INA250: Reding different values in current monitoring

According to test and given circuit below, it seems the gain ration can differ between same ICs'. is this normal? Again, according to test gain ratio differs between minimum 1,53 V/A and maximum 1.97 V/A. 

Hi Harun,

These numbers appear to be too far from datasheet spec.

I suggest you measure the load current for each data point. If you have already done so, please forward the data data and associated calculation.

Regards, Guang

Hi Harun,

is what we see the complete schematic?

Kai

Hi Kai,

These are the gerber views:

Top:

Bottom:

This is what the PCB looks like:

And this is the test circuit:

I'll send you detailed test results when possible. But now we are looking for equivalent products as quickly as possible.

Hi Harun,

are there are some shorts between the solder pins?

Also, have you carried out a star point grounding in your measuring setup?

Kai

Hi Harun,

I assume a DC load current is applied? When possible, please forward test data and scope measurements.

Regards, Guang

Hi Guang,

Yes, the supply current is DC. For a quick test that can reveal the problem, we chose two subjects to represent the perfect circuits(Sample 1 and sample 2) and two subjects to represent the circuits that we thought were defective(Sample 3 and sample 4).

This is the test result of four INA250A4:

And these are the test subjects:

Hi,

Thank you for the data. The contrast is clear.

Maybe you could insert a current meter in series with the load as anther means to measure the current. In addition you could try to swap INA250 between a good sample and a bad sample. Repeat the measurements to see if the defect travels with IC unit.

Regards, Guang

-"you could try to swap INA250 between a good sample and a bad sample. Repeat the measurements to see if the defect travels with IC unit."

We already tried this: we doubted the thickness of the stencil(0.12mm) if a short circuit may occur due to excess soldering. Several times we tried to switch ICs between good and bad samples but defect replaces with IC. Then we thought we should report this to this forum. We doubt whether the ICs we use are original or not. We cannot contact the place where we supplied the ICs. I think because of the virus. 

-"Maybe you could insert a current meter in series"

Another available time, of course. But for this test, we focused on why it yields different values rather than the accuracy of the product.

Best Regards
Harun

Hi Harun,

Thank you for the confirmation. It does sound like these samples are due to defective IC. What is the rough percentage that is bad?

I would say swap them out with brand new ones and say what happens. Meanwhile save the bad ones for future replacement and analysis once you get hold of the distributor.

Regards, Guang

-"What is the rough percentage that is bad?"

7 of 35 samples had this problem.

-"I would say swap them out with brand new ones and say what happens. Meanwhile save the bad ones for future replacement and analysis once you get hold of the distributor."

This is exactly what we will do.

Regards
Harun