INA126: Sources of Output Noise

Hey Cafain,

Could you share the circuit you are using with the INA126?

If you are unable to share a schematic, could you provide the following information?

1. What is the input impedance on IN+ and IN-?
2. What is providing the supply voltage?

Popcorn noise at that gain would likely only show up with high input impedance. TI does screen for popcorn noise on their devices, but due to the random nature of this behavior, and the random intervals that it occurs, it is not possible to screen this out 100%.

Best,
Gerasimos

Hi, Gerasimos san

Thanks for your support.
Since this is a private circuit for a user, I would like to send it to Gerasimos-san's private email.
What do you think?

regards cafain

Hello Cafain san,

I have sent you a friend request, you may message me directly on E2E to share the schematic.

Best,
Gerasimos

How can I check my friend requests? I haven't received any.

Hello Cafain-san,

Is the oscilloscope AC coupled or DC coupled? What gain resistor is in use when the noise is being observed? Typically popcorn noise will be a current noise on the input transitioning between two discrete states. Please see below image for an example. This is most frequently seen with high input impedance, and in high gains.

Best,
Gerasimos

Gerasimos san

Thank you for testing it.

It's DC coupled.

Gain=23 (G=5+80k/4.53k).

Is it possible that there's a problem with the device?

regards

cafain

Cafain-san,

With a gain of 23, your input noise will have to be 6.21mVpp to show up as 143mVpp on the output. With a typical 0.1-10Hz noise of 0.7uVpp, this is very unlikely. If this is caused by a current noise, a 6.21mVpp with a 10kOhm input resistance would mean that your input current noise would have to be 621nApp. The typical noise of this device is 7.3pApp. This means it would have to be 100,000 larger than typical noise. Both of these feel unlikely, but may be thrown off by the high frequency components of the noise.

Is the output filtered at all to remove high frequency components of the noise? I would recommend a 120kOhm and 10nF capacitor on the output connected like the below image:

The behavior does not appear to follow the typical characteristic of popcorn noise in the oscilloscope, but this may be due to the high frequency components of the noise.

For an occurrence rate of 1 out of 14,520 it is a similar occurrence rate to popcorn noise. TI does test this device for popcorn noise, but due to the random nature of popcorn noise, it is not possible to screen out this behavior 100%. If this is popcorn noise, this is an inherent risk when designing with any bipolar amplifier.

Just for confirmation, the customer has performed an ABA swap, and the behavior follows the device?

Best,
Gerasimos

Gerasimos san

It would be very helpful if you could verify the circuit.
Connecting an output filter is not practical because so far no noise has been confirmed to occur.

We have checked the A-B-A swap.
As a result, the noise symptoms disappeared when the device was replaced.
There have been no other similar issues in the past.
(No noise was confirmed in any other device.)

regards
cafain

Cafain-san,

I do not see any obvious issues with the circuit. Unfortunately, popcorn noise is a random occurrence, and may happen in any bipolar device randomly in any lot. As I said, we do try to screen this behavior, but due to the random nature, it is not certain that we can capture 100% of this behavior at the screen.

Best,
Gerasimos