ADS8920B: Failure mode for ADS8920

Hello Marco,

These types of failures are usually caused by electrical overstress. One or more of the ABS max ratings, shown in section 6.1 of the datasheet, have likely been exceeded. Once the ABS max ratings have been exceeded, the internal ESD diodes will turn on, and if the source of the transient can supply large amounts of current, the internal diodes will overheat and fail as a short (can be less than 1ohm to several 100's of ohms depending on the severity of the transient).

Many times, the transient voltage occurs on the inputs, but in your case, since multiple devices were damaged, it could have shown up as a transient on the power supply pins.

1. Are the ADS8920 devices powered by a system supply, or are they powered by a lab supply?
2. Do all of the damaged ADS8920 share a common circuit topology and are these different from the other ADS89820 that survived?
3. I assume you are driving the input of the ADC with an external amplifier. Are these external amplifiers powered from the same supply as the ADS8920B? If not, these amps could damage the inputs of the ADC if their outputs exceed the ABS max input voltage rating of the ADC.

If you can provide a screenshot of the ADS8920 schematic, including the reference circuit and input amplifiers, that would be helpful.

Thanks,
Keith N.
Precision ADC Applications

Hi Keith,

Thank you for your answer, I would agree on the transient on the power supply pins.

To answer your questions:

1) There are three regulators on the board, common for all the ICs, (for analog, Vref and digital supply). There are still operational

2) They all share the same circuit topology

3) There is no amplifier, just a sensor again connected to the same rail. However the analog inputs still have an impedance >10Mohm to the rails, so I guess there was no over-voltage there.

Unfortunately I can't share the design, anyway I don't think there is much to learn from it.

I was thinking: because RVDD failed at a significant lower impedance (10Ohm) than REFIN (100 ohm), would it be sensible to assume that was RVDD to cause all of this (or at least to fail first) ?

Regards

Hi Marco,

RVDD seems to be the primary path.

Are you using an RC filter on the input of the REFIN pin, similar to Figure 110 in the datasheet? 1kohm and 10uF will help protect this input. Also, did you follow the board layout recommendations in Figure 121, especially the ground and REFM pins?

If yes to both of the above, you may want to add a clamp and additional capacitance on the RVDD pin. Without schematics or a board layout, it is hard for me to make any other suggestions.

Regards,
Keith N.