S.E.D is disappearing

ESR & ESL will complicate that of course.

NT

That usually turns up as you go along.

(Anyone noticed that the more "well specified" the problem, the less useful the response. I think it starts to look more like hard work and you get responses like "if you want me to do your homework you'll have to pay me".)

Maybe, or a new memory. But, from memory, krw hardly starts *any* threads, let alone OT ones. "Cursitor Doom" does little else (and then can't bear to see the responses of course).

--

John Devereux

If he barely stays at max current with no alteration of output voltage, he's definitely going exceed the average current limit with a boost.

That works.

I'm skeptical of the LED terminal voltage. It's unlikely to be 6V @ 10A.

Cheers, James Arthur

To break a boost converter's direct input-to-ouput path, just replace the boost's inductor with a dual-winding choke and use the second winding in flyback for output. I did that on my last supply, for robust short-circuit tolerance.

Cheers, James Arthur

Hah! I have no idea what happens when you push 10A through an led. (for 200 uS?.. the time constant of a cubic inch of stuff is like a second, or more...?)

George H.

piglet, Re: the zany alternative. I tired to draw this up tonight. I had the capacitor in the wrong place. Now I've drawn a parallel combo of an LED, inductor and cap. with a switch in front, fed by a 250 mA current source,

7V compliance. (I'm not sure the 7V matter's so much but you do have to know when to throw the switch. ) Is that close to what you were thinking?

George H.

Hmm, well when I don't specify enough I get a lot of spurious responses. Which may be interesting to others, but never seem to help me so much, at least in the short term.

So I try to set parameters... but certainly don't want to discourage different thinking, I often take a wrong path in search of a solution. There's a need for lotsa crazy idea's at the beginning of a project. ('cause the first one might not work so well.)

George H.

If you charge a discharged capacitor from a power supply, through an inductor and a diode, it charges to twice the supply voltage, at 100% efficiency. That must be useful for something.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

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

I think when I was at the FEL (Vanderbilt) we used that circuit (multiple times?) to boost the voltage that eventually fed the klystron.. which made microwaves for micro seconds (tens-hundreds) There where some other LC bits in the power chain.

The diodes were in these ~2' long black square sticks.

George H.

You can get almost any voltage you want if you do this...

...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 

             I'm looking for work... see my website. 

Thinking outside the box...producing elegant & economic solutions.

Actually I don't know either - what issues do electrolytics have with high dI/dt?

--

John Devereux

[Snip!]

Electrolytics are reputed to have high parasitic inductance, which may upset your circuit. For what it's worth, I measured ballpark

100nH for axial wired 100uF units, and 4nH for radial wired ones with 5mm lead spacing. That last value is actually very good. A Ta cap with the same lead spacing also has that inductance.

Jeroen Belleman

Hmm, I thought he was implying a specific part failure mode (like I think SCRs have). Yes one would specify a low impedance part, that was already assumed I think.

Say the circuit switched the 10A in 1us dI/dt = 1e7 A/s.

I suggested a radial part

but even with your 100nH

V = L.dI/dt = 100n . 1e7 = 1V.

which would be fine AFAICT.

With the more likely 4nH this becomes 40mV.

--

John Devereux

Nah. This is a place where inductors work better than caps, you can ramp up the current slowly and with the opening of a switch, the whole stored energy zaps into that low-impedance load from a suitable secondary winding. Needs a custom inductor, though, because of the time constant spec.

Blumlein pulser might do, it's about right for 30 feet of Cat6 cable... that's quicker to pull out of the storeroom than a custom inductor.

There's ESR, ESL and ripple current rating. Then there's the fact that lytics on high di/dt are more prone to lose capacity & fail, so if you want reliable be conservative all round.

NT

Hi George, this sketch shows what I meant:

piglet

High *current*, sure (as in the rms ripple current spec). That is why I suggested the Al. Poly. Which are good for low ESR and ESL too. But is there some capacitor failure mode caused by high *rate of change* of current specifically? I was not aware of one.

--

John Devereux

More like 30 kilometers.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics