Offline, Single stage, isolated Cuk PFC'd converter with no mains rectifier bridge

Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

Hello,

We are doing a 3kW battery charger (Vin = 230VAC, Vout = 250 – 420VDC)

We wish to use the modified Cuk converter to do this. This converter does not need a mains diode bridge before it, and does not need an isolating SMPS stage downstream of it. The Cuk converter does the whole job by itself and needs no other circuits.

Do you think this is possible? Dr Slobodan Cuk seems to think so (as in link below)


Why is the current in the L1 always discontinuous? Can this circuit not operate in Continuous conduction mode?

Also, the series capacitor, C1, suffers 15 Amps of ripple current, and has to be rated to 600V (and must not be of polar type like an electrolytic), and has to be at least 100uF. Why does this capacitor have to be of such big Faradic value?

The attached is a simulation of our work so far of the Cuk converter. –Also the schematic of our work so far.

Here is the link to the pioneer of this breakthrough power supply circuit, and the details of it….

https://www.power-mag.com/pdf/feature_pdf/1310569074_Teslaco_Feature_Layout_1.pdf

The link is 4 years old, howcome nobody is doing this now?….why bother with a mains diode bridge and a downstream isolating converter if its not needed.?

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

The attached LTspice simulation and schematic show an improved version of the special PFC circuit, but the voltage and/or current waveforms are too peaky in this converter…
Also, the transformer is suffering enormous voltage spikes. I believe that this needs a energy recycling snubber to be used with it….so has Dr Cuk not given us enough information to implement his fantastic converter with? Perhaps he is understandably only divulging the basic detail? –not enough to achieve implementation?

The circuit is a resonant circuit, and L3 is the resonant inductor, and C1 is the resonant capacitor. I thought that the LC resonant frequency would need to be roughly at the switching frequency but that gives poor performance….the power throughput cannot be made high enough that way.

If C1 is too low Faradic value then C1 has too high voltage across it. If L3 has too low value then the peak current is too high. If L3 is too high Henry-ic value, then the current throughput goes down and you can’t get the power throughput.

The transformer primary would need a heavy snubber, and this would be very dissipative, unless some kind of energy recycling snubber was implemented. The problem is, where would you recycle the energy too?….since there is no storage capacitor on the primary side, and recycling to the secondary needs another isolation transformer.

The circuit would obviously need a controller to control it…but there is no “primary ground” to use to ground such a controller to. Therefore, an output of an isolated SMPS would need to be used to “ground” the controller to. Then the sensed current and output voltage signals would need to be referred to this isolated ground. The necessary fast current sense in the power FET loop will be difficult, because this will have to be transferred across an isolation barrier to the control chip, and this will be difficult. A hall sensor would have to be used, but they have poor bandwidth which would mess things up.

Does anyone know how to set the component values for this converter? The only advantage of this modified cuk converter appears to be that you can dispense with the mains diode bridge. But in exchange, one needs to have a lot of extra housekeeping circuitry to manage the control of this converter. This seems unwise, since a mains rectifier bridge only cost about 30p, and only dissipate about 1.5% of the total power budget.

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

According to an article I read, electric utilities would love to see more of any type of power supply that pulls a more continuous current waveform.

It is because our typical familiar mains-to-DC adapter draws a current spike at every sine peak. (This describes the common full diode bridge and smoothing capacitor.) With so many of such power supplies in use, electric generators must provide a massive Ampere 'whump' onto the grid, at peak sine voltage. This is only at the peak and it is out of all proportion to Amperes drawn during all other parts of the AC cycle.

Because of that massive 'whump', power companies need to build larger everything (generators, wiring, relays, sensors, etc.), so that the grid waveform does not become distorted. Therefore they are likely to welcome any topology which reduces this massive 'whump', meaning it should pull during all parts of the AC cycle.

Clever use of inductors/ transformers might help somewhat in this regard. (Perhaps it could be a sepic/ Ćuk type. Why not?)

In any case manufacturers will make whatever type of power supply is easy to make, and which is profitable.

It would be interesting to find out whether Mr. Ćuk is entitled to royalties. Manufacturers sometimes wait until a patent expires, before implementing its technology.

But in exchange, one needs to have a lot of extra housekeeping circuitry to manage the control of this converter.

Click to expand...

This could limit its acceptance. There seems to be a rule for power supply topologies: Every advantage is offset by some disadvantage.

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

The transformer primary would need a heavy snubber, and this would be very dissipative, unless some kind of energy recycling snubber was implemented.

Click to expand...

Why? The circuit is designed as a resonant converter, the switch should see an almost real load impedance at the switching frequency without generating much overvoltage.

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

ya you're right, i remember when that came out. never heard anything since then till now

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

Why? The circuit is designed as a resonant converter, the switch should see an almost real load impedance at the switching frequency without generating much overvoltage.

Click to expand...

..Thanks, but the simulation shows Megavolt level overvoltage spikes on the switch and the transformer primary.
At the end of the day, the di/dt is very high when the switch turns off and the L4 inductive current suddenly finds that it has to flow through the transformer magnetising inductance....the di/dt in that at this instant causes massive overvoltage spikes..the simulation bears witness to this. I am not sure how this circuit is resonant....if one makes C1/L3 resonant frequency equal to the switching frequency then the circuit doesn't work well at all.....that's not like resonant behaviour.

As you know, there is a "rule of the simulator"......'if one can make a circuit work on the simulator, then it doesn't mean the real circuit will work, but if something doesn't work on the simulator, then it absolutely definitely wont work in real life'....this holds very true for SMPS's in LTspice

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

Yes, you are right. The switch voltage transient problem is brought up by the current-switching toplogy. It needs additional means to keep the voltage during switch-off within transistor safe operation area.

My consideration about the resonant circuit that loads the switch is only valid for the fundamental frequency.

I didn't notice before that the paper is authored by Slobodan Ćuk himself. But unlike usual Ćuk converter implementations, the circuit misses a path to cut the leakage/series inductor flyback voltage.

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

But unlike usual Ćuk converter implementations, the circuit misses a path to cut the leakage/series inductor flyback voltage.

Click to expand...

..yes, and the problem is that the current in the transformer, and indeed in the switch, is AC, in other words for 10ms the one side is positive and the other negative, and for the next 10ms its the other way round, etc etc...so making a snubber for this kind of circuit is going to be a challenge...the snubber would be more of an invention than the circuit itself I believe.

Its quite something when you think that there is no "ground" reference point, (because of the absence of a diode bridge). I'd be tempted to put a diode bridge in there and then just use the modified cuk as a single stage PFC from there.

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

Dr Cuk's ckt is interesting, there is some turn off overvlts if not done right, it been built to high powers (>1kW) to prove the concept, there are (can be) minimal switching losses, the only issue is a largish amount of output C is required to reduce o/p ripple (100/120Hz) as with any single stage implementation of a corrected psu (e.g. controlled flyback)
The ckt is patented for the next 16 years or so...so a licence required.

- - - Updated - - -

Just to clarify the above, linking the input choke to the Tx as shown in Dr Cuk's circuit substantially reduces any snubbing needed on the mosfet - we have verified this in practice...

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

Just to clarify the above, linking the input choke to the Tx as shown in Dr Cuk's circuit substantially reduces any snubbing needed on the mosfet - we have verified this in practice...

Click to expand...

the attached LTspice simulation couples the input choke to the transformer, but this does not stop the high voltage spikes, and also, when so coupled, the converter doesn't work as well as when they are uncoupled, the power throughput appears to get throttled back.


I know its only a simulation, but their does appear to be a golden rule of simulators..."if something works on the simulator, it doesn't mean the real circuit would work, however, if something doesn't work on the simulator, its virtually guaranteed that the real circuit will not work"

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

dot notation correct? do you model the C across the switch?

- - - Updated - - -

the non linear C across some mosfets (e.g. coolmos) seems to make something of a difference in real switching ckts...

Re: Is this this circuit going to revolutionise Switch Mode Power Supplies worldwide?

As attached, the ltspice of the cuk pfc with fet capacitance added, and with the inductor coupled either way, makes little difference, there are still huge spikes.

Hello,

Do you know what the current waveform looks like in the inductor L of figure 8 (page “20”) of the following. Its shown as a sine wave, but how can this be so?

It’s a single stage PFC’d, isolated Cuk converter, with no mains diode bridge.

https://www.power-mag.com/pdf/feature_pdf/1310569074_Teslaco_Feature_Layout_1.pdf

The attached is a LTspice simulation of it, but it cant be right because of the enormous circuit voltages, can these be reduced by coupling the input inductor into the transformer?

"There seems to be a rule for power supply topologies: Every advantage is offset by some disadvantage."

From time to time someone comes out with with a novel SMPS topology.

If the new topology does provide REAL WORLD advantages, which include reliability, cost and component availability, then the topology becomes widely accepted.
Otherwise it languishes.
Another factor is that design engineering community must be able to understand it. We as engineers may not want to accept it, but a topology which requires very complex loop compensation for instance, will not enjoy widespread popularity.