isolated DC/DC converter

Given base charge storage, if the driver rise/fall time is fast, the design could have shoot through.

ir2153x and a dual fet instead? but then I guess you might as well pick a standard converter ic

-Lasse

But only by one Vbe above or below.

...and given magnetizing energy, emitters will be pulled above and below the rails. Shottkys can discourage this. The driver still may not be close enough to the rails to prevent base current and conduction in the switch, CE, to the opposite rails.

Consider using a driver or controller intended for the function, with schottkys.

RL

I want to avoid custom magnetics, and I want it to be cheap, in case we build a lot of them. I'd like to trade off efficiency for switching noise, which this topology allows. There will be a 12-bit, 250 Ms/s ADC a couple of inches away.

And besides, I'm a circuit designer.

John

If 370kHz fixed is ok maybe these?

Around 50c in quantities. Problem is that your two transistors then become unemployed :-)

--
Regards, Joerg

http://www.analogconsultants.com/

I've done some sims of totem-pole fets, and it's a little tricky avoiding shoot-through. It can be done, but takes more parts. The complementary emitter follower has a lot of nice properties, like controllable edge rates and no snubbers.

John

the ir2153x has build in dead time, somewhere between 0.6 and 1.2us depending on which one

-Lasse

P.S.: Note that the pull-down FET is wimpy, 10ohms. So if you decide to lose the external transistors make sure this is ok for the load.

--
Regards, Joerg

http://www.analogconsultants.com/

There is nothing wrong with buying a controller chip. You really don't have to roll your own.

But I need isolation. 5-watt dc/dc bricks are expensive and none give me the output I need.

John

I meant to drive either the transistors with this or drive directly into the cap and primary. You'll have the isolation. In other words, use it as a square wave generator instead of a buck. If you don't want to use FB you might be able to ditch the compensation node stuff, reducing parts count. One of those unorthodox or "wacky use" cases. I do this a lot with switcher chips.

You didn't mention the power needed. Keep in mind that the low side device in there is around 10ohms Rdson so for heavy loads you might have to keep your NPN/PNP transistors.

I am sure there are integrated sync buck chips with more power but not at the 50c price level.

--
Regards, Joerg

http://www.analogconsultants.com/

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

How much current? I have driven a Wurth(CS?) transformer with a TL494 ,

0.17c a piece. I can get about 50ma out of it at 100khz in a push pull config. It's one of those T1/T3 communication transformers.

The 494 is good for 40V. Maybe you can totem pole the outputs like the

3525. Or use a 3525 they are only 0.60c

Cheers

You might want to check IR2085S

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

Gee, I'm kind of disappointed. Doesn't really drive the transformer all that well (as Legg mentioned), no way to adjust voltage, and still a thousand transistors (well, maybe a hundred in an average op-amp / comparator / driver, and you can't see them, but still).

I already solved this problem for cheaper than the bricks and with higher efficiency. Finding the magnetics is the hardest part, but I've found some very cheap and common parts that will easily deliver a few watts with moderate (open loop) regulation, with minimal capacitance, and they hi-pot at a whopping 14kV, not bad for a teeny PC mount transformer rated for

2.5kV peak. I'd provide my circuit, but it counts as proprietary information by now.

The only downside is the limited ratios. Want 12V output from a 5V circuit? Too bad, you need a boost to get the primary voltage up to ~13V. On the upside, at least you can put the boost (or buck or whatever you end up using) inside the feedback loop (if you regulate the whole thing).

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

"John Larkin"  wrote in 
message news:dk24b7hs2qoimekeurpq3v0s0mnnoaode1@4ax.com...
> This
>
> http://www.panoramio.com/photo/61564837
>
> might work for powering a thing I'm doing. I can use a standard cheap
> Coiltronix dual-coil inductor as the transformer. The complementary
> emitter followers will have no shoot-through and can have controllably
> slow switching edges, since they will just follow the base drive.
>
> Anybody got ideas for the base driver device? Ideally it would be
> self-oscillating, set with some R-C; have a moderate slew rate; swing
> to the rails. I'm thinking roughly 150 KHz maybe, a few watts output.
>
> Maybe an LM8261 opamp? I'd have to see if it winds up when it rails.
> It might not.
>
> Some sort of fet gate driver would be OK, but few go to 24 volts.
>
> Something discrete maybe, like a 2N7002 to 24-, and maybe a
> bootstrapped pullup?
>
>
> John
>
>
>
>

All I want is something vaguely in the 24 volt p-p range, to get across the isolation requirement. The small box I'm building has 9 power supply rails:

+75 +6 +5 +4 +3.3 +2.5 +1.2

-5

-6

all of which I can derive from something like +10 and -10. Seems like half my life these days is designing power supplies.

Don't need to

Transistors inside an opamp are cheap!

Coiltronics makes cheap 1:1 transformers, which will work here. Custom magnetics is a real nuisance.

I'd provide my circuit, but it counts as proprietary

Oh.

John

Wow, an eschatological chipmaker!

Cheers,

Phil Hobbs

Self Oscillating? Make a colpits oscillator, no drivers needed.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

Except chips don't have an afterlife when that lower FET blows. They can't repent :-)

--
Regards, Joerg

http://www.analogconsultants.com/