Thanks Roman and Russell,

I am going on vacation for a week and will experiment with your advice when
I get back.

--BobG

-----Original Message-----
From: pic microcontroller discussion list [mailto:PICLIST@MITVMA.MIT.EDU]On
Behalf Of Russell McMahon
Sent: Thursday, August 29, 2002 6:35 AM
To: PICLIST@MITVMA.MIT.EDU
Subject: Re: [EE]: Switching Regs and EMI (was: Crowbar protection)

> > I am considering using your switcher as a pre-regulator to go from the
12V+
> > to 3V+ and use a linear 3V reg to make it clean.
>
> I would suggest using an extra transistor, cost
> of about $0.04, will improve the regulation a lot
> and efficiency a bit. You will be spared the 3v
> regulator unless you specifically need current
> limit and thermal protect etc. I think Russell
> McMahon had a good 3-tran circuit.

In its 3 transistor form my circuit probably wasn't as efficient as Roman's
AFAIR BUT it has extremely good regulation considering its simplicity -
AFAIR about 0.1v variation in output across wide load and input voltage
variations. Also, in the form I used it it had an "immense" input voltage
variation capability by normal standards - say from about 11 volts to 200+
volts input for 10 volts out. In its 11 - 200 volt input production form I
use an extra transistor in the high side driver and a FET as the main
switch.

Attached is a circuit which is an *** EXAMPLE ONLY *** as an experimental
starter. Cap CBUK2 is usually much smaller than this. 0r3 in output is
probably unneeded - was used as part of a current limit which I have
omitted. Actual output voltage is close to zener ZBUCK1 value due to zener
knee. CBUK1 adds formal static hysteresis to the feedback path but is
actually not needed in practice. Diode DBUK1 replaces a low side drive
transistor and produces about the same result. LBUK2 provides an extra
"almost free" output but this only gets fed during flyback part of cycle so
will tend to sag somewhat as Vin approaches Vout. Note that FET is (of
course) a P Channel device! PNP transistor can be substituted with suitable
component changes. V100 is nominally a <= 100 volt rail but the cct still
operates OK at 200v but at lower efficiency. ZBUK2 protects FET gate. While
the circuit and method of feedback caused much debate when first introduced
starting is very reliable in practice.

Inductor value is quite flexible. The circuit shown provided about 600 mA
out nominal (over an amp on demand) but can be scaled down with suitable
component changes (left as an exercise for the student :-) ).



        Russell McMahon

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu