200V @ 100mA from 3.2V

So, I've gotten one more change from a customer, that makes the current circuit into more of a creative pile of revisions than something that's actually well thought out.

Before I go and exert myself to design something, does anyone have any suggestions for a reliable, efficient way to get 100mA at 200V from

3.2V? I'd like to stick with off-the-shelf parts, but I'm willing to entertain the notion of a custom coil if that's going to make me happiest.

TIA.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott
Loading thread data ...

I'd consider using a flyback transformer in reverse. Pick one that is geared towards 3.3V or 5V supplies with high current, 10 amps or so. You can use a regular flyback driver chip plus external FET but you'd have to generate the start voltage with a charge pump or the like. After it started you can supply this IC from the helper winding that almost all flyback transformers have. So the charge pump can be very wimpy, charging up an electrolytic that stores enough energy to "crank it up".

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

I considered doing just that. I may consider it even harder now.

--
www.wescottdesign.com
Reply to
Tim Wescott

That's 20 watts from a 3.2 volt source, or 9.4 amps, assuming 80% efficiency. And that assumes the source accurately maintains its 3.2 volt level with a 10A load.

Despite the standard suggestion of a high-turns-ratio flyback-transformer scheme, I'd suggest that you use two stages, each with a modest 8x step-up ratio. Or two 10x stages (excess) with tracking PWM controlling.

10x xfmrs and 20 watt converters are not difficult.
--
 Thanks, 
    - Win
Reply to
Winfield Hill

Why kind of power source can even supply 10A at 3.2V?

Reply to
edward.ming.lee

Strange input voltage, lithium battery ?

That is about 8-10 A at the input side. It would require quite a big battery to run for more than a few minutes and even then you had to handle the battery voltage variations.

Reply to
upsidedown

lithum polymer battery

this gizmo sounds about right for lightsaber based round a flourescent tube, although 20W would probably be too bright.

--
  \_(?)_
Reply to
Jasen Betts

Photomultiplier supplies aren't usually two-stage. A big turns ratio does t end to turn into a low resonant frequency for the transformer, but you can help that a bit by breaking the secondary windings into banks.

The Baxandall Class-D oscillator seems to have been invented to cope with l ow resonant frequency transformers in photomultiplier supplies. Baxandall's paper doesn't mention it, but one of my bosses who had worked for Baxandal l at the time said that it was the motivation.

formatting link

--
Bill Sloman, Sydney
Reply to
Bill Sloman

tend to turn into a low resonant frequency for the transformer, but you ca n help that a bit by breaking the secondary windings into banks.

low resonant frequency transformers in photomultiplier supplies. Baxandall 's paper doesn't mention it, but one of my bosses who had worked for Baxand all at the time said that it was the motivation.

Christ. The Baxandal again. Did you ever build it? :-)

I would use a flyback, or to keep peak currents lower at forward, custom wo und

Cheers

Klaus

Reply to
Klaus Kragelund

es tend to turn into a low resonant frequency for the transformer, but you can help that a bit by breaking the secondary windings into banks.

th low resonant frequency transformers in photomultiplier supplies. Baxanda ll's paper doesn't mention it, but one of my bosses who had worked for Baxa ndall at the time said that it was the motivation.

wound

Or perhaps even a two stage boost...

Reply to
Klaus Kragelund

A two-stage boost, exactly. Simplicity reigns. Boost because it's exceedingly simple, an ordinary inductor. Two stage to avoid an extreme on/off ratio, but hey, both "stages" can be run from the same PWM driver. Ooops, I feel an ASCII drawing coming on, arrghh!

. Two-stage boost convert, one IC, two MOSFETs . . 3.2V 20 to 30V 200V . PWR in -------, ,---+---, ,---+---+-- HV out . | | | | | | | 200V . L _|_ C L _|_ C | 0.1A . | /_\ | | /_\ | fdbk . | | GND | | GND . boost +---' +---' . controller | | . ______ |--' Q1 |--' Q2 . | |_____|| __|| . | | | |->, | |->, . |______| | GND | GND . | | '-----------' . GND '----- feedback

Q1 is a low-V, high-current, logic-drive MOSFET and Q2 is a 400V, modest-current logic-drive type. Oops, I don't think you'll find 400V logic-drive MOSFETs, so a small 3.2V to 10V boost switcher may be needed to power the two-stage boost converter. Or better, the intermediate node can be diode-OR'd to power the controller IC once things get going. Table 9.6 in AoE III has candidate controller ICs and seven in the list can run on 3V or less.

Also, a reminder, chapter 9 is free on the web,

formatting link
We need to put up a corrected version, but meanwhile I show corrections as comment notes to this file.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

es tend to turn into a low resonant frequency for the transformer, but you can help that a bit by breaking the secondary winding into banks.

th low resonant frequency transformers in photomultiplier supplies. Baxanda ll's paper doesn't mention it, but one of my bosses who had worked for Baxa ndall at the time said that it was the motivation.

I've built several, and at least one variant, for a variety of jobs.

It's actually a rather popular circuit - Jim Williams used it in his cold c athode back-light driver as described in Linear Technology's application no tes AN45, AN49, AN51, AN55, AN61, AN65). Jim Williams did describes the inv erter as a "current driven Royer" which is isn't, but Peter Baxandall had p ublished the circuit long before that (in a UK journal that isn't easy to g et hold of and no American has ever read) and there are suggestions that Ji m Williams got the circuit from somebody in the UK who hadn't bothered to t ell him where it came from.

wound

Storing energy in an inductor and dumping out out again isn't an elegant so lution.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

That's correct, but it avoids a custom transformer.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Funny sort of simplicity.

John Larkin seems to have been frightened by a transformer as a child, but they aren't that complicated. By the time you've worked out how to stop you r inductors from saturating - I always found myself having to gap mine - th e simplicity of the inductor becomes rather less evident.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

400V logic level:

formatting link

A suitable candidate:

formatting link

Cheers

Klaus

Reply to
Klaus Kragelund

The turns ratio isn't so critical in a flyback as long as you don't have to max out any heat-space-size budget. Just make sure the (now driven) lower voltage winding can take the current and the core won't saturate. If form factor is a problem or gets tight talk to Renco.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Or use a core with the "distributed gap", such as the Super MSS/Sendust ones.

Best regards, Piotr

Reply to
Piotr Wyderski

Why not stack it up for higher voltage?

And won't last very long.

Reply to
edward.ming.lee

John? I'm the one trying to avoid a transformer. I've spent many thousand of $$ on cores, bobbins, and have wound more carefully-design xfmrs than you can shake a stick at, that's why I'd rather go with an inductor, or perhaps a stock 1:1 xfmr.

Yes inductors usually need gaps, but thankfully distributors have good inventories of Coilcraft, etc., inductors with good gap and turn selection, as evidenced by their saturation-current specs.\

That said, I'm always on the lookout for 10:1 and higher stock HF transformers, and discover suitable p/n in IC manuf eval designs - parts from Wurth Elektronik or Pulse Engineering, etc. to discover they're special order, large qty.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

There are some Wurth parts in the Digikey catalog that come close to being suitable, but only go to 15W. I'm tempted to use a pair of 12W pieces, to keep the height down.

Or, I'll try your 2-stage idea for contrast.

--
www.wescottdesign.com 


----Android NewsGroup Reader---- 
 Click to see the full signature
Reply to
Tim Wescott

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.