On Mon, 10 Dec 2001, Roman Black wrote:

> Vasile Surducan wrote:
> >=20
> > On Sat, 8 Dec 2001, Dave Dilatush wrote:
> >=20
> > > Alexandre Guimar=E3es wrote...
> > >
> > > >    No getting back to the SMPS. I will have an application where I =
will
> > > >have to make a supply with 74 volts DC input and 24 v at 300 ma and =
5 v at
> > > >30 ma. I am thinking about using 2 of your 2 transistor designs modi=
fied to
> > > >the my specific needs. When I finish it I will let you and the list =
know how
> > > >it goes and what modifications are needed.
> > >
> > > With those voltages and currents, why not use Russell's design to ste=
p
> > > the 74 volts down to 24, then Roman's circuit to step the 24 volts do=
wn
> > > to 5?  Russell's circuit is probably more suited to the high-voltage
> > > part of this thing, especially if you need 300 milliamps out of it.
> > >
> >   I wouldn't do it like this. For me sounds like a standard switching
> > supply using a driver ( tl494, UC3842 ), a mosfet or bipolar switch and=
 a
> > transformer, then a good rectifying diode pair and finally a filtering
> > stage.
> > To increase the efficiency the driver could be supplied directly from
> > output voltage after the oscillation has begun.
> > I know that many piclisters are afraid about designing feritte core
> > transformers for switching supplys. So why many of you are using standa=
rd
> > coils. However for dedicated applications we need dedicated solutions.
>=20
>=20
> Vasile, 24v at only 300mA is well within the
> range of the cheap circuit with a pre-wound
> inductor costing a few cents. Two transistors,
> cheap inductor, no need for a chip or otherwise.
> I've spent years winding ferrites too and its
> nice sometimes not to have to do it. :o)
>=20
> If REAL efficiency is needed I would PWM the
> solenoids direct like Alexandre suggested, as
> you can give them full current for just long
> enough to "pull-in" and about 1/3 the current
> for holding. I have a one R one C circuit
> that does this that i've been using for many
> years, (but with the added PWM this would get
> very very efficient for solenoid driving.)
>=20
> The non-pwm circuit:
>=20
> + ---R-------------
>          |        |
>          |        |
>          |        SOL(+diode)
>          |        |
>          C        |
>          |        |
>          |        Transistor
>          |        |
> - -----------------
>=20
> This circuit allows full 12v etc to pull-in
> the 12v solenoid, and you simply choose the
> R value to give enough holding current. You can
> hold a 12v 100mA solenoid reliably with about
> 35mA, and just make sure C is big enough to
> provide the pull-in energy. I invented this
> when I was about 12 playing with my first
> large solenoid. I don't think i've ever used
> a solenoid since without doing this, AND its
> great for keeping solenoid hash away from your
> PSU. :o)
> -Roman

  Roman, 70V to 24V and 24V to 5V using two=20
separate supplys, sound for me a little strange
even at 300mA.
One transformer with two output coils, one for=20
+24V ( and/or driver) and another for +5V sounds
better. At this current, efficiency of up to=20
85%...90% can be easily obtained.=20
The solenoid ( or transformer as I suggested,=20
to isolate the output ) is not a big deal, small
core, tiny wires. A HV driver can drive directly
the transformer. Resonant solution may minimise=20
the input current and increase the efficiency.
Just a thought, don't kill me [grin]
I'm glad you're OK.
Vasile=20

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.