boost mode switcher design

[eccerr0r]

Has anyone tried designing a high voltage boost supply?

For characteristics I tried connecting a 555 to a fairly high voltage transistor from a junk pile (2SC2979), and used one of those mains filter coils from a dead PSU (I think it's the one where there are two parallel coils for the two input lines - except I hooked the windings up in series for more inductance. My Sencore measures about 40mH when they're in series but 10mH for each winding, very weird...) The 555 and the switch is powered by about 12V for now.

Come to think of it, I should question the inductance some more... 10mH is fairly high for that inductor... hmm.

Switching frequency is about 7.5KHz, probably should try to bump it up some but I just randomly picked some parts up from my parts bin. Currently running open loop, except for a current limit on the 2SC2979, in which if it exceeds about a hundred milliamps it will shut down the 555.

I got about 280 volts or so from a multimeter and this can definitely light a neon bulb through a 1M ohm resistor (with the DC caveat problems).

I'm wondering if I can possibly get more out of this. I've seen a lot of things indicating that once I add in voltage regulation, it will be excruciatingly hard to stabilize - which is OK, this was meant for a Geiger counter and I suppose as long as it's in range (targeting around 400V +/- ~30V) @ 0.5 mA, it should be fine.

If anything I need a better inductor...

[PeteS in CA]

Re: boost mode switcher design

The A(L) for a magnetic core is expressed as nH or uH per turns squared. So if you double the number of turns, you quadruple to inductance.

[eccerr0r]

Re: boost mode switcher design

Ah yes that would explain it, did not think of it as a shared core which it clearly is. I was thinking it of independent cores but in this case that's not true.

[kaboom]

Re: boost mode switcher design

Has anyone tried designing a high voltage boost supply?

For characteristics I tried connecting a 555 to a fairly high voltage transistor from a junk pile (2SC2979), and used one of those mains filter coils from a dead PSU (I think it's the one where there are two parallel coils for the two input lines - except I hooked the windings up in series for more inductance. My Sencore measures about 40mH when they're in series but 10mH for each winding, very weird...) The 555 and the switch is powered by about 12V for now.

Come to think of it, I should question the inductance some more... 10mH is fairly high for that inductor... hmm.

Switching frequency is about 7.5KHz, probably should try to bump it up some but I just randomly picked some parts up from my parts bin. Currently running open loop, except for a current limit on the 2SC2979, in which if it exceeds about a hundred milliamps it will shut down the 555.

I got about 280 volts or so from a multimeter and this can definitely light a neon bulb through a 1M ohm resistor (with the DC caveat problems).

I'm wondering if I can possibly get more out of this. I've seen a lot of things indicating that once I add in voltage regulation, it will be excruciatingly hard to stabilize - which is OK, this was meant for a Geiger counter and I suppose as long as it's in range (targeting around 400V +/- ~30V) @ 0.5 mA, it should be fine.

If anything I need a better inductor...

I tried something like this years ago. IIRC, those common mode chokes don't seem happy with such imbalance and current. You'll find, most likely, that it'll saturate over even half an amp. Makes sense- they normally have no DC, and operate in common mode, not normal mode. Also, the cores are ungapped.

You can try a -26 or -52, commonly seen as "coupled output inductors" in PC SMPSes. Better yet if there are parallel/bifilar windings for either +5 or +12. Again, in series, they'll yield 4x their individual inductance. Make sure you increase your freq as well.

Ah yes that would explain it, did not think of it as a shared core which it clearly is.

Their filtering action depends on this.

[ben7]

Re: boost mode switcher design

I think I recall making a decent PSU (maybe 15W output) for high voltage before, with an LM555.
Traditional ways of regulating the output use the control pin on the 555, which I find is a bad and in-efficient way of doing it. The best way I found was by adding a bit more circuitry to interrupt the pulses going from the 555 to the mosfet, thus having a pulse skipping sort of regulation. I got it to run quite efficient too, the mosfet (T0-220 style) didn't need a heatsink at all, compared to other methods of hooking up the 555 (and yes, fiddling around to try to make it work the best I can). I just used a single 100uH inductor, and ran the 555 at about 100kHz. A filter cap is important on the input side of the PSU too.

[cheapie]

Re: boost mode switcher design

The best way I found was by adding a bit more circuitry to interrupt the pulses going from the 555 to the mosfet, thus having a pulse skipping sort of regulation.

Do you mean like those old hit-and-miss engines?

[ben7]

Re: boost mode switcher design

Do you mean like those old hit-and-miss engines?

Yep!

I love those

[tom66]

Re: boost mode switcher design

Main problem with that type of design is capacitor ripple current can be through the roof, so make sure you use good capacitors on the output, ideally poly-film type or very large electrolytic caps.