Re: Transistor booster design

On Thu, 02 Aug 2007 09:15:27 -0700, linnix <me@xxxxxxxxxxxxxxxxxx>
wrote:

On Aug 1, 4:25 pm, John Larkin
<jjlar...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
On Wed, 01 Aug 2007 13:44:22 -0700, linnix <m...@xxxxxxxxxxxxxxxxxx>
wrote:



On Aug 1, 1:15 pm, John Larkin
<jjlar...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
On Wed, 01 Aug 2007 07:06:49 -0700, Jim Thompson

<To-Email-Use-The-Envelope-I...@xxxxxxxxxxxxxxx> wrote:
On Sun, 29 Jul 2007 12:15:53 -0700, linnix <m...@xxxxxxxxxxxxxxxxxx>
wrote:

I can't find the old thread. But here is the two transistors booster
model and schematic, as promised.

For BC847B, 6V to 15V @ 1ma:

http://linnix.com/power/main.php?vcc=6&coil=330&load=15

Now that's REAL engineering :-(

He is probably referring to the auto query engine (still going on
every 3 minutes),
but it's been IP filtered.

...Jim Thompson

Blocking oscillator boost converter:

http://s2.supload.com/free/BlockOsc.jpg/view/

The problem is that both inductors are on and off at the same time, so
it could not be more than 50% efficient.

It's a transformer with two windings, demonstrated by the "T1" on
each. The windings are not equal, being a pretty hefty step-down in to
the emitter. The feedback winding can be in the base, too, just about
as good. A bit of series resistance in the base may be needed to
better define peak base current, but a slick custom transformer could
handle that, between copper ohms and leakage inductance.

One of the many cool things about a blocking oscillator it that the
transformer is driven to saturation every time it's used, making
optimum use of the core.

This should be very efficient if designed right.

And manufactured right. I have to check into making small enough
transformer. Using 80% 1000uH:10uH mutual inductors, the model is at:

http://linnix.com/larkin

It takes longer to reach the target voltage, but I guess that's OK.
I had to limit the spice run to 20 ms.

Looks like the oscillation frequency is very low. Try reducing the
base cap.


On reflection, it would probably be best to ground the emitter and put
the feedback winding from the R-C to the base, with maybe a bit of
series resistance.

http://s2.supload.com/free/Block2.jpg/view/

Blocking oscillators are analytically a bit tricky. Very tricky,
actually... there's a lot going on.

John

.

Quantcast