Vasile Surducan wrote:
> Maybe we should start with naming the most important parameter which
> has never been mentioned here: which is the switching frequency and
> which is the maximum current load?

The switching frequency varies, but will be around 100KHz.  The maximum
current should be a bit above 500mA (I don't have my notes where I am right
now).

> it does not need such big engineering in driving the power FET.

What "big engineering"?  This high side FET driver uses exactly 3 cheap
bipolar transistors, 3 0805 resistors, and 1 0805 ceramic cap.  That's about
$.20 even in moderate volumes.

> Compute how big is the 2K * Ciss compared with the switching time
> (which is producing Q3 dissipation). Ciss is about 2nF at a switching
> frequency of about 1MHz and Vgs=0V (I doubt you can switch Q3 from
> PIC10F at this frequency, so probably it's less than half). So the
> estimated time constant Tau is about 4 microseconds, does the
> 4microseconds switching delay killing your FET ?

You are forgetting that the Q2,Q4 emitter followers greatly lower the
apparent impedence of R4 as seen from the gate of Q3, and the time constant
is significantly shorter because of that.  This is precisely why Q2,Q4 are
there.

> R4 and C11 could dissapears.

C11 only speeds up the switching edges a bit, and could be deleted resulting
in a small loss of efficiency.  However R4 is central to the design.  It
converts the current of the current sink Q6 to a voltage, which is then
presented to the gate after buffering by Q2,Q4.

> Instead those two a resistor in the Q6
> base. If need a switching acceleration scheme could be there in the
> base of Q6.

You are missing the same point as Russell.  Q6 forms a constant current sink
when on.  This allows the switching step size accross R4 to be the same
regardless of input voltage.  In other words, the constant current sink
decouples the switching signal from its ground reference.

> I bet the schematic is working the same without Q2 and Q4

The switching edges would be much slower.  Too slow for it to even work
properly given the rest of the parameters.

> I presume the real supply never excedeed 15-20V.

Nope.  It is about 25V if the recommended wall wart is used, and I've tested
it accross the full range of course.

> Also if the Q3 it's
> packed in SO8 package, than have only 2.5W maximum dissipated power.
> So it's not a power supply.

Huh?  The point of a switcher is that the switching element spends most of
its time either full on or full off, thereby dissipating little power.  Q3
doesn't dissipate much power and doesn't need to be in a large package.  In
practise it's hard to tell it got warm.  In fact, the input diodes get
warmer than any part of the switcher.


******************************************************************
Embed Inc, Littleton Massachusetts, (978) 742-9014.  #1 PIC
consultant in 2004 program year.  http://www.embedinc.com/products
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist