INA233: Abnormal behavior when changing the temperature rapidly

Hi Yoshikazu,

heating up a chip or part of a circuit board with a hair dryer has several disadvantages:

1. The hot air stream does not show an even temperature profile over the whole air stream area but there are spots which may show a considerably higher temperature.

2. The temperature extremely depends on how close you go to the outlet of hair dryer.

3. An enormous amount of static charge can develop due to the air streaming causing severe ESD issues.

Because of all that a chip should not be heated up by a hair dryer. This merely results in a highly uneven heating of chip because the pin wires of chip, the epoxy package and the printed circuit board heat up unevenly and at different paces. This means an enormous stress (temperature shock !) for the chip and you can even damage the chip by this torture.

With good reason the whole printed circuit board has to be placed in an oven when performing temperature tests. Only an oven can guarantee an even heat up of the whole circuit.

And if in your application the printed circuit board would see asimilar uneven and sudden temperature changes, you would have to thermally shield the whole printed circuit to prevent temperature shocks and damage.

Kai

Hello Kai-san,

I know what you mean, but it's quite difficult to heat up rapidly if an oven is used.  The temperature they used is in the supported range shown in the datasheet and INA226 didn't have any issues, but INA233 did(I hear the hair dryer was around 20cm from the chip).  Such temperature rise could happen in the actual application, so I think it has to be supported.  At least the mechanism should be provided.  Otherwise most of the customers wouldn't want to use INA233.  Would you please test in your side as well and provide the mechanism?  They saw 7 put of 17 units, so you would be able to reproduce the issue if you test 3 units unless it has lot dependency.

I had a similar issue in the past on FPD-Link devices.  FPD-Link team supported us, provided the mechanism and implemented test program to screen such units.

Best Regards,

Yoshikazu Kawasaki

Hello Kawasaki-san,

Thank you for using the TI forum. I would not expect the device to fail at a rapid temperature change. I have a few follow-up questions to look into this further:

1. Can the customer look more closely at timing to make sure it is robust?
   1. Can we get a zoomed in scope shot to see how SCL looks relative to SDA?
   2. From the abnormal operation photo, how long does it take for SDA to recover? (which would be some time to the right of the image)
2. I noticed that SCL looks different between the normal and abnormal regions also. Was that supposed to look the same between both regions? SCL is driven by the MCU, so I wouldn't think it would look that inconsistent. So maybe the problem is with the MCU/timing?

Regards,

Mitch

Hello Mitch-san,

Thank you very much for your comments and I'm sorry for my late reply, but we're checking your comments to the customer now.  Please wait for a while until we get information from them.

Do you think you can check if you can reproduce the issue in your side as well?  I'm not sure what the root cause is yet, but if it doesn't have lot dependency and is a common issue, you would be able to reproduce the issue in your side as well if you check 3-5 samples.

Best Regards,

Yoshikazu Kawasaki

Hello Kawasaki-san,

I'm not sure if our ovens will bring up the temperature fast enough to reproduce the error. Can you tell me how long it takes them to get their device up to temp? I'll look into seeing if we have a faster way that would be safe for the parts, but I think we may need to just wait for the info from them.

Regards,

Mitch