How's This Smps Patent?

It's a resonant converter These have been around for some time but his transformer configuration seems unique. I'm not sure if this design brings much to the table or not. I'd have to see a specific design example, look into the transformer specifics, circulating currents, capacitor specs, regulation, etc. and compare it with a more traditional design in price and performance.

By way of comparison he has included a Cuk converter in the prior art section. If you read their patents and papers, the Cuk is the best thing since sex, yet they were never popular. The peculiar, hard to make magnetics and other engineering nastys make them production nightmares. The patent circuits have some of the same look to them. So, who knows?

It looks *exactly* like one of Ed Herberts symmetric converters, but with an AC switch.

Cheers Terry

Yup.. somehow the classic Cuk has taken the back seat. (Based on the small amount of research I've done.) Maybe everybody is just dodging the patents or the design is just too bloody complicated. Dunno... Over and over again I see the classic Cuk config...with no feedback control. It's like: 'This is a Cuk...good luck stabilizing it!'

I've made some working classic Cuk supplies and yup the magnetics are a pita. For example, the classic non-isolated non-resonant Cuk design may need magnetics with a high core saturation due to unipolar operation (1st quadrant of BH plot).

However, the above patent (6343021) claims no RHP zero and no unipolar operation of the magnetics. It's like a super Cuk design..

D from BC British Columbia Canada

Late at night, by candle light, "Bob Eld" penned this immortal opus:

Where's genome off to? ISTR he's pretty knowledgeable about these thingies.

- YD.

--
Remove HAT if replying by mail.

Dunno... I suspect there's a few retired designers spending their last days on SED.

'...and the best switchmode supply topology considering all compromises and current component technology is a;liddufga'dodudijjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj... .'

(Collapses on keyboard due to heart attack) :P

D from BC British Columbia Canada

It is a Cuk converter topology. The capacitors are going to be an interesting issue in the real design.

You can drive the MOSFETs through gate drive transformers. If it is a narrow range supply just hooking the tranformer to the gate is good enough.

For wide range:

D1 ---->!---- Q2 A ! ! !!------ -----+-- -+---!! !!( s! ! !d !!--+--- !!( ------- ! !!( Q1 ------- ! !!( B ! ! -----------+------------

Q2 is the big N MOSFET or IGBT

Q1 is a small N MOSFET. When B goes positive WRT to A, it turns on and discharges the gate of Q2

D1 is a schottky. It prevents current in the substrate diode of Q1 so you don't have recovery time issues with it.

Although the primary switch and coupling are arranged like a Cuk and have the same phase relationship in isolation transformer and coupling inductor (regardless of which serves what function), it can't perform the same conversion cycles: C3 is not in series with the rectifier and load; it can only be sized for a resonant or quasiresonant function without actually inhibiting power transfer.

In the Cuk, the high-frequency voltage across the coupling capacitors is not intended to reverse, while passing full load current. In fact it maintains an average value equal to the sourc/output voltage respectively (or their difference, in the non-isolated case).

Converters that impress the fw bridge rectification function on the actual HF switch, in such a crude manner, surrender the basic efficiency provided by low frequency rectifiers or SCRs (and in this case the Cuk single rectifier output efficiency also), and impose low frequency filtering requirements on the low voltage load filter.

You can do a lot of silly things at low power levels, but there's no reason for them to pay, unless part REDUCTION results, or part stress is signifigantly reduced.

RL

Yes, you are right. I have my doubts about the usefulness of the circuit now that I look more closly.

C2 is nearly across the incoming mains voltage from an AC point of view. and the same with C3.

The author describes it as a resonant converter which could be true is

1A and 1C had differing numbers of turns.

Neato.. That's less parts than an isolated supply (linear or smps[1]) + a digital isolator (opto or other) + a mosfet driver.

I'm not sure what you mean by narrow and wide range. Do you mean small dV or large dV upon the gate Q2?

I suppose a mosfet driver can be used to drive the transformer primary.

[1]Kinda goofy to make a smps to power an smps.

Reminds of a loop I once got into.. I wanted to make a good voltage source. I figured I could make one with a good current source.. But, to make a good current source I needed a good voltage source.. :P

D from BC British Columbia Canada

Ahh..If I understand.. The high frequency switching in the bridge e (with the power mosfet) will make those diodes more toasty compared to low f rectifying of the line as traditionally done with offline converters..

D from BC British Columbia Canada

Near 50-50 duty cycle, you can just use a transformer. If you are making a wide-range supply (DC in from 90 to 300V or Vout from 0 to

100V) you need to vary the duty cycle far from 50-50. In that case the little trick with the MOSFET helps.

You only need a DC blocking capacitor and a small resistance in series.

I have done it. A very small switcher can make the supply to run a floating switcher chip.

Such loops sometimes lead to good ideas. More often not but sometimes.

Consider: I need a stable current to make a better voltage reference. I can make a stable current if I have a good reference.

This leads to the bootstrapped reference.

Parts optimized for speed trade off in their Vf.

Anything in resonant branches experience higher peak to average stress ratios.

Output low voltage rms is ^^. Two series'd output rectifiers in a circuit path that previously got by with ~1/2 of a rectifier is suicidal, when efficiency is an issue.

RL