Designing an active rectifier

[Varde1234]

Hi, I was laying out the PCB for my active rectifier but the design is not finished yet. Currently the circuit receives a 1.12kV AC and rectifies it to DC. It uses a TEA2208T/1J. I want my board to output a 48 VDC signal instead. Do you think I should adjust the windings of the transformer. I will have my multisim circuit and my board. Please give me advice what to do with the connectors because its my 2nd board.

How can I cheaply make a board that's essentially what is on the NXP TEA2208T/1J website? The image is also attached... I want to use these super GaNFETs from transphom.

Maybe I should have a buck-convert to step down the DC-DC?

[mariush]

I would be very concerned with having 1kW+ through just those 4 through holes. I'd double the holes, at least for the DC positive and negative.

I also would try to flip a couple of those mosfets so that the negative and positive traces on the board would be much wider and I'd consider arranging them so you could add a small heatsink on top, and maybe have two holes for an eventual heatsink clip. Kinda like chipset heatsinks.

[Varde1234]

Thank you for responding. I can wind down the receiving coil so theres only like 160VDC going through but the power efficiency is dog water. It shows that it consumed 5kW on the AC supplying board and its barely over 1kw on the load resistor. I also notice in the recommended application data sheet I see resistors for the TEA2208. What are these for? I will attach the datasheet for it.

[Kleinstein]

In a full wave bridge retifier with a resistive or similar load the diodes need to withstand the peak voltage. With a capacitive load (e.g. with a flilter capacitor and little other load) the diodes even see the peak to peak voltage.  Unless really desparate at the rather high voltages just good normal diodes and a passive rectifier may be the more obvious choice.

[jonpaul]

 You build a 1.2 kV >> lower voltage  SMPS ? Complete specs line/load, etc?

please post entire system schematic. The snips and ref to an app note are insufficient.

Why your fixation on GaN and Class E?   

j

[Varde1234]

Dear Jon Paul,

My fixation with the class E converter is because I am driving the left side with a class E amplifier. I thought the topology matched for the project. I just was trying to make an efficient WPT system that wasn't already traditionally used. I also saw resources for the project a couple months ago and it got me interested. Do you have a discord we could chat about this? Varde1234#0272

[jonpaul]

Rebonjour

Terminologie.....
How do you define "active rectifier", to me it's a mains frequency AC rectifier with transistors instead of  diodes for higher effecincy in very low V output supplies, where a diode drop 0.6 Vf causes significant losses.
Perhaps your terminology differs?

Class  E....

Our old friend and colleague Dr Nathan SOKAL (RIP) invented Class E and patented it decades ago.
He received an IEEE award.

The class E topologies have advantages in certain applications, and depending upon the switch frequency either  BJT FET GaN or SiC devices can work.  But GaN is relatively expensive and may have availability issues compared to the others.

If your objective is to  learn about Class E or GaN, bravo!

But for practical designs, cost, availability, magnetics, efficiency dictate a logical engineering approach.

We advocate considering all topologies, devices types, switch frequency etc. and choosing the right compromise topologies, switching devices and frequency.

Just the ramblings of an old retired EE

Jon

[Varde1234]

Dear Jon,

I have seen two instances of an active rectifier. The first one is the NXP sample board for the TEA2208T, the controller is just one IC to hook up with +14V. The second was the attached article with the professors building a feedback loop system with a microcontroller ("MainArticle.pdf"). That one was a bit over my head, so that is why choose first option for synchronizing the FETs.

I went on mouser to order my the GaN parts and ended up costing $33 for each of them. After contacting their marketing VP on their website. They were able to send me 5 samples for free. This would save most of the cost of the project.

I understand if you would rather chat on this forum about this topic instead.

The class E amplifier was desirable for me because I had another project trying to get a spreadsheet for the values given a desired Pout, Coss, and RL.

Thanks,
-Chris .Y

[dietert1]

The NXP part is for mains rectifier, followed by active PFC circuit. Datasheet 8.1 says that it must be followed by an active boost PFC circuit, not by a large capacitor. This is also clear from the timing diagram figure 4. There is no turn off depending on DC output, but a 50/50 scheme.

What i remember from building an active rectifier with DC output is that a solution with a second AC input for timing control is much safer and one still one needs to compare voltages between AC input and DC output. Especially when using low Rdson mosfets and at reduced load, one needs voltage regulation of each mosfet in order to avoid "glueing" input and output together. If you can no longer detect current direction in a mosfet, there will be additional losses due to late turn off.

Regards, Dieter

[mariush]

The Youtube channel Great Scott did a video about active rectifiers where he included this TEA2208 and mentioned that recommendation, to have some active pfc circuit after it as it's not designed to work with a lot of bulk capacitance connected directly to it




SDG Electronics also did a video about ideal diodes / active rectifiers, LT4320 is presented in the video but the downside of that chip is price and fact that it only turns on at around 9v if my memory is correct, so below that threshold you get no output...