INA126: 1/f noise

Hi Kanji,

this can also have to do with contamination on the PCB caused by the soldering. Have you carefully cleaned the PCB after soldering?

Kai

Hello Kanji,

That sure looks like classic popcorn noise, Some parts are screened for this (not 100% effective) but not sure this, or any of the INA parts, are.

Dear Kai-san

Thank you for your reply soon.

I believe our customer has log experience for assembly, but I will check to customer.

Do you have any article/document  in regard to contamination at soldering  and 1/f noise?

Or that is from your rich experience?

Best Regards

Dear Michael-san
Thank you for your comment. I think same thing as you say.

Hi Kanji,

I remember that we had a similar issue. Cleaning the PCB could fix it.

What speaks for popcorn noise, is the nearly constant amplitude of the noise peaks. So, it could actually be popcorn noise in your case...

Kai

Hello Kai and Kanji, 

One slight clarification, -in the many popcorn measurements I have seen or done, it is rarely the case the amplitudes are the same. Being noise, popcorn is random steps in time, duration, amplitude and polarity of the input offset voltage - normally, there are also cases where those steps manifest themselves in the offset current. there might be cases where it is the same polarity and amplitude, but don't count on it. 

Attached here is some example bench data on a different part - this is not the INA126. NOrmally the next step would be to return a suspect part for bench popcorn verification. That involves a specialized post amplifier with a lot of gain and active filtering built around the OPA627. 

Example popcorn noise measurement.docx

I also found this older thread on the THS4531A on popcorn, there are some good measurement traces in this one, 

Hi all,

the first two photos are from the link given by Michael. The last photo is from Kanji. The photos shall show the similarities:

Popcorn noise can even be seen in voltage regulators like the LM317. And I remember that we had to use the NE5534 in an audio project once. We had to hand select the NE5534 for ultra low noise because we found that even many of the NE5534A (low noise version of the NE5534) showed excessive popcorn noise. To our surprise it wasn't possible to detect all the popcorning NE5534 by the help of noise measurements. Even parts not showing any popcorn noise on the scope failed in a simple audio test carried out by our ears. So, after selecting the NE5534 for lowest noise we additionally performed an audio test with our ears to eliminate all popcorning NE5534. At the end of the day only 5% of all screened NE5534 could be used...

Kai

Thanks Kai, good background, to add a little more color -

We often talk about popcorn noise issues as random surface trapping states in the input stage due to better or worse processing cleaniness - that can be related to the overall process and a lot by lot thing -

One of my earliest experiences here was to screen CLC505's for no popcorn for an NSA project. I had found a particular wafer that was relatively good for that and carefully stored it in a benign environment for later order support.

There are some modern products that actually get a modest popcorn screen in final test. The problem there is you literally cannot wait long enough in ATE to screen really slow popcorn. And no one likes to screen as it quickly increases product cost and few want to pay for that. But it does happen to some extent on some parts - just need to ask as it not usually published.

Hi Michael,

I enjoy your contributions! Yee seem to have a huge bunch of experience and knowledge... :-)

Kai