Discrete boost converter with only 9 components, does anyone have a schematic?

[Refrigerator]

So i saw Bruce from RCmodelreviews review a very simple yet effective boost converter.

But for the life of me can i not find a schematic for it. I'll be boosting lipo 1S to 5v for a little audio amp (LM4871) and this seems like a very simple yet effective way to do it, the amp would only draw 300mA at the very most.
From what i saw in the video the booster only uses two semiconductors, four capacitors, one diode, one inductor and one 0 ohm resistor.
The signal wire seems to have nothing to do with the booster since it has it's own little trace going around the whole circuit.
To me this is pretty much black magic, i have no idea how those chinese people did it. 

[wraper]

Discreet boost converter won't be good nor in efficiency, nor in the parts count/size.

[Refrigerator]

Discreet boost converter won't be good nor in efficiency, nor in the parts count/size.

But it is good for experimenting and getting the basic undertanding of how it works.

[wraper]

But it is good for experimenting and getting the basic undertanding of how it works.

But you wrote that you want it for small battery powered device (where you normally would want max possible battery life and small size). For making a device which you are going to actually use, this is not a good choice.

[Refrigerator]

But it is good for experimenting and getting the basic undertanding of how it works.

But you wrote that you want it for small battery powered device (where you normally would want max possible battery life and small size). For making a device which you are going to actually use, this is not a good choice.

I want the amp to give it's max power, the battery life isn't such a big concern for me. For the battery i think about using a 1.5Ah LiPo from my old Xperia Arc, or a Sony 18650 3000mAh LiIon ( salvaged from a genuine sony battery pack from a camera ).

[mikerj]

http://www.romanblack.com/smps/conv.htm

[rsjsouza]

Something like the TPS61220, perhaps?

[TerminalJack505]

You could probably adapt this circuit to suit your needs.

It's fairly minimal and completely discrete.

[Refrigerator]

Thanks for the replies.
I've just tried the two transistor mikerjs circuit but just connecting the amp without any audio completely kills the oscillation.
The IC seems like the best way to go. It's really interesting how the HK booster actually regulates the voltage, the current being drawn is not constant.

[T3sl4co1l]

I don't like the Black circuits, they're poorly regulated, usually hysteretic, and know little or nothing about short circuit protection or current limiting.

I've got a few that are short on transistors,



This one being made for a high voltage output, but you can do whatever with the coil and what's after it.  Like the Black circuits, this depends upon switch node voltage, so is supply sensitive (most importantly, too high voltage will cause it to run full-tilt as the 2N3906 discharges the capacitors via E-B avalanche), and the frequency will vary all over the place as supply is varied.

The 2N4401 can be replaced with a MOSFET for slightly better performance or capacity, but you're still limited by the 0.7V drop on the shunt resistor (2.2 ohm).  It takes two additional transistors to reduce that.

Since you didn't specify if the components should be integrated or discrete, I would better suggest a SimpleSwitcher LM26xx or similar device.  Those do around 5 components, and work very nicely; as long as you don't mind that there's a dirty IC on your circuit.

Tim

[Refrigerator]

I don't like the Black circuits, they're poorly regulated, usually hysteretic, and know little or nothing about short circuit protection or current limiting.

I've got a few that are short on transistors,



This one being made for a high voltage output, but you can do whatever with the coil and what's after it.  Like the Black circuits, this depends upon switch node voltage, so is supply sensitive (most importantly, too high voltage will cause it to run full-tilt as the 2N3906 discharges the capacitors via E-B avalanche), and the frequency will vary all over the place as supply is varied.

The 2N4401 can be replaced with a MOSFET for slightly better performance or capacity, but you're still limited by the 0.7V drop on the shunt resistor (2.2 ohm).  It takes two additional transistors to reduce that.

Since you didn't specify if the components should be integrated or discrete, I would better suggest a SimpleSwitcher LM26xx or similar device.  Those do around 5 components, and work very nicely; as long as you don't mind that there's a dirty IC on your circuit.

Tim

I don't mind having an IC boost my voltage, in fact i've already looked up some booster modules on ebay. But it's always fun to make something using just a couple transistors, it's really satisfying seeing it work ( or sometimes go pop ).

[Zero999]

I don't like the Black circuits, they're poorly regulated, usually hysteretic, and know little or nothing about short circuit protection or current limiting.

Have you seen the Buck regulators by Black? They're more efficient than some ICs and current limited can be added too.
http://www.romanblack.com/smps/a04.htm

However the boost one isn't so great. It's just an unregulated boost converter with a zener shunt regulator on the output, making it extremely inefficient.

Connecting the zener so it cuts the output when it switches on will improve the efficiency, although regulation and noise will still be poor but what do you expect from only two transistors?