Thread: Flashing Battery Status Monitor

I would like an L.E.D. to flash every couple of seconds
when a sealed gel-cell battery is fully charged, then flash
more frequent as it discharges, to eventually staying on
continuously when it needs to be charged. I understand
I'll have to set the voltage levels with a potentiometer to
"meet my needs". I would also like to incorporate a quick
charge into it. 12V20Ah, and I would use the 12V700ma
wall charger (as the quick charge bypass), and when not
in bypass I would use it to float the battery. How big of a
resistor (value and wattage) would be ideal to reduce the
transformer for float charging? How about a circuit that
oscillates the L.E.D. as the battery goes down? I know, it
may be better to have the L.E.D. light less as the battery
depletes, but I want it to grab someone's (my) attention.
Even if it can't be electronically performed, I s'pose I would
settle for the reverse (fully on at full charge, flashing slower
as discharging and going off when in need of a charge), but
I would much rather go the other way (more attention when
in more dire need). Anyone???

My ideas so far about the single L.E.D. battery
status monitor are as follows, and my questions are
included...

The sample timer would be an astable variable
duty timer (lm555) that would incorporate an
isolation transistor to turn on a different astable
timer (lm555 on a 50/50 cycle) that has it's voltage
to the R reduced in such a manner that it slows
the charging of C accordingly. But this would cause
the L.E.D. to flash less frequent as the battery dips.
How could I reverse this, so as the voltage dips the
C would charge faster (causing a shorter duration).
Might I incorporate a network of Zener/general
purpose diodes with transistors to "install" different
resistors for the C, for say 8 different voltage levels?
But how could I readily do this with the least current
draw and smallest footprint? Also, I'm aiming to have
the oscillation vary at steps less than 0.7v apart!
I'm so stuck...How could I do this???
Then, I need some sort of a circuit that senses
the lowest voltage setting and keeps the L.E.D. "on"
until V rises. The float charger has an L.E.D. for when
the battery is fully charged, and I could inverse this
signal (to override the "monitor") with an isolation
transistor, to "disconnect" the L.E.D., hence leaving
it "off". But I need a circuit to keep it on when a low
threshold has been reached. Anybody with any ideas?
I was also thinking that maybe for the slow charge
option, I might have to use a diode and not a resisitor.
I know the diode reduces the voltage and the resisitor
limits the current, so which would be better for a slow
charger? Reducing the adapter by 0.7v or the current
being supplied? Say I want to lower the 12V700ma to
12V200ma, does this mean that I need a 60Ω resistor
to do this or a 24Ω? How many watts in size minimum?
What would be most ideal as far as eliminating blowout
and/or it being too hot to the touch? Or should I do
the diode (or two?), and if so, w/ sinking up to 700ma,
how many amps would be the most ideal for such (re
eliminating blowout, keeping it cool)? Would a 1N4001
suffice <I think it's 1A 50V>, or do I really need to kick
up the "size" of the diode? 2A?? 3A??? More?????

Well, I have the timer circuit finished. It is
actually two variable duty timers I'm using. But I
still need a means of selecting which resistor to
"insert" between pin7 (of the second 555, which
controls the L.E.D.s flash rate) and VCC. I am
still wondering about creating a transistor "ladder"
but also don't know how I could keep it out of
"bar graph" mode (so only one transistor is on
at a time). Can anybody help me?
I was thinking of resistors for measuring the
voltage levels, and somehow having each drive a
transistor to pass VCC and a resistor for the timer,
through to the timer. But I am guessing I would
also need another transistor that somehow cuts
off the Vin from subsequent transistors on the
ladder so only the circuit just above the cutoff
voltage will activate. Any ideas anyone??? I was
also thinking of a "bargraph voltmeter" chip (I do
know there's one out there but I don't remember
it's number), set it to "peak" (no bar graph) and
have each of it's outputs do the transistor and
resistor thing (it would need a Vcc out from the
chip, and not ground). Help...

Well, it's the LM3914 voltmeter chip. This would do the beauty.
How could I inverse the chip's outgoing gnd to Vcc??? (then I could
place different resistors between each transistor(?) and pin7 on the 555).
Would I need an inverter for this, and if so, can anyone suggest the
device (Vcc is 5v, Vin is up to 13.8v or so) most ideal, and how I
should wire it?

Vin => LM3914 => each level pin (Gnd out) => inverter? (+5v out)=> R
(all level pins arrayed to pin7 of 555 after each levels resistor).