Blars Blarson wrote:
>
> I saw the black regulator discussed here in the messages before I was
> cut off for being a few weeks behind reading piclist.
>
> This intrigued me, and mostly as an intelectual chalange I worked on
> designing it to function better over a wider input range.  My design
> is at:
> http://www.blars.org/blackblars1.html


Hi Blars, you made some interesting changes to
the circuit. The simple constant current source
to supply the zener is the system I also chose
to give a good (large) input range, and is actually
the best general improvement to the regulator
as it reduces current wastage at higher input V,
and stabilises the charge time into C1 which is
the monostable timed-off period so you find the
regulator freq becomes less dependant on input
voltage and efficiency improves. Your values of
100k and 390 ohms are very similar to what I used
in the 3 amp version of the regulator, ie you are
spot on. :o)

But I really can't see the second adaptation working
well at all. I'm guessing you tried to use the same
simple constant current idea (with the 2 diodes etc)
with Q2 so that Q1 base is driven at a constant
current.

In theory this is a good idea but adding R2 in the
emitter circuit of Q2 wrecks regulation as the B-E
voltages need to be hard coupled to give fast switching
and accurate regulation. It also won't give the
constant current effect you desire, not until the
output voltage has dropped a full 0.6v which won't
happen as Q1 will be turned on long before.

The solution I found to the problem of getting
even Q1 base current over a wide input voltage range
was fairly simple, to use a darlington for Q1
and choose a resistor for the base of Q1 that
enables hard saturation at low voltages, which
for your 600mA example would be about 0.2mA. With
higher input voltages Q1 just saturates better
and you waste a mA or so.

The best 4 transistor solution I found was to keep
Q2 optimised for good regulation and hard switching,
and add the extra transistor at the top in a
configuration that gives constant current "on" drive
to the base of Q1 and increases gain.

But even that system required a couple of volts
overhead and for a 5v output the input minimum is
still around 8v and with a tendency toward slower
switching or even going linear with low input
voltages.

To get really good performance at low input voltages
a system is needed to give guaranteed square
switching at all times, like adding a comparator
chip etc but that introduces new issues. :o)
-Roman

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body