Multidevice interleaved bidirectional boost converter

[orcuns91]

Hi everybody,
I'm an MSc student and I'm writing my thesis subject about dc dc converters of electric vehicles. I having a problem because I couldn't get enough help from my advisor in this pandemic period. I'm trying to evaluate the performance of SiC mosfets in multidevice interleaved bidirectional boost converters instead of Silicon Mosfets. The converter must be bidirectional because of increasing efficieny of EVs. However, I could not find an article on analysis of the circuit expect for the unidirectional one. (https://ieeexplore.ieee.org/document/7054946) Normally I know I should replace diodes with mosfets, but I'm confused about gate signals because of the circuit's structure.
In conculusion,
1. Do you have any article about the analysis of multidevice interleaved bidirectional boost converters?
2. How can I convert the unidirectional MBIC converter to bidirectional one?

 

[BradtheRad]

Bidirectional suggests stepping up in one direction (boost converter)...
then step down (buck converter?) in the other direction. Is that your plan?

It's one challenge to bias transistors at the right times so they conduct in one direction. Mosfets have the body diode which conducts in the opposite direction. It's another challenge for you to arrange the circuit correctly, so current flows the way it's supposed to.

There is such a thing as installing two transistors anti-parallel, so you can make one conduct in one direction, or make the other conduct in the other direction.


Here is a simple boost converter. It's synchronous type as you seem to prefer (not switched by a diode). Only one clock signal is sufficient to control it, because it turns on the N-device and P-device at different halves of the cycle. To make two interleaved converters, add a second identical circuit with an identical gate signal 180 degrees later than the first.

 

[orcuns91]

Bidirectional suggests stepping up in one direction (boost converter)...
then step down (buck converter?) in the other direction. Is that your plan?

It's one challenge to bias transistors at the right times so they conduct in one direction. Mosfets have the body diode which conducts in the opposite direction. It's another challenge for you to arrange the circuit correctly, so current flows the way it's supposed to.

There is such a thing as installing two transistors anti-parallel, so you can make one conduct in one direction, or make the other conduct in the other direction.


Here is a simple boost converter. It's synchronous type as you seem to prefer (not switched by a diode). Only one clock signal is sufficient to control it, because it turns on the N-device and P-device at different halves of the cycle. To make two interleaved converters, add a second identical circuit with an identical gate signal 180 degrees later than the first.

View attachment 163573

Thank you for your reply. I need a dc-dc converter which works as a boost converter when EV is in traction mode and works as a buck converter when EV is in regeneriton mode ( gain some energy when braking ). So, I chose the topology of multidevice interleaved bidirectional dc dc converter. My purpose is evaluating the perfonmance of silicon carbide devices by using matlab simulink. For this reason, I'm trying to draw the circuit in the program. However, I don't know how the gate signals should be.

 

[FvM]

Consider synchronous switching buck/boost half bridge (either high or low side is conducting with 150-200 ns dead time during commutation). PWM duty cycle = V1/V2. And evenly spaced phase shift for interleaved paths, e.g. 180 degree for shown two phase topology.

 

[zenerbjt]

The power stage of a synchronous buck converter..is exactly the same as a synchronous boost converter....you will see this if you draw it out....depending on which end you have as your input.

Just have your current sense in place for each topology....eg could be a resistor in the source of the low fet for a boost...and a CST upstream of the hi side fet in the sync buck.

SiC Mosfets are good because the internal diode is schottky and has no reverse recovery.

As you know, non transformer isolated topologies can suffer hideous reverse recovery...but Sic fets solve this well...but they can be pricey and often have higher rdson.

Your supervisor has asked you to do it bidirectional…of course…you could just do two separate unidirectional ones, and changeover switch them…but presumably it would take up a bit more room. EV chargers are very often not on the car itself…because its unnecessary extra weight….so your bidirec converter will be stationary, in the garage based charger ….so why does it need to be small?....i’d say the biggest thing about EV chargers is that they are built to last for years and years…to cut electronics waste…so I hope your supervisor has got you to do it without wear-out components such as electrolytic caps, fans and optocouplers….if not, then ask him why not……if you are using electrolytic caps…then are you putting them in a separate cool chamber?…that can be disconnected when the electro’s fail…so that the whole charger doesn’t get thrown away just because the electro caps have failed……this is more important than doing bidirectional.

 

[BradtheRad]

My purpose is evaluating the performance of silicon carbide devices by using matlab simulink

Start simple with a few components. Get the basic circuit working, then add more components.

At some point you'll replace each plain mosfet with a SiC type. If you're lucky Matlab has a shortcut to make it easier.

how the gate signals should be.

Do you plan to write code for a programmable device? In any case here's one method to generate dual staggered pulses. Change duty cycle for all outputs simultaneously via one potentiometer adjustment.

The pot was moved from max to min during the run.

OR gates provide overlapping duty cycles (51-99%). Use NOR gates to obtain short duty cycles (1-49%).