isolated DC/DC converter

Check out the LT3439 (slew rate limited). I found it worked very well as a DC transformer (its datasheet shows you how) with an off-the-shelf transformer. Couple of points -

- wrap a shorting turn round the transformer to reduce radiated flux

- use a transformer with about 500-1000uH primary inductance to avoid wasting power / heating the IC unecessarily. If you have one of those standard 6 coil types, you can put 2 coils in series for 4 times the L (not twice, 'cos they are on the same core)

- you can get by with 10uF ceramics for the output, it switches around

- follow up with C-multipliers on the output for an unregulated but very quiet rail.

Nemo

But I am sure John won't want to spend the two bux in the BOM :-)

if it sits next to a 12bit 250MHz adc I'm not sure 1-2$ will make such a big difference

-Lasse

Actually, I ordered some IRS2153D's, a very simlar part. Maybe if I use biggish gate resistors and keep the frequency down I can get tolerable efficiency and low noise. It may need a snubber, not sure. I'll build one on Monday.

My little emitter follower thing seems to work, but it's not very efficient.

A couple of bux of parts cost is not a big deal here. Getting it to work the first time is.

John

... just enough to cause the transistor to conduct the magnetising current to the opposing rail - clamping the winding in the lossiest way possible, Imag x VCC losses being absorbed by the bipolar transistor that is supposed to be 'off'.

RL

just wait till jf hears you are playing with what is basically a 555 and a halfbridge driver :P

-Lasse

But yesterday you wrote that you wanted it cheap in case you build lots of them :-)

Reminds me of a time when I was designing something for the guys where things must have cast-iron robustness. I sat there in agony and a manager asked me whether I had problems with my design. "Not technical ones but these multipliers are over $6, a piece". A whole group of people around me had listened in. They and the manager burst into thundering laughter ...

So, getting a self oscillating thing to work seems too complicated ?

Use a PIC.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

It should be pretty easy to throw some randomness into the frequency with a PIC for a spread-spectrum converter.

Best regards, Spehro Pefhany

The magnetizing current tends to turn the "wrong" transistor on, but the load current mitigates that. A snubber helps, too. It's not hard to make all the currents play nice.

I did find a weird asymmetry that turned out to be a feature of LT Spice. If you specify a pulse generator of

Ton 5 us Tr 1 us Tf 1 us Tper 10 us

you get an asymmetric waveform. You need Ton of 4u to get symmetry.

John

Within reason. I don't want to buy a 5-watt potted dc/dc converter for $30, or custom magnetics for $10. The IR thing is under $1 in quantity, and doesn't need a lot of external parts. This one even includes the bootstrap diode. I still want to use a stock dual-winding inductor as the transformer. Our per-part kitting/placement cost is high, so we take that into account too. Our customer, probably wildly optimistic, estimates maybe 800 boxes a year, so it's worth spending a day considering alternate ways to do this. Besides, playing with circuits is fun.

Our customer just decided to use a 600 watt single-mode fiber laser to

John

Actually, that does look complicated.

Where would the PIC get its power from?!

Just getting a square wave clock isn't hard. The power drivers are what I'm playing with. I want a quiet, efficient, cheap, low-risk

John

Unfortunately it guzzles 5mA Iq though. It's a self-zenered device so you'll probably be burning 200mW by it just sitting there.

That's why I suggested the A&Q part. It even has the timing RC integrated (or probably some sort of ring oscillator). I try to get companies away from kitting, never believed in it and always saw it as a waste of time. At Endosonics which was a fairly big company with around

I sometimes have that as well. Major project, 6-7 figures, and then someone occasionally worries about an additional travel expenses to the tune of a few hundred bucks.

I did a box that used a bunch of the LTM8023 bricks. I built a TinyLogic schmitt oscillator and used the triangle to FM all the switchers a little. It spread the switcher noise spectrum beautifully.

John

Uh, we're talking 10 figures here.

John

Oh, sounds like courtesy of the taxpayer. As in "free" :-)

John Larkin a écrit :

Arghhhh, why did I forgot that for my last project? (PCB has just been sent to fab...)

Thanks for the reminder.

On a sunny day (Sat, 05 Nov 2011 10:11:37 -0700) it happened John Larkin wrote in :

I do not get it, you can use that same circuit you recommended and I used for the PMT HV supply. Just one transistor, 1 resistor, 1 transformer, 3 caps, and some diodes. I have used it in the past with the feedback in the base to power many things, for example my old EPROM programmer, from 5 V.

What exactly did you do? Dither the voltage?

If you were using a bunch, wouldn't multi-phase synchronous operation be better?