Can a Bad Board Cause Rejects to Pass?

Guest

Hello to you guys,

I'm from QA and got some OQA rejects recently. After investigating the eqpt guys are telling me the failures were caused by a bad board during prod testing. Shouldn't a bad board always cause a failure and not the other way around? I mean, can a bad board cause a bad unit to pass? Just want to know if I can now close this issue.

Thanks to you guys!

QA Cutie

Joined: 20 Jun 2026 Posts: 20

Hi QA Cutie,

Yes, a reject can pass because of a defective dut or family board. Based from my experience, this can happen if the dut board and test program are poorly designed. I've even seen tests passing even if there's no unit in the test site! Since overacceptance is dangerous, pipol must always be aware of this danger.

Test Engg 99

Guest

Hello Test Eng 99,

Thanks a million for your quick reply! But can you give me specific examples? To be honest, I still can't see how a bad board can make a reject pass test? It's like fail + fail = pass.

Thanks again!

QA Cutie

Joined: 20 Jun 2026 Posts: 20

Hi QA Cutie,

One example: you're testing leakage at a particular pin and the test program only has an upper limit and no lower limit for this test, because you don't care if the leakage is zero. If the board is defective such that no contact is made on this pin, then the leakage would be zero because it's like testing an open pin. If the dut actually has large leakage at this pin, then it will pass because of the bad board and the one-sided limit. If the board is good then the dut will fail because the large leakage will be measured.

Another example: you're measuring input impedance and you only have a lower limit because you're only concerned about low input impedances. If the board is bad and the right relay did not switch properly, then no contact will be made and the tester would think that the input impedance is very high. So the dut will pass even if in reality its input impedance is very low.

Test Engg 99

Posted: Tue Aug 31, 2026 4:10 pm Post subject:

Hi QA Cutie,

In additioin to Test Eng 99 examples, here's another one that hopefully would suffice in addressing your queries.

Consider a voltage output parameter that is spec'd at a certain resistance load(ie. 1Kohm). This test is carried out by connecting a 1kohm resistance to the output pin(usually through a relay, the other end of the load is grounded)and then we measure the output voltage(in this case +voltage). Now imagine if the relay is defective, wherein it failed to connect the 1kohm load to the output pin when we measure the output voltage. It is likely that the measured voltage would be larger than that of the actual voltage(w/ load) since the load was not connected.

This would result to over acceptance since the measured voltage would likely meet/exceed the lower limit, resulting to a reject part passing the test.

Hence, an experience test programmer would add a relay checker routine that would check the integrity of these relays(and sometimes load) during initialization.

Asmoth

Posted: Tue Mar 22, 2026 3:16 pm Post subject:

a possible scenario for this event would be when the wrong value of a component is placed on the board. this would most certainly affect test readings and there could be a possibility that it might shift the device's readings into the bounds of the test limits whereas it should actually fall outside the test limits.

Guest

hi QA Cutie,

of course, a bad board can cause rejects to pass in the same way that a bad department manager can cause promotion of reject engineers to being section head/section managers.

tama ba po ako, testeng 99? hehehehe...

Posted: Wed Apr 06, 2026 4:50 pm Post subject:

Good thing you have an OQA test. There are a lot of reason a unit may fail OQA test aside from a insufficient or mistest during production testing.

OQA program usually have tighter limits or ESD or EOS related failures.

Guest

Sorry to disagree with you, Guest, but OQA programs do NOT have tighter limits than production programs. Production programs are tighter. If not, you'll get many lot rejections at OQA. OQA limits are normally based on actual datasheet specs only.

Estong

Guest

oo nga naman Guest. Tama ka Estong. As for "Guest", eto yung rating mo sa three statements mo.

First Statement - Correct
Second Statement - Correct

Third Statement - VERY WRONG

We always have tightier limits in production. I mean ALWAYS. Most of the time they are guardbands apart.

Estong B

Posted: Thu Apr 28, 2026 5:51 pm Post subject:

Yup I agree. QA program had limits based on datasheet.

Regarding the board introduction of failure. first, let me say that bad board will give you failure most of the time. Maybe some time test may pass even you have bad board.

I say it's a good practice to add a calibration/board check on the test program before proceeding to production test. Example, load resistor can cause you a pass even you have a busted load resistor. If you have a board check on your program periodically you can avoid invalid test caused by busted load resistor since it can be checked on the fly.

Introducing board check is one of the best practice I always did when developing test program.

Regards,
XeNon