Very high gain transformerless boost converter - is it possible

Possible, yes. It's possible to shoot onesself to the planet Mars. That doesn't mean it's a good idea, or economically sound. :-)

The best is a tapped inductor, and second best is ordinary flyback (isolated at the transformer level, but you don't need to isolate it, it can be tied common ground).

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website:

Nomadic Electron typed

I would not use the word 'gain' here, you meant efficiency I guess.

It's simple (tm)

350 / 12 = 29.166667 So we need 2 x 29 makes 58 12V batteries, oh why not use 11.1 V lipos then we need 350 / 11.1 = 31.531532 say 2 x 31 makes 62 3 cell lipos.

Use a bunch of antique relays, and put half the lipos parallel to your 12V battery, with some charging control in series with each lipo. The other 31 lipos switched in series to the 350V output

Once in a while when output lipos go low, switch the first 32 to the output (in series) and the second 31 to the battery (in parallel) for charging.

Have a small electrolytic >350 V cap on the output to cover the relay switch-over time. make sure you use break before make relays,

This also should work with capacitors and MOSFET switches in place of LIPOs. It is a well known method but forgot what it is called, switched capacitor converter perhaps? In the case of caps and MOSFETS you need to switch a bit more often.

All depends on what you have in the junk box... Note the isolation requirement for the MOSFET drivers.

;-)

A switched mode boost converter would probably work and doesn't need a transformer but does need an inductor:

Why do you want to avoid transformers? If you are concerned about the size and weight, a high frequency flyback regulator would be a good option. It does require a transformer but it could be fairly small compared to the half brick sized things designed to work at 60Hz.

If you want to avoid magnetics altogether then it's going to be more difficult. A Cockroft-Walton voltage multiplier may be an option but I don't know what the efficiently would be:

Obviously you would need to generate an AC voltage to drive it but that can easily be done.

What are you actually trying to do?

It's possible yes, but the choke would run at a low duty cycle. So to get 85% you'd need a big choke to get low copper loss at short high current spikes.

NT

Switched capacitor converters do exist, but probably not at this power level. There are also some experiments with piezoelectric transformers, but it's still a niche:

But what's the reason not to use perhaps the most robust electronic component in the Universe? Coilophobia?

Best regards, Piotr

It's also probably possible to parallel up the outputs of some boost converters you can buy off the shelf, e.g. I have a bunch of these modules which claim to be capable of 40 watts (not had a chance to test actual full-load capability but the ones I received seem well-built so real-world performance may be fairly close to what is claimed)

Two of them seem to share load OK with no further effort but to get the output power levels as he wants you'd need like four or five, and my intuition is they'll wreck each other trying to naively load-share without some kind of closed-loop to make sure they share the load current equally.

Which you could probably do with current sensing on the input side to equalize the power for each, there are a number of schemes online but...at that point your "solution" is already rube goldberg and total amount of magnetics much larger and heavier and efficiency will deffo not be anywhere near 85% so you start to wonder once again what the point of these "make a tasty sandwich without bread" exercises are.

As others have said, the low duty cycle of an inductor-only boost converter, under 4%, spells big trouble for obtaining high efficiency. However, there's an easy trick that can get around the 28x step-up problem: reducing it to two sets of sqrt 28 = 6x step-up. A two-stage boost converter, but with one controller, driving two MOSFETs, inductors, diodes and capacitors. Use separate resistors for each gate.

The first MOSFET & inductor creates an intermeduate voltage. It runs at higher currents, but at lower voltages.

--
 Thanks, 
    - Win

Get a dynamotor or HV photo flash battery

That's cool. Thanks.

Combine that with my center-tapped inductor trick maybe to get another

4x.

--

John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Oh BTW, this is relevant:

just a toy I made. Toy in that it does what a $30 automotive inverter box does. The left board is a 12V 300mA (isolated) and 170V 1A supply; middle is controller, and right is inverter.

I wouldn't suggest doing anything other than a push-pull forward converter. A current sourcing inverter (constant current buck supplying PP chopper) is better especially at the high voltage output (which will suffer from excessive winding capacitance).

Flyback is bad because of the large peak currents required, and single stage boost is right out because of the fuckoff enormous inductor and semiconductors (and peak currents) required.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/ 

"Nomadic Electron"  wrote in message  
news:8eeae69c-e937-4471-9340-130584cdbd61@googlegroups.com... 
> Is it possible to build a 130-150 Watt, 12V input (for example, car  
> cigarette lighter) to 350V output transformerless boost converter. The  
> efficiency should be at least 85%. Assume a resistor load only. 
> 
> If it is not possible, Would there be any other possible tranformerless  
> solutions?

Wow thanks very much for all the replies. I will think about what to do bec ause there are so many different ideas here. Right now i will probably try to do a simulation of the cascaded boost converter idea of winfield hill. M aybe ill also try simulations using a transformer if it is the best way to get the highest efficiency at this power level. Gotta try to understand wha t some of these ideas are about firstly.

Nomadic Electron

And do not forget the 12V DC motor driving a 350V DC generator solution :-)

100 relays switching 100 Li-ion cells from parallel to series.

--
 Thanks, 
    - Win

Dynamotor. For regulation, control the field of the generator. I think that's an amplidyne or something. They used them for things like aiming the guns on battleships.

Overkill. Two banks of 30 relays each should do it.

--

John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

The best bet would be Dickson triplers stacked.

The first one takes it to 36, the next one takes it to about 100, the next one takes it to just about 300. More like 290 I would say depending on the parameters you give the convertors. Can you get by on 290 ? If not then sta ck another one on top but fed from the original 12 or maybe just a doubler off the 36. Caps will not be cheap because they need low ESR and high volta ge rating.

I would not attempt to multiply it all at once but am at a loss to explain exactly why. It just seems the current in the choppers would be too high fo r good efficiency. Though more junctions means more drop, they are not as s ignificant at higher voltages.

For each stage of the convertor, you are talking 4 good power transistors,

4 very good capacitors and 6 low loss diodes, Shottkies in the first stage at least. In the first stage you might try to find something like the 1N609 5, but those do lose at higher currents.

One nice thing about using Dicksons is that the main output filter has very little to do because it is fed with 2 polarities of a square wave, and I m ean a real square wave not some rounded spiked and "rung" out waveform.

However the cost is ridiculous compared to just biting the bullet and using a transformer. Also with a transformer it can have isolated output.

It is so much easier and cheaper just to go to Digikey and find a transform er with a ~1:20 ratio, feed it with a totem pole at its resonant frequency and use a doubler on the output. Use a chopper before the main filter for r egulation, or synchronous rectification if you can handle designing it. (mo re efficient)

At 150 watts that will be a fairly good size transformer and will cost mone y, but still cheaper than the stacked Dicksons. And much less designing, be cause with the Dicksons each stage has to be different. Different outputs, different diodes and different caps, you must choose them all. With this ma ny components you don't want too much overkill. With a transformer you desi gn ONE circuit and call it a a day.

Your call. I like to avoid transformers because they can become unavailable . But sometimes you can't always get what you want.

I just checked Digikey and no suitable transformer is readily available. Wi nd your own. At the right frequency you only need a few turns in the primar y. So 4 turns there and you want 80 turns on the secondary. For a one off i t may be alot of work but handleable. For production forget it, the Chinese probably have something.

Do you have a few good references for the Dickson voltage multipliers? I only found two recent locked Academic papers (used my Harvard access to get them). No mention of why called Dickson, no paper by him, etc. Looks like Cockcroft-Walton.

--
 Thanks, 
    - Win

;)

Wow. I invented the thing (not first though) and can't find the Spice file right now. I guess I'll have to make another one.

I actually posted a hand drawing of it back when I had hosting. Someone replied with a Spice file. I remember it.

I'll get to it.

That's not what I designed.