Electronics > Projects, Designs, and Technical Stuff

MOSFET Gate Driver not working as datasheet describes (solved)

(1/2) > >>

IMI4tth3w:
Edit:

Just wanted to put here that i was using BNC extensions for my scope probes that completely threw off my compensation. Rookie mistake.

Original post below:



Greetings everyone!

My question today is about MOSFET gate drivers. The IC i am using is the Microchip TC4427A MOSFET gate driver.

I am trying to use a 5V PWM signal from a uC to drive the gate of a MOSFET (Vishay SQJ974EP) at 12V.

The main issue i am having is that i understood from the TC4427A datasheet that the MOSFET driver should work effectively as a schmitt trigger. However, i find that the output almost directly matches the input.

Here's the schematic (12V bypass caps not pictured. But they do exist)


The image below shows channel 1 as my 5V input from the uC. PWM frequency is 1600Hz with a 50% duty cycle.
Ch2 is the output of the gate driver. The gate driver is just breadboarded with no outputs connected.

As you can see, the output directly follows the input, just at a higher voltage. From the internal schematic of the TC4427A with all the buffers, it seems like this should be impossible. Shouldn't the output be a nice square wave? Datasheet says that any voltage over 2.4V should be logic "1", and give full 12V on the output.

I have also swapped the TC4427A with a MIC4469YN (another gate driver IC) i had laying around, and i get the EXACT same output.

Are these IC's not applying any hysteresis? Should i just switch to an actual schmitt trigger IC? Surely two gate driver's can't both be wrong.

And the more important question: Should i care? I feel like its not ideal for my MOSFET to spent 400us every pulse not in saturation as i have up to 20A of current coming down the pipe.

Thank you for your help.

Matt

T3sl4co1l:
Check probe compensation?  Or is that a bad ground connection?

The output is delayed by maybe 30ns, check the datasheet.  You can't see that delay on this scale.  You need a triangle wave to measure the hysteresis loop.

Tim

IMI4tth3w:

--- Quote from: T3sl4co1l on January 22, 2019, 11:39:01 pm ---Check probe compensation?  Or is that a bad ground connection?

The output is delayed by maybe 30ns, check the datasheet.  You can't see that delay on this scale.  You need a triangle wave to measure the hysteresis loop.

Tim

--- End quote ---

Hi Tim,

Probes are set to x10 and properly calibrated to the scope test connector.

Grounds are also solid. I have a PCB and breadboard version of this circuit. Both give identical waveforms.

The 5V PWM waveform is unchanged when disconnected from the gate driver input.

The 5V PWM waveform is nearly a triangle wave. I'll get more screen shots with cursors tomorrow.

The main issue is the expected square wave output, which i am definitely not getting.

Doctorandus_P:
Probe compensation problems sure look like the first thing to check.

Gate drivers need to deliver high current pulses, which have to come from nearby decoupling and bulk capacitors, because the inductance of the power leads is usually to high. I do not see any capacitors in your circuit.

I would not be surprisd if the the gate driver pullls it's own supply voltage down to a voltage that it does not work properly anymore.
What waveforms do you measure on the power supply pins of the gate driver?

Also: With high current flanks PCB layout becomes important. You can not expect these gate drivers to work correctly if you build something on a breadboard.

MasterT:
TC4427,  Vdd = 12V, Input in series 1 k, Output loaded 2.2 nF.
Sine, rectangular, triangle.  Slow slope unstable. 
Last picture shows hysteresis, about 150-170 mV. Switch threshold 1.65 V.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod