Active bridge rectifiers use MOSFETs in place of diodes for reduced voltage drop. I suppose the simplest way to achieve this would be to turn on the MOSFETs whenever the source voltage is greater than that of the opposing MOSFET, but that wouldn't work if there's a capacitor on the output of the bridge rectifier, as that would result in the capacitor getting discharged by the MOSFET whenever the voltage is falling.

How do active rectifier circuits prevent this from happening?

Do they use a comparator to compare the MOSFET's source and drain voltage? If so, then how is this configured in the case of higher-voltage (100V+) bridge rectifiers? Do they have the comparator supply voltages tied to the MOSFET source voltage, or do they have a comparator tied to ground?

 

how is this configured in the case of higher-voltage (100V+)

Active rectification is not used in high voltage applications. They are used in high current low voltage applications. The reason; a 3.3V 1000A supply will have 0.65V X 1000 = many watts of loss. A 1000V 3.3A supply will have 0.7V X 3.3 = loss.

There are many different ways to make active rectifiers. The last one I used. looks for the voltage drop across the diode in the MOSFET. If the voltage is slightly positive, turn on the MOSFET. If the voltage is 0 to slightly negative, it means that current is going backwards so the MOSFET is turned off.

 

Are You asking for a detailed description of how such a Chip or Circuit works ?
That would be very impractical.

You need a reasonable knowledge of how many diverse components operate,
and how they might be combined to produce a very low Drop-Out-Voltage Active-Rectifier.

This will also depend upon the circumstances and specifications of all of the
Components or options that could be involved or employed to achieve a specified end-goal.

Start-out with ........
What is the desired overall function of this Project ?
Why do You think You need or want an Active-Rectifier for your project ?
What is the Voltage-Source ?, ( Frequency, Voltage, Current ).
What is the desired Output-Voltage and Current ?

For reducing Heat-Dissipation in a High-Power, Mains-Voltage Project,
You might try something along the lines of the Schematic below ........
.
.
.

.

 

Active rectification is not used in high voltage applications. They are used in high current low voltage applications. The reason; a 3.3V 1000A supply will have 0.65V X 1000 = many watts of loss. A 1000V 3.3A supply will have 0.7V X 3.3 = loss.

I doubt this is entirely true. Surely for a trolleybus or railway power supply they would prefer active rectification to using diodes with a whole lot of cooling.

There are many different ways to make active rectifiers. The last one I used. looks for the voltage drop across the diode in the MOSFET. If the voltage is slightly positive, turn on the MOSFET. If the voltage is 0 to slightly negative, it means that current is going backwards so the MOSFET is turned off.

But how did this work? You must have used a comparator. What voltages were the comparator power pins relative to the MOSFET? Was it powered from ground to a voltage higher than the output voltage of the rectifier? Or was it at a voltage relative to the MOSFET source?

 

Here are two methods I've thought of myself for how to achieve this:


The first one uses a charge pump and a comparator with their voltages tied to that of the MOSFET source. This means that the voltage across their power pins will be roughly constant, but their voltages relative to ground (not the "GND" pin, but ground) will move up and down in a sine wave. It may be possible to get the same result using an ideal diode controller.

The second one instead uses a resistor divider and a current transformer for the comparator inputs, so that the negative supply pin on the comparator can remain at ground. A hall effect sensor of some other kind of isolator may also work as an alternative to the current transformer.

Are there any MOSFET driver ICs that already provide this function?

 

Active bridge rectifiers use MOSFETs in place of diodes for reduced voltage drop. I suppose the simplest way to achieve this would be to turn on the MOSFETs whenever the source voltage is greater than that of the opposing MOSFET, but that wouldn't work if there's a capacitor on the output of the bridge rectifier, as that would result in the capacitor getting discharged by the MOSFET whenever the voltage is falling.

How do active rectifier circuits prevent this from happening?

Do they use a comparator to compare the MOSFET's source and drain voltage? If so, then how is this configured in the case of higher-voltage (100V+) bridge rectifiers? Do they have the comparator supply voltages tied to the MOSFET source voltage, or do they have a comparator tied to ground?

The basic idea is very simple. If the voltage falls below a small threshold voltage then the MOSFET is turned off. It's as simple as that although the implementation may be more complicated. The threshold voltage can be low like 0.1 volts or even 0.050 volts or even lower. When the voltage across the MOSFET falls below that level, the MOSFET is turned off.
There could be a little more to it though, where that threshold voltage is actually regulated so that it stays at that level (like 0.050 volts) until the regulator drops out of regulation when the voltage naturally falls below that level, and that is when the MOSFET is turned off.
I suppose there would be some hysteresis also so that the circuit does not oscillate. When the MOSFET is turned off, it could be natural for the output to fall to a lower value rather quickly and that would cause the MOSFET to turn back on right away. A small hysteresis value would help keep this to a minimum although it could not be too high.
A lot of power circuits work at only 50Hz to 400Hz anyway so the circuit would have to be especially fast for those applications.

 

Read data sheets.

UCC24612
diodes

 

Read data sheets.
View attachment 314049
UCC24612
diodes

Thank you. This looks like a rather cheap way to get an active rectifier for my project. I will probably use this or a similar chip. It looks like it has everything I want, aside from the package being small and probably difficult to solder. It wasn't in the section of Digikey that I expected.

View attachment 313987
The second one instead uses a resistor divider and a current transformer for the comparator inputs, so that the negative supply pin on the comparator can remain at ground. A hall effect sensor of some other kind of isolator may also work as an alternative to the current transformer.

This design that I posted earlier probably won't work, as the current at the primary side of the current transformer is mostly flowing in one direction. If my understanding of transformers is correct, than the direction of current on the secondary side won't reflect the direction on the primary side. I don't know much about hall effect sensors, but I suppose it might work with those, though the limited response time may be a problem for some circuits.

 

Your understanding of Transformer-Polarity is incorrect.
A change in Primary-Current-flow direction always results in the Secondary-Current-flow reversing.

This includes Hall-Effect-Current-Sensors.
.
.
.

 

Here are two methods

I have used two methods.
1-The IC watches the voltage or current in the MOSFET and drives the Gate.
2-Some PWM have outputs that drive the Gate(s). The PWM controls the timing and does not look at current in the MOSFETs.

 

Active bridge rectifiers use MOSFETs in place of diodes for reduced voltage drop. I suppose the simplest way to achieve this would be to turn on the MOSFETs whenever the source voltage is greater than that of the opposing MOSFET, but that wouldn't work if there's a capacitor on the output of the bridge rectifier, as that would result in the capacitor getting discharged by the MOSFET whenever the voltage is falling.

How do active rectifier circuits prevent this from happening?

Do they use a comparator to compare the MOSFET's source and drain voltage? If so, then how is this configured in the case of higher-voltage (100V+) bridge rectifiers? Do they have the comparator supply voltages tied to the MOSFET source voltage, or do they have a comparator tied to ground?

Have a look at this datasheet:
https://www.diodes.com/assets/Datasheets/ZXGD3113W6.pdf
I think it explains it quite well.
It is important to note that active bridge rectifiers are less likely to be used at high voltages as the voltage drop from a single diode becomes insignificant in comparison to the power supply voltage, and also that MOSFET on resistances increase with their voltage ratings, so the gains in efficiency become insignificant.

 

Have a look at this datasheet:
https://www.diodes.com/assets/Datasheets/ZXGD3113W6.pdf
I think it explains it quite well.
It is important to note that active bridge rectifiers are less likely to be used at high voltages as the voltage drop from a single diode becomes insignificant in comparison to the power supply voltage, and also that MOSFET on resistances increase with their voltage ratings, so the gains in efficiency become insignificant.

Interesting. Unlike the UCC24612, this one has an output voltage linearly proportional to the drain-source voltage, which means that the voltage at which it turns on would depend on the MOSFET's threshold voltage.

Active rectification is not used in high voltage applications. They are used in high current low voltage applications.

Despite my initial skepticism, I believe you now. I've looked at high-voltage high-current MOSFETs and diodes on digikey, and I've found that doing a high-current 750V power supply would be far more cost-effective with diodes. There are diodes that have only 0.5V drop at over a thousand amperes, and these are cheaper than the MOSFETs of equivalent current.

 

We use active rectification on our 100kW 750VDC:750VDC isolating converters - it is CT based
We use 1200V 8m-Ohm SiC mosfets, 2 in an half bridge, 3 of these in parallel to form half of each H ridge, just over 99% efficient at full power. We are currently building 750V : 1500V, non isolated, 250kW, these have active rectifiers using 1700V SiC mosfets ( we call them controlled diodes ) again current controlled.
In some 3kW telecom supplies the boost stage - to get unity power factor, use 600V SiC mosfets - i.e. no diodes, all active rectification - usually timing based, based on observed voltages, similarly for the LVDC side - timing used in the high speed uP ( PGA ) to handle the switching - estimation methods to avoid reverse current.

 

We use active rectification on our 100kW 750VDC:750VDC isolating converters - it is CT based
We use 1200V 8m-Ohm SiC mosfets, 2 in an half bridge, 3 of these in parallel to form half of each H ridge, just over 99% efficient at full power. We are currently building 750V : 1500V, non isolated, 250kW, these have active rectifiers using 1700V SiC mosfets ( we call them controlled diodes ) again current controlled.
In some 3kW telecom supplies the boost stage - to get unity power factor, use 600V SiC mosfets - i.e. no diodes, all active rectification - usually timing based, based on observed voltages, similarly for the LVDC side - timing used in the high speed uP ( PGA ) to handle the switching - estimation methods to avoid reverse current.

Interesting. I've looked at various MOSFETs on Digikey to see what components may be used for a 750V, high-current version of the power supply that I designed. It looked like it would be crazy expensive. I found that the diodes offer a similar voltage drop for a much lower cost than those SiC MOSFETs. For switching, the IGBTs appear much more cost-effective than MOSFETs at this voltage.

Given that the diodes are many times cheaper for similar voltage drop, why did you choose MOSFETs? Was the cost difference not that much from your supplier?

 

The 1200V 100A ( 200A ) Sic half bridges are about USD 150 each, here is a link to similar items

https://www.digikey.co.nz/en/produc...MI1uPQrIWnhAMV9YVLBR0-nQWREAQYASABEgLzZfD_BwE

obviously a good quality SiC diode is about 2V @ 100A, but 6 milli-ohms in parallel is < 600mV so 140 W saved for the time the " diode " is on.