UCC5350-Q1: The Vgs waveform of high side SiC MosFET using UCC5350 by bootstrap power supply in motor control application

A.W

Part Number: UCC5350-Q1
Other Parts Discussed in Thread: UCC5350, UCC21551

I'm testing UCC5350 for SiC driver in PMSM control application(600VDC input).

I use bootstrap method to give the high side UCC5350 Power supply.

The waveform of the Vgs is a little strange, it has a voltage overshot（about 3.5V） when turning on and negative negative transients when turning off.

I use the high voltage differential probe(TMDP0200) of Tekronix for Vgs measurement.

Is this a normal phenomenon,or is there something wrong with my test process?

Regards

Arrow

10 months ago

0 Don Dapkus 10 months ago

TI__Guru**** 189399 points

Hi, Arrow,

Your waveforms look close to being correct, but the switching edges do look a little abnormal. Reasons for this might include:

1. Bad probe

2. Uncompensated probe

3. Improper probe bandwidth

4. Excessive gate inductance loop on your PCB

5. Capacitive loading of the gate loop

6. Damaged component on your PCB

If you have more prototypes, I'd suggest testing multiple boards to see if this is the same across different boards to help narrow the root cause down.

Best regards,

Don

0 A.W 10 months ago in reply to Don Dapkus

Intellectual 330 points

Hi Don

I find that the probe has much influence with the test result.

Is the CMRR vey important for high side SiC MOS's Vgs monitoring?

I use two differential probes, one for Vgs，another for Vg（with positive and negative clips short).

The single Vg point also has ripple when SiC switching.

If I use Vgs-Vg as the new waveform, it is much reasonable.

Is there any other way to test the Vgs. It seems that the Optical-fiber Isolated Probe is very expensive.

Regards

Arrow

+1 Don Dapkus 10 months ago in reply to A.W

TI__Guru**** 189399 points

Hi, Arrow,

Yes, it is difficult due to the large voltage offset and trying to measure a 15V signal sitting on top of 600V.

This has some pretty good suggestions - https://www.eevblog.com/forum/projects/how-to-probe-high-side-mosfet-gate/msg3088297/?PHPSESSID=4oae39abudcabca4pus5vmjs56#msg3088297

You may also be able to rent or borrow one of the optically-isolated probes. Those are what we use in our lab.

Best regards,

Don

0 A.W 10 months ago in reply to Don Dapkus

Intellectual 330 points

Hi, Don

one more question about the high side driver bootstrap structure.

I take the bootstrap structure as UCC21551 datasheet shows, like Figure 10-1 and Figure 10-2 . I use Vz for negative bias voltage for SiC Mos.

I think Vz will perform 2 state ,one is forward conducting when low side turn-on and the other is reverse clamping when high side turn-on

So will this limit the dead time?

The zener is not a fast recovery diode.It has a large trr than the bootstrap diode.

If I limit the dead time to 1us, will the zener diode heat a lot?

Regards

Arrow

0 Alexander Mazany 10 months ago in reply to A.W

TI__Genius 11760 points

Hi Arrow,

Sorry for the late reply; yesterday was a US holiday. Don is also out of office, so I will assign this thread to someone else. They should be able to reply by tomorrow.

Thanks,

Alex M.

0 Krystian Plaskota 10 months ago in reply to A.W

TI__Expert 4065 points

Hi Arrow, what diode are you using? I think it will be fine as 1us is a pretty long deadtime. What is the trr of the diode you are using?

0 A.W 10 months ago in reply to Krystian Plaskota

Intellectual 330 points

Hi Krystian

I'm using 1SMA5913BT3G as the zener diode.It seems that the trr is not in the spec.

/cfs-file/__key/communityserver-discussions-components-files/196/1SMA5913BT3_2D00_D.PDF

Regards

Arrow

+1 Krystian Plaskota 10 months ago in reply to A.W

TI__Expert 4065 points

I would suggest testing it, but 1us is a long deadtime and I don't expect the zener to overheat during switching.

0 A.W 10 months ago in reply to Krystian Plaskota

Intellectual 330 points

Krystian Plaskota said:
deadtime

Hi Krystian

I will test the deatime setting. For now I find that the different dead time will lead to the different bootstrap voltage.

The bootstrap voltage in 2us deadtime will higher than in 1us.I'm going to close this thread and do more test.

If coming with another problem, I will open a new thread.

Regards

Arrow

Power management

Power management forum

UCC5350-Q1: The Vgs waveform of high side SiC MosFET using UCC5350 by bootstrap power supply in motor control application