Alt-BEAM Archive
Message #03281
To: Wilf Rigter Wilf.Rigter@powertech.bc.ca
From: Steven Bolt sbolt@xs4all.nl
Date: Tue, 11 May 1999 18:34:16 +0200 (CEST)
Subject: [alt-beam] Re: My Pagerbot
The common BEAM SE still is a two-transistor latch/driver
combination, which uses a zener, a led, an fled or a 1381 as
voltage trigger. The latter variant can be modified with a diode to
increase the switch-off level as described above, from about 0.8V
to 1V. The small change indicates a reasonable 3906 base current at
switch-off.
Your experiments indicate that the fled-variant, when modified like
that, has it switch-off level increased by more than 0.6V. As there
must flow about 1mA before you see that kind of Vf, I submit that
your 3906 (which is not the driver) is getting excessive base
current. It seems safe to predict that a LED in that position will
actually flash :)
Of course I still maintain that the latch and driver have to be
separated to make the SE a well-performing, predictable circuit;
look at Ken Huntington's Micropower SE and the 4-transistor
SunEater for ideas.
> I indicated in my email and schematic that my circuit uses a 1K base
> resistor instead of the conventional 2.2K in part to compensate for the
> lower voltage drop across that resistor and to improve Vsat. If you look
> more closely at my schematic you will notice that I separated the FLED load
> resistor and feedback resistor functions. That was the improvement I
> referred to.
Noted, but that doesn't make the 3906 base current look any better.
> At the 2N3906 base, a 10K (or larger resistor) base emitter
> resistor shunts less current away from the base than the original 2.2K and
> should make the circuit more sensitive and less likely to "hang". Moreover,
> simply adding the 1N4148 in series with the base of the 2N3906 as you
> suggested would appear sub-optimal since it requires an even higher turn-on
> current ie the voltage drop across the 2.2K resistor (and shunt current) at
> turn-on must be greater (Vbe+Vf).
Using the 1381C as voltage trigger, switch-on voltage and current
remain about as they were. The switch-on current, specifically, is
a single digit uA figure.
> In the original FLED SE design simply changing the 2.2K resistor to 10K
> would reduce turn-on current but also reduce available base currents and for
> larger motors this would cause the 2N3904 to come out of saturation.
Apparently, the 1381 version doesn't suffer this trouble. The
driver (a BC337 in my experiment) saturated well enough; to a VCE
of 0.1V, while driving a small recorder motor.
> Adding the diode at the NPN collector and isolating feedback paths by using
> a separate FLED larger load resistor for low turn-on current and a separate
> smaller feedback resistor for low Vsat during would appear to give the best
> of both possible worlds.
No harm in trying to improve the fled SE. Though I'd like to see
the last of it. The basic design just isn't that good, causing
those new to electronics all sorts of trouble which can't be
properly solved. The same goes to a lesser degree for a
two-transistor latch/driver with a 1381 voltage trigger. Separating
latch and driver would be a very good idea in this case as well.
And in most cases, a monostable will outperform the latch by a
large margin.
Best,
Steve
----------------------------------------------------------------------
# sbolt@xs4all.nl # Steven Bolt # popular science monthly KIJK #
----------------------------------------------------------------------
------------------------------------------------------------------------
eGroup home: http://www.eGroups.com/group/alt-beam
http://www.eGroups.com
- Simplifying group communications
Home