UCC27712: UCC27712

Hey Jay,

Thank you for reaching out to TI with your question regarding the UCC27712.

Looking at your schematic, I don't see anything that would be of concern but I do have a few questions.

1. Can you take a waveform capture of HI, LI, HO-HS and LO?

2. When is this load being applied to the circuit? Is this in a startup condition when you are seeing this failure or is it in steady state? It would be good if you could explain what was happening up to the point of failure.

3. If you replace the driver, does the issue persist?

4. Can you take a measurement on the damaged driver on impedance with respect to VSS of all pins and then do the same for a known good part with the gate driver on the bench and off of the PCB?

Let me know if you have any questions.

Thank you,

William Moore

Hi, William,

Thanks for that detailed response!

1.

I am getting the waveforms you asked for.

I only have a 2 channel scope, so it is going to take a minute...

2.

It happens when I turn on the circuit with an inductor in the load (at HS)

3.

When I replace a driver, the problem goes away

4.

I checked out a "bad" device against a new one and the resistances are IDENTICAL!

A few K ohms here and there but almost exactly the same...

Please give me a day to get all the waveforms for you

Thanks for your help!

Hey Jay,

1. I understand that 2 channels is your limit. To simplify it, please take these captures. HI and LI, HI and HO-HS, LO and HO-HS, and LI and LO. It would also be good to take a capture of HO-HS and HS.

2. What is your startup sequence when you do this?

3. If you performed an A-B-A swap to determine if the problem follows the driver? This would mean taking the damaged part off, verifying the circuit with a good part and then putting the damaged part back on. Has any other components been damaged?

4. Can you please share that data with me?

Once you get all this data and scope captures, please update here as well as with responses to the above questions.

Thank you,

William Moore

Hi, William!

Here are the traces you asked for (attached)

This forum doesn't allow me to load the original bmp file so I had to print it to a pdf.

hope the resolution is goo enough to be clear....

I did the A-B-A test and the device IS dead

I suspect that the reason the impedance measurements don't vary between good and bad devices is because the hig side PFET in the UCC27712 is blown OPEN, not shorted...

Your EXCELLENT question about sequence of operations got me thinking that perhaps it is an inductive kick from the inductor that is causing a spike to feed back into the HO line of the UCC....

BTW, I was turning on the whole shooting-match in one swell foop rather than waiting for the voltages to stabilize before connecting the load to the driver

I am now going to add a switch between the HS line and the inductor of the load filter.

I am going to add a 12V tranzorb between the HO and HS pins of the UCC so it can catch and kill any large transients...

Please let me know if the pictures make any sense to you - I have annotated them the best I can and put them all in one picture so you don't have to switch between pics.

I will let you know how the further testing goes.

Thanks again and have a great weekend!

7534.Waveforms.pdf

Hey Jay,

Thanks for gathering these scope captures for this. Just so I am clear, is the left column of the damaged device and the right column of a good device? What operating condition are we seeing in these? Is this no load?

Yes, there are a lot of transients when starting up supplies and equipment in a test bench. So you must be careful in your startup sequence as this can cause damage to circuit components.

I am unfamiliar with the tranzorb solution as we do not see this in typical systems. So I cannot comment on its affect on the gate driver.

Once you have completed some further testing please update here and also please let me know about your comments to the above questions.

Thank you,

William Moore

Hi, William,

So I jiggered the circuit to add a switch before the load, so that I could turn it on AFTER the circuit stabilized.

I also left in the tranzorb between HO and HS.

I turned the circuit on, observed the 108V square waveform on the HS pin.

I then switched the L-C filter circuit into the system.

UNfortunately, my power supply browned out.

BUT, when I disconnected the L-C circuit from the system, the HS waveform was back!

So, either the tranzorb helped and / or the sequenced switching of the load helped...

Now I have to add a series resistor to change the time constant of the LC filter...

Fun and games!!

Did the waveforms tell you anything about what happened to the UCC driver chip?

Thanks!

Hey Jay,

I'm sorry to hear you have damaged your power supply, but it is good to hear that you have some confidence in a potential solution.

I did not see anything in particular with the waveforms but I would like to verify if the left column of waveforms is the damaged driver and if the right column is a good driver. What operating condition are these waveforms in? Is this a no load condition?

Let me know what your responses to the above are and if you come across anything else in your debugging.

Thank you,

William Moore

Hey Jay,

Have you been able to resolve this issue or can you provide responses to my above questions or an update to the status of this issue.

Thank you,

William Moore