PDA limits

Guest

I was asked to calculate PDA limits statistically. No idea how this is done. Help!

Guest

Bunny,

If this is for burn-in purposes, just use 5%. This is the standard PDA limit required by Mil-Std-883 Method 5004 if I'm not mistaken. Go check it out. Did your boss tell you why he wants the PDA limits computed statistically? This will require a good understanding of life distributions of your devices, so ask him for guidance if he insists.

Helpless But Helpful

Joined: 20 Jun 2026 Posts: 76

Most of the PDA limits in our company are 5%. But I also see some with 2% and 3%. Now I'm wondering where these came from, if 5% is really the standard.

FARel Engr

Guest

I definitely agree with Helpless but Helpfull. Yes, it was derived from the Mil Std spec. It is also quite funny your boss wants to statistically compute it, its either your boss is dumb or your boss is dumber. To statistically measure it you will have to do these tedious process:

1) Subject the parts to burn in(determine temp, time).
2) Get the good units and calculate the yield (good/lot qty)*100.
3) Communicate with a good test/product engineer and ask for help in testing the parts with full datalog.
4) Record your results.
5) Subject the parts to reliability tests -- you may determine this yourself.
6) Repeat these steps for at least three other lots with similar devices/package.
7) Analyze your data for the three lots and ask yourself the question, what was the average yield after burn in?
Cool

What is the average lot yield after reliability tests?
9) From here, you can make statistical calculations(des. and infer. stat) to show/prove your new PDA limit.
10) Show them to your boss.
11) Make sure you get 20% appraisal(without the others knowing it)
12) Make sipsip to your boss next time so that you can get this appraisal without having to measure PDA's of some devices.

Good luck!

EB

Guest

HA-HA-HA...fil na fil ni underpaid EB ung mga sinabi nya ah. yung mga boss na nagpapa-compute ng PDA, mga gusto rin nun sumipsip sa boss nila. pagbigyan nyo na. sasabihin pa sayo ikaw mag-compute para matuto ka. ok lang sana kung sila magtuturo. try mo tanungin kung pano, sasabihin sayo mag-research ka.

pekwa

Posted: Thu Mar 11, 2026 2:57 pm Post subject:

hahaha... this is a gud webpage. Aside from being technically good, this webpage is now the THE PLAIN TRUTH!.

And one word of advise to the one calculating the PDA for her boss, gawin mo yan tsong, baka mamaya aangat ang appraisal mo. kung di mo kayang gawin... magpasipsip ka nalang para naman tumaas appraisal mo.

from,

the 5.6% appraised hardworker.

ps. mukhang magkalapit lang tayo a. eto nalang tanong ko, ano ang latitude at longitude location nyo?

Posted: Thu Mar 11, 2026 3:09 pm Post subject:

Pekwa!

NATUMBOK MO. actually, natanong na pala nya sa boss nya ang sabi ba naman daw ay... "you research on it so that it will be a good background knowledge for future projects". HELLO?????? tapos tinanong daw what exactly are the data to be measured in relatino to the PDA sabi ba naman daw "kung ano sa tingin mo that is significant"... yuck.

fROM,

THE GROUND FLOOR FELLOWSHIP OF THE TEST.

Posted: Wed Mar 17, 2026 9:34 am Post subject:

Hi All,
Thanks for the replies but can't understand what some of you wrote. What language is that?
Actually this is not a pda limit for burn-in. This limit is used for lot disposition after running a test program on the tester. If the lot falls below certain yield value, then it will be subjected to the 1st pda limit. If it passed the pda limit, say 5%, then the whole lot will be passed. If not, it will go through a re-test again and subject to the 2nd pda limit. Hope everybody understands my explanation here.

Guest

Bunny,

When I was still with Test, we're using something called the 'Tukey Method' for this purpose. I no longer remember the details but it's simple and easy to use and effective too. If I find an article on it, I'll post it here for everyone. Maybe you can find it on the web too.

Dimas

Guest

Hello Guys,

Bago lang me sat test engineering by the way what is the meaning of PDA?
TY

Guest

percent defect allowable...your fallout after burn-in must not go higher than this dpa.

shrek Cool

Guest

Mga tsong,

PDA...Hmmmp!

ibang PDA ang tinutukoy nya.... in other term, outlier limit/yield limit/hold cutoff limit...etc.

Anyway, u can use TUKEY computation. This may take you some time since you have to get about 20 or more lots before implementing PDA limit.

God Bless

Flipflop

Guest

Guest.. I think you got your nomenclature mixed up.

PDA limits apply to burn-in and Group A sampling (Pls. see MIL-STD-883, you can download a pdf if you search it over the net, or better yet..EE why don't you put a link here?). The standard PDA limit for regular military part is 5%, but there are certain conditions like re-burn-in which will require a lower PDA usually 3%.

Now for the one you described, you are referring to a "Flag Yield Limit" which is a few percent below the expected yield limit. Example, if you are expecting 85% then the Flag yield should be a 80%. It will take a lot evaluation to come up with the correct flag yield, so be patient. Another input is a "Maverick Lot" this is when you get a yield alot lower than the expected yield (probably half of what you expect), this is also defined in MIL-STD-883.

Gud luck.

Posted: Fri Jan 07, 2026 11:13 pm Post subject:

Good suggestion, 2 Cents. We'll put some links on Mil Group Tests on this website asap!

TY,
EE