (Process FMEA)

nitros1

Starting to get Involved
#1
Hi all greetings.

Working in one of my first PFMEAs I come across with tester stations. My question is how are these stations treated? They are part of the process but no operation is being made, or modifications are being made to the product they are just testing some characteristics of the process, using sensors for example, so the failure modes of these stations would just be not testing well? if these is the case then why in other non tester stations the malfunction of sensors is not considered failure modes? or for example if I want to the test the presence of a screw, should I include the failure mode * absence of screw* even though I considered this stations before in the station that indeed puts the screw on?

Secondly, what is the best way to use * visual inspection* as a detection method? An engineer told me that I shouldnt use visual inspection unless I had somebody inspecting the condition, what about if the operator of the next station in the paragraph above notices the lack of screw even though is not an inspector checking 100 per cent that condition.

Thank you for your help.
 

Miner

Forum Moderator
Staff member
Super Moderator
#2
Working in one of my first PFMEAs I come across with tester stations. My question is how are these stations treated? They are part of the process but no operation is being made, or modifications are being made to the product they are just testing some characteristics of the process, using sensors for example, so the failure modes of these stations would just be not testing well?
When including an inspection control in a PFMEA as a process step, the potential failure modes are:
  • Accepting a bad part
  • Rejecting a good part
A specific malfunction of a sensor (e.g., failed open) would then be a potential failure cause for these failure modes.

if these is the case then why in other non tester stations the malfunction of sensors is not considered failure modes? or for example if I want to the test the presence of a screw, should I include the failure mode * absence of screw* even though I considered this stations before in the station that indeed puts the screw on?
When sensors are a built in part of another operation considered in the PFMEA, they are considered part of the process controls for the specific failure mode/cause for which they were included to check. If there is a high failure rate for the sensors, I would downgrade the Detection rating accordingly.

Secondly, what is the best way to use * visual inspection* as a detection method? An engineer told me that I shouldnt use visual inspection unless I had somebody inspecting the condition, what about if the operator of the next station in the paragraph above notices the lack of screw even though is not an inspector checking 100 per cent that condition.
Visual inspection, both at the station itself, or by successive checks (next station), is a control, just not a very effective control. Reducing it from 100% to x% further reduces its effectiveness. Again, you can include it in the PFMEA as a control, but you should downgrade the Detection rating accordingly.
 

yodon

Forum Moderator
Staff member
Moderator
#3
Just to add on a bit to Miner's excellent post, you asked how test stations are treated. You do want to be sure that the test stations do identify the defects they're intended to identify. You can (should) do a validation of the test station to provide confidence that they perform as expected.
 

nitros1

Starting to get Involved
#4
When including an inspection control in a PFMEA as a process step, the potential failure modes are:
  • Accepting a bad part
  • Rejecting a good part
A specific malfunction of a sensor (e.g., failed open) would then be a potential failure cause for these failure modes.



When sensors are a built in part of another operation considered in the PFMEA, they are considered part of the process controls for the specific failure mode/cause for which they were included to check. If there is a high failure rate for the sensors, I would downgrade the Detection rating accordingly.
Thank you pretty much for your help, but if a test station has the risks/potential failure modes you mentioned, what do I put as a detection method since its already a detection station? A loop can go forever.
 

nitros1

Starting to get Involved
#5
Just to add on a bit to Miner's excellent post, you asked how test stations are treated. You do want to be sure that the test stations do identify the defects they're intended to identify. You can (should) do a validation of the test station to provide confidence that they perform as expected.
Thanks yodon.

For what I know R&Rs and CPks are a way, what else, would I need to validate?
 

Miner

Forum Moderator
Staff member
Super Moderator
#6
Thank you pretty much for your help, but if a test station has the risks/potential failure modes you mentioned, what do I put as a detection method since its already a detection station? A loop can go forever.
The most common control for an inspection process would be a calibration program. Others might include an R&R study, red rabbit testing, etc. You would stop at this point because these are no longer unique process steps.
 

yodon

Forum Moderator
Staff member
Moderator
#7
Thanks yodon.

For what I know R&Rs and CPks are a way, what else, would I need to validate?
R&R is close... you want to validate that the station detects bad product and passes good. There are often a lot of 'edge' cases that can be assessed in a test station.

Of course, if your test station has software, that needs to be validated as well, ensuring it's under proper control, etc.
 

nitros1

Starting to get Involved
#8
The most common control for an inspection process would be a calibration program. Others might include an R&R study, red rabbit testing, etc. You would stop at this point because these are no longer unique process steps.
ok, but what kind of controls would these be? Im guessing prevention controls, but what do I fill the detection controls column with? a detection for the detection...
 
#9
In the APIS-IQ Software, I have bridged the concepts and tools developed for DFMEA called "Mechatronics" (Error Detection & Error Response objects, and Operating Conditions) to be used in the PFMEA. The idea is that these inspection devices are to the Process, the same as a software diagnostic is to a product in the field. The same way that software detects a problem in your car, lights a lamp to warn the driver, and then puts the vehicle in a safe mode is directly the same as what the testing station is supposed to do on the manufacturing floor. As previous writer mentioned, we have Type 1 & Type II error to worry about.

DISCLAIMER: Mr. Johnson works for APIS-IQ software.
 
Last edited by a moderator:

Miner

Forum Moderator
Staff member
Super Moderator
#10
ok, but what kind of controls would these be? Im guessing prevention controls, but what do I fill the detection controls column with? a detection for the detection...
It depends. Some are both, like calibration. Done up front it is preventive, but it can also be detection when you find the gage out of calibration. Red Rabbit tests are definitely detection.
 

Top