UCC23513EVM-014: UCC23313: Switching frequency range of the optocouplers

Hi, Shay,

Welcome to e2e, and thanks for your interest in our products!

Sometimes when driving IGBTs and SiC MOSFETs, you want a negative voltage from gate-to-source when they are turned-off to insure they stay turned off, especially if the switch node is slewing voltage (say in a half-bridge application).

So, we provide two options - you can input two separate supplies to generate this negative bias, or, you can use the zener diode circuit to generate a negative supply on the board to provide the negative bias.

Let us know if you have further questions.

Hi, Shay, thanks for your question

Shay Borenshtain said:

1. What is the difference between single and double supply? What is the purpose of a single supply if we are talking about Split Supply operation?

UCC23513 is a single-output driver (one OUT pin). This section is talking about the difference in the external components required for a Unipolar supply

A unipolar supply only has a single, positive supply. That means when the driver turns the IGBT off, it pulls the gate down so VGE=0 (or in the case of Mosfet, VGS=0)

For example, a unipolar supply could look like:

• VCC pin -> +15V
• VEE pin -> GND (tied to IGBT Emitter/FET source)

A bipolar supply, there is a positive and negative supply, referenced to ground/Emitter/Source. That means when the driver turns the IGBT off, it provides negative bias, meaning VGS/VCE are actually negative, so you could say the IGBT/FET is MORE turned off than with a unipolar supply.

For example, a bipolar supply could look like:

• VCC pin -> +15V
• VEE pin -> -5V (referenced to IGBT emitter / source)

The purpose for providing a negative bias at turn-off? When the IGBT is switched off, the IGBT/FET gate is pulled low each cycle, the VCE can spike quickly which creates a large dv/dt transient that can couple through the Miller cap of the device, injecting some current to the gate cause the gate voltage to swing up.

While dv/dt on its own has the potential to damage the device by exceeding VCE limits, there are other problems high dv/dt can cause when switching.

This can kind of be understood by Q=CV ->  I=C*dv/dt. When the transistor is switched "harder" (more sink current), or the Bus voltage is larger, or Miller cap is larger (size of device), the Injected current is larger and it has increasing potential to swing the gate voltage up and falsely turn the device on.

If the lowside device falsely turns on, for example, there is essentially a short-circuit from VBUS to ground, which is very bad. At very least, it will damage the switches. So you really want to ensure that those switches stay off.

This is why a Bipolar supply is important, by pulling the gate voltage below the source, can ensure the device stays off despite Injection current thru miller clamp. Miller clamp also helps this (not a feature of this device though). Miller clamp can also be used in conjuction with bipolar supply.

• Excellent Explanation of dv/dt Induced turn-on and how to prevent it

Shay Borenshtain said:

2. In the first method (Single supply), what's the purpose of the zener diode and resistor branch? From the diode's specification, it seems it's protecting against VSS voltages lower than -5V, so why is the resistor in series with VCC?

Having a bipolar supply can add to cost and size of the system compared with a unipolar, so sometimes designers might not want to implement a bipolar isolated supply.

The Zener diode option in the EVM is a kind of workaround to get Negative turn-on bias with a unipolar supply, basically C6/C7/D2 are populated. I've screen capped it from the EVM document below.

A better way to look at it is from a different devices datasheet, show below, so we can get a better idea how its connected to the power switch.

The true purpose of the zener doesn't have anything to do with protection. Pay close attention for the unipolar supply (25V), Ca1, Ca2, Rz, and zener in below.
Zener essentially Forces VSSA to be referenced a zener drop below the Source/Emitter of the switch. (or whatever zener clamp voltage of that diode)
As you can see, that means that we can effectively have a negative pull down bias with just a Unipolar supply.

Notably, this Config does requires some current to "leak" through Rz in order for the voltage between Emitter and VSSA to be clamped like that. So the key drawback with this setup is that it can't work with bootstrap supply, which is commonly used on highside drivers.

A topology to get negative turn-on bias with a SINGLE unipolar supply, and with bootstrap on highside, (which is not implemented in the EVM), looks as below.

However it has challenges. CANT be used unless duty cycle is fixed (50%), and requires some fine tuning for sure.

• These figures came from this datasheet section (UCC21540 dual channel driver)
• For your purpose, you can compare them to using 2 Seperate drivers, just cut them in half where it says "functional isolation"

Hopefully this answered your questions. If it did, please let me know by clicking the green button. Please, feel free to ask us any more questions as they come up while youre working with the UCC23513EVM.

Best,

Dimitri