Need Help with FPY Data in Assembly Process

jsh298

Registered
Hey, everybody. I am the QE for a company and need help regarding an AutoTester. So we have a product we assemble in a U shaped cell. Once it is all assembled, they run it through HiPot in an AutoTester. If it passes, they then put a decal with the serial number on the unit. If it doesn't pass, then they either run it again until it passes, or try to determine why it failed, then rerun it. Once it passes, they then stick the serial number decal on it. The AutoTester day only shows a count of units passed and units failed. So if they have one unit that failed three times, then passed, it would show as 3 units failing and 1 unit passing. Management seems to be okay with this data, but I am not. I am trying to determine how to get an accurate count of fails and passes per unit so that its 1:1 not 4:1 as referenced above. Can anyone please help me so that I can fix this FPY issue? Thanks in advance!
 

optomist1

A Sea of Statistics
Super Moderator
I beleive you are refering to what I am familiar with...FTC, or first time capability?
 

jsh298

Registered
It is FPY that I am looking to improve. Management wants to see better FPY numbers, but to me, FPY isn't being recorded properly to get an accurate count of good vs bad. Showing that we have 3 bad and 1 good when they failed one part three times and it passed the fourth isn't reporting correctly in my opinion. I would love some feedback on thoughts as to how to improve upon this. If anyone has any knowledge of HiPot testing with an AutoTester, that would be very beneficial too since this is where they are pulling all of the data.
 

blackholequasar

The Cheerful Diabetic
We use a Cirrus tester for our HiPot. There is some good data that you can collect from it via an Excel export. We have an internal calculation that we run to show a more accurate FPY. For us, the issue was opposite - we would have a cable with 40 connections and 10 of those would have faults so it showed one cable failed 10 times! Using DPMO helped us quantify it better. Does your HiPot have any sort of reporting applications that you can use? If so, I may be able to share the calculation excel that we use.
 

jsh298

Registered
I will have to check. I am new here and I get the FPY report from a daily report sent out. I need to dive deeper into that rabbit hole on Monday. Once I do, I will post back on here. I would appreciate any excel calculations. Thanks!
 

Bev D

Heretical Statistician
Leader
Super Moderator
Welcome jsh298! It’s always nice to have accurate metrics as it typically can stop the wasteful arguments and questions regarding the accuracy of the report. But I think it’s important to remember that it’s not necessary. You can and should focus your efforts on solving the failures. The entire point of the metric is to help the organization identify and prioritize when there are an overwhelming number of failure types. But when it’s internal failures. You have immediate access to the failed units and the failure type and you can begin root cause analysis. Remember that FPY i sa one dimensional metric - multiple repeat failures are just as wasteful maybe even more so. Go fix the causes and your metric will take care of itself.
 

jsh298

Registered
I haven't been able to go look at the autotester this morning, but I have a side question regarding FPY. Thanks Bev D for the info! I am focusing on the failures and diving into those, but management is solely focused on FPY numbers (what boss says goes....) So, I realized this morning that they are using =(#passed autotester / #total tests) to get FPY. With the autotester just reporting # good and # bad so you cant differentiate how many times one pump got ran through the autotester before it passed, I think it should be =(1-(#failed / #good) to get FPY. I used my formula at my last job and it was always correct. Which one is right?
 
Top Bottom