LM22679: Oscillation is unstable

user1881876

Part Number: LM22679

Dears, We have been using LM22679TJE-ADJ in our products for 9 years.
Until now, we had cleared the shipping inspection without any problems, but from this lot, the ripple has increased.
This is the first time I have been manufacturing for 9 years.

The specifications of the power supply are as follows.
Input: 12V Output 5V / 3A
Ripple increases when the load is about 2.7A.
I checked by changing the overcurrent threshold, but the load current value at which the ripple increases is the same.
Therefore, it seems that the oscillation is unstable, not the malfunction of the overcurrent.
I'm using it with an RFBB value of about 2kΩ. If this is the cause, can it be a problem now that 9 years have passed?

Please tell me if you want to try it.

over 2 years ago

0 Daniel Li14 over 2 years ago

TI__Mastermind 27040 points

can you do a comparison with good and bad part of LM22679? Pls take the waveform of SW in same input and output condition .

Thanks

0 user1881876 over 2 years ago in reply to Daniel Li14

Prodigy 190 points

Thank you for your reply.
I will send you the waveform.
As a symptom, the load current appears at about 2.7A.
It becomes noticeable when the load current decreases from the heavy load current (3.0A) to 2.7A.
Installing the previous lot IC on the board where the problem occurred will fix the problem.

Please let us know if you find any cause or countermeasure.

LM22679_Waveform observation.pdf

Thanks.

0 Daniel Li14 over 2 years ago in reply to user1881876

TI__Mastermind 27040 points

do you have a way to test 2.7A loop, if anything difference from new and old part which will cause this issue, it may be the internal compensation . But as you mentioned, it only have issue when 3A down to 2.7A, can you add a small 10pF capacitor at FB node and see what will happen. I just want to see whether it is more sensitive to FB noise?

0 user1881876 over 2 years ago in reply to Daniel Li14

Prodigy 190 points

Thank you for your reply

The loop test cannot be done right now because there is no device.
I added 10pF to the FB node.
It has become a good trend.
Waveform fluctuations occur for a moment, but they subside after a while.
I will continue to experiment with different capacities.

Thanks

0 Daniel Li14 over 2 years ago in reply to user1881876

TI__Mastermind 27040 points

Don't put too higher value, it looks FB is sensitive , but hard to explain why. any passive component change, like inductor or capacitor?

Thanks

0 user1881876 over 2 years ago in reply to Daniel Li14

Prodigy 190 points

Hello

There are no changes to the passive components.
I mounted it on an old board and confirmed the operation.
Only the IC of the new lot will be reproduced.

Thank you

0 Daniel Li14 over 2 years ago in reply to user1881876

TI__Mastermind 27040 points

OK, then I think FB node may be more sensitivity

0 user1881876 over 2 years ago in reply to Daniel Li14

Prodigy 190 points

Hello

Do I need to adjust it for something with good sensitivity?

I will summarize this matter.
-This board itself has a small phase margin.
-Therefore, it happened due to IC variation (FB sensitivity).

Is there a way to prove that the old Lot IC has a small phase margin?
There is no problem even if the load sudden change test is performed with the IC of the old Lot.
(I want to know a simple method other than using FRA)

Thank you