Re: Simple (but long) solar charging question


Lars Torben Wilson wrote:
Hi all,

(Just reread and grasped the length of this--sorry. Not looking for
specific answers, just insights into things I am perhaps
misunderstanding--any pointers to good reading help. :) )

Another problem from an experimenting noob.

I have a schematic for a simple solar-recharging garden light. This
thing uses a very simple blocking oscillator to power a bright white
LED (3.9V) from a 1.25V AA NiCad cell, and works wonderfully.

1N4004

.--------->|---------o--------------------.
| | |
| | |
o-----' _'_ |
| | .-------UUU-----o-----. |
| | | | | |
| | | 20-T-20 | | |
/+\ | .-. | | /+\
( ) | | | | | ( )
\-/ | | | 1K | | \-/
1V PV | | '-' | | | 1.25V NiCad
| | | |/ | |
| | '-------------| V -> |
| | | |> - |
| | |/ Q1 2N4401 | | 3.9V |
| '---| | | |
| |> | | |
| Q2 2N4401 | | | |
| | | | |
'-----------o---------------o-----o-------'
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

Short description: Q1, the 1k resistor, and the inductor form a
blocking oscillator which pulses into the ~1.25V from the battery
into fairly high frequency ~3.5-4V pulses which are sufficient
to power a bright white LED. Q2 disables the oscillator when
the solar panel hits enough voltage. The 1N4004 prevents the
battery from discharging though the solar panel.

Full explanation from someone wiser than I:

http://www.cappels.org/dproj/ledpage/leddrv.htm#Solar_Powered_Garden_Light

As I said, it works great--at least, it powers the LED quite nicely
from the battery.

Here's the problem--as you see from the schematic, I only have a
1V/400mA solar panel. Sources I have read seem to indicate that you
want more voltage from the charging source than the battery can
provide in order to charge the battery. Obviously the above circuit
falls short of that goal.

Other things I have read indicate that amperage is also important, as
far as charging rate goes--charge at C/10 for a fairly safe long-term
charge etc. But it seems to me that the amperage won't mean much if
the voltage is insufficient to push it into the battery. Am I way off
track here? Can a good charge into a 1.25V NiCad be had from a 1V
400mA panel (in reasonable light)? Or should I look into getting a A)
higher-voltage, lower-current panel, or B) converting some of that
amperage into voltage?

That was the first question, in a very roundabout format. :)

The second question involves me having chosen B) above to work on. I
reasoned that I might be able to have the solar panel, instead of
simply disabling the oscillator when it hits a high enough voltage to
do so, reverse the connections and instead pump its output through the
oscillator into the battery.

Now, one problem has to do with my switching topology, but my question
about *that* will have to wait until I'm a little more familiar with
AACircuit. The real question I have is: will that even work for
charging a battery? As I understand it, the output from a blocking
oscillator is a series of very fast, very short pulses. However, it
looks fine (solid DC line) on my scope. My scope may well be too slow
to show it, though (EICO 460 (pawn shop special)). But my DVM shows
around 4V when the solar panel has enough source light to power the
LED.

Am I mad? Yes, I am testing this but I have a shady yard in
Vancouver--which is pretty cloudy these days. It seems 4V should be
plenty to charge the battery, if I can work out the switching--if the
pulsing doesn't affect the charging.

So am I barking up the wrong tree here? Will the tiny pulses from the
oscillator (which I have read about but not personally observed) not
provide enough average power to charge the battery? Or is the
DVM--which I think gives an average DC voltage reading--showing me
what the battery would be seeing--enough to charge it well?


Thanks for any insight,

Torben


Hi, Torben. Quick summary:

* There's no way you'll get the circuit above to charge the NiCd
battery from the solar panel you have. Even if you use a low forward
drop schottky diode, the panel voltage will have to exceed 1.4V before
the diode becomes forward biased enough to conduct anything into
charging the battery. And you'll need more than that to start getting
a good charge. Battery voltage rises as it's being charged. Sorry.

* Your idea B) of "convering some of that amperage into voltage" is
commonly called a step-up switching power supply. There are ICs you
can use which will control a step-up switcher to boost your 1V into 3V
or more. If you really wanted to pursue this, you would look at the
Linear Technology line of micropower step-up and flyback regulators.
You can actually have the solar panel voltage drop well below 1V, and
still get a regulated output. Of course, since TANSAAFL (There Ain't
No Such Thing As A Free Lunch), your power into the regulator can't
exceed your power out. Since you can only supply 400mW (1V * 400mA),
your 3V output can't exceed 133mA.. Since switching regulators are
typically 80% to 90% efficient, you realistically won't expect more
than 100mA or so at 3V. However, that might be enough for your NiCd
battery, especially if you can get more inductance into the circuit to
slow down the oscillator.

Three big difficulties here. The first is, putting together one of
these will certainly cost quite a bit more than just swapping out the
solar panel. Second, once you've developed the step up switcher,
you'll need another circuit to limit current into the NiCd battery (a
simple resistor won't cut it here). Third, switchers are inherently
tricky, and if your big gun is an Eico 460 with a non-functioning
trigger circuit, you're not going to get very far if something goes
wrong (and with switchers, something almost always does, especially for
those who don't carry the scars of battle). From a newbie perspective,
this tells us to slowly put down the scope probe and step away from the
concept.

It might be worthwhile for you to look at a Linear data*** to see how
it's done. Go to the Linear website:

http://www.linear.com/index.jsp

and punch in LTC3400 under search. Look at the product page and the
data ***, and see what you think. But remember, even though the
typical apps circuit looks simple, there are big problems lurking under
the surface. At a minimum, you'll need a well thought out etched
circuit board with a ground plane to make this happen. And a good
scope.

You're not "mad", you've asked a couple of really intelligent newbie
questions. If the fate of civilization depended on it, it could be
done. Unfortunately, A) looks like the only realistic newbie option.
And I'd go with a larger 6V solar panel capable of supplying 200mA
minimum, so I could just use a series resistor to safely and reliably
charge the battery.

Vancouver is one of the great places to be in early summer -- get out
and enjoy the day.

Good luck
Chris

.

Loading