DRV8353R: Variation in gate rise times

Hey Oliver,

It seems like screenshot #2 seems to be nominal turn off of the FET, while screenshot 1 and 3 are abnormal. What changes did you make between these screenshots?

The FETS you're using have a nominal Qgd of 34nC, while using a IDRIVE setting of 300mA, the resulting rise time is 113ns (actually it's a little bit slower because of your series gate resistor, if you let me know your VBAT voltage, I could tell you exactly what the IDRIVE strength is). As a general rule of thumb we ask customers to design their system such that their rise and fall time is between 100ns-200ns. Rise time can be calculated with the following equation:

Risetime = Qdg/IDRIVE strength

Just as a test, can you try lowering your sink and source current to 200mA/300mA

Thanks,

Michael

Hi Michael,

There were no changes between screenshots 1-3. I was running a motor at constant load and just pressing the single capture on the scope. I am planning on doing a capture of a number of consecutive waveforms to see the frequency of good vs abnormal waveforms.

VBAT is 48V but could drop to as low as 38V.

I'm on the hardware device so will have to do 150/300mA bu I'll get that test done today too.

Thanks

Hi Michael,

I have done some testing with IDRIVE at 150/300mA. This has a big impact on the consistency of the mosfet switching on but not quite as significant as the switch off.

I took a video of the scope to show the variation in case that is useful. But below are screenshots of the scope at 150/300mA IDRIVE and 450/900mA.

Above is the switch on of  150/300mA Idrive. The turn on time is constant at 700ns. The shape of the wave also seems consistent.

Above is the switch on of  450/900mA Idrive. The turn on time is varies between about 500ns and 1200ns. The shape of the wave varies with a much larger miller plateau in one.

Above is the switch on of  150/300mA Idrive. The turn off time varies between 600 and 1200ns. The shape of the wave also varies slightly.

Above is the switch on of  450/900mA Idrive. The turn off time varies between 400 and 1000ns. The shape of the wave also varies slightly. In the video the wave shape changes much more frequently than with the lower IDRIVE.

Did you expect the turn on time to be more stable for a lower IDRIVE?

The mosfets are also still switching a lot slower than required and what would be expected using the gate charge/IDRIVE equation. I would be expecting to see just over 230ns rise time when taking into account the resistor.

Thanks,

Hey Oliver,

Apologies for not getting to this today. I'm going to review this with the team tomorrow, so I'll get back to you then. If possible, can you provide the rest of your schematic? You may send this to me via direct message on E2E if you don't want to post it on this public forum. I've sent you a friend request.

Thanks,

Michael

Hi Oliver,

Sorry for not noticing this earlier. When we are talking about the rise and fall time of the FETS, we are referring to the time it takes for the MOSFET to turn on, so we want to see the VDS time, not VGS. VGS isn't as important in the system because we don't really care how long the gate capacitor takes to charge. 

As for whether or not the turn on time is more stable as a lower IDRIVE:

Typically yes, because given the average Qgd of the FETS in motor drive applications, we don't recommend customers using the highest IDRIVE setting. Doing so results in rise/fall times that are out of range from what we recommend (100ns-200ns).

Good job with the schematic, I don't see any issues with this.

I will say that bulk caps only need to be rated to 1.5x - 2x the input voltage. 

Let's take a look at the VDS rise time and see if it's an appropriate value.

Regards,

Michael

Hi Michael,

Sorry it's taken a while to do the tests. Had a few issues which took longer to resolve than expected.

I have switched to measuring both gate-source and drain-source on the low side MOSFET as it gives a better picture of what is happening I think. The screenshots are below.

Above we have gate-source (yellow) and drain source (blue). In the left image, the behaviour is as expected. The drain-source voltage drops when the gate threshold is reached.

On the right, the drain source voltage drops before the gate is switched. 

These two screenshots are taken on the same mosfet under the same load Just pressing the single waveform button on the scope.

The behaviour is the same for the gate switch off.

On the left is the correct behaviour, on the right, the voltage doesn't rise until about 600ns after the gate threshold is passed.

in terms of the drain source switching time, it is stable when operating correctly, but can take a long time to switch in the case when it starts dropping before the gate is switched, see below:

thanks,

Hey Oliver,

I'll back to you on this tomorrow morning. 

Regards,

Michael

Hi Michael,

Thanks for this, it makes sense. I will have a current probe next week so should be able to see which way the current is flowing when this behaviour occurs.

Sounds good Oliver. Let me know if you have any further questions once you get the current probe.

Regards,

Michael

Hello Oliver,

Were you able to run your experiments and do you need any additional help on this thread?

Thanks,

Matt