UCC27201A-Q1: Shorting HB and HO pins during operation.

Hi Shinu,

I have sent your question to the appropriate Team for follow-up and you should hear from them soon.

Hi Shinu,

I am sorry to hear that customer is having failures. We will try our best to resolve the issue.

I would recommend that you submit the failed parts for failure analysis. The local sales person should help in this regard.

May I ask does power MOSFET fail when driver fails?

In other words, did you have instance where you just replaced driver and everything worked ok.

May we also know the failure rate. Number of drivers used so far and number of drivers failed.

Is there a particular operating mode when the driver fails or customer does not know when the driver fails?

LM5109B-Q1 datasheet details how to calculate the bootstrap capacitor value.

For 6kHz operation with 35-85% duty cycle, there is enough time for 1uF bootstrap capacitor to charge and there is enough time for bootstrap diode to recover.

But I would be able to tell more if I know the power MOSFET part number.

When you say HO and HB shorted with 14 ohm, do you mean there is 14 ohm resistance between HB and HO?

Are these pins or any other pins shorted to any other pins?

Do all failed drivers have the same failure signature?

There are few possibilities, such as over voltage on HO or HB with respect to HS. It is also possible that HS pin experiencing excessive negative voltage.

It is also possible that the ESD cell is getting damaged somehow.

What is the bus voltage as I see 160V electrolytic capacitor on the bus.

I doubt this is the case but it is also important to check whether there is any cross conduction on input or output.

If there is excessive negative voltage on HS, it could cause device to latch or device might put wrong pulse, which might cause cross conduction.

The solution for this problem is better layout and fast low leakage schottkey diode close to chip across HS to VSS.

I see that there is a zener from HO to HS, but if there is fast transient from HB to HS which would translate to HO and might cause HB to HO short.

Improving the layout can help reduce ringing both on HS and HB.

If possible please provide waveforms of HB to HS and HO to HS in worst case condition, such as start-up, load transient, short-circuit, etc...

Also, let us make sure that we FA the failed parts.

If excessive negative HS voltage is root cause, customer might want to try UCC27211A-Q1, which has better negative voltage handling capability (and it is pin to pi compatible with UCC27201A-Q1).

Let us keep investigating till we get to the root cause.

Thanks for bringing this to our attention and your patience in this matter.

Regards,

Ritesh

Hi Ritesh,

Thanks for the reply.

Please find the attached plot of the voltage at HS pin while the failure had occurred.

On the time line at the end you can see the HS pin shows 9V which is near to driver's power supply (10V). (When fails)

Moreover HS is never crossing above 16V or below -5V.

UCC27201A-Q1_HS PLOT.pdf

Find below layout for your expert views and comments.

UCC27201A-Q1_Layout.pdf

Thanks in advance for your help and support.

Regards,

Shinu Mathew.

Hi Shinu,

It is difficult to diagnose the problem looking at just one waveform.

As I recommended earlier, it would be good if customer can support some of the failed parts for FA.

I also observed some negative voltage on HS for more than 100ns, which might also cause damage to the part.

It seems that one or more parameters are exceeding its recommended limit and that is causing the failure.

We need to look at HB, HO, and HS during normal operation and see if any of the these parameters are exceeding recommended min or max limit.

Let us keep investigating.

Regards,

Ritesh

Hi Ritesh,

Thanks for the reply.
The negative voltage in waveform is not even crossing 1.5V it is far above IC withstanding capacity. We are triying to get few samples of failed IC. How we can analysis from FA as always the end result shows "Electrical Over Stress". We will collect the samples and will send to you for further analysis.

Regards,
Shinu Mathew.

Hi Shinu,

The -18V spec on HS is only for 100ns. The DC spec is -1V.
Therefore, if there are instances where voltage is more negative than 1V for significant amount of time such as microseconds, the part might get damaged.
I also suggest to use special trigger so that we can capture waveform at faster speed, such as 100s of ns per division.
Try to trigger on HB voltage and keep the triggering point above Vin+Vdd (if measuring with reference to VSS).
This way we can capture any voltage spikes that might occur on HB and therefore on HO as well.
Now let us talk about FA.
From FA we would be able to know which section of the IC is getting damaged, whether it is output stage or ESD cell, or something else.
Sometimes this help in reverse analyzing possible failure modes and then we can ask customer to reproduce the failure.
If signature matches, then we can say that somehow IC is getting into that particular failure mode.
I hope this helps.

Regards,
Ritesh

Hi Ritesh,

Thanks for the inputs.
Can we send failed parts to you for analysis or should through local FA team.
Once again thanks for your help and support.

Regards,
Shinu Mathew.

Hi Shinu,

I have never been involved in FA process and therefore, it would be better if you could go through local FA.

Ask them that you need to know which part of circuit is damaged. It would be great if they can tell us whether it was an instantaneous over voltage event or extended over current type of failure.

Regards,

Ritesh