Michael Ross wrote:
Freakish, eh? How much does the state of charge change with a variation of
0.05V on a LiFePO4 cell?
As many have observed, the no-load voltage of LiFe cells barely changes
at all from 20% to 80% state of charge. Voltage is worthless as a way to
gauge state of charge in this range. You'll measure about 3.31v no
matter what the cell's state of charge or condition. The differences are
as likely to be caused by brand, model, age, batch number, temperature,
etc. as they are by state of charge.
Voltage is just a crapulous way to measure SOC on LiFePO4 cells.
That's a great word for it! :-)
Here's my advice on storing LiFe cells: When they arrive, you really
don't know if they are good or weak or bad, and you don't know their
state of charge. Voltage alone is worthless. *Nobody* has ever tested
them, including the factory (they ship *everything* to someone, good or
bad). Like they say in poker, if you don't know who the patsy is, then
*you* are the patsy.
If you're lucky, the vendor may have shipped 100% good cells, all at the
same state of charge, and all with the same self-discharge rate. Since
good cells have a low self-discharge rate, such cells can indeed just be
sat in the corner for a year or more.
However, this has never happened to me. I have *always* found large
variations in the initial state of charge (even though the cell voltages
are the same), and there have always been weak or defective cells in the
batch.
So, it is only prudent to test them. You don't need expensive equipment;
it only requires time. If the cells are going to be stored for a while,
then you have lots of time.
You're only going to spend about 5 minutes per cell. The rest of the
time is just waiting for it to discharge, and then waiting for it to
recharge. You just check and write down the results at the beginning and
end.
Equipment:
- An adjustable regulated power supply that you can set for 3.6v.
Anything from 1 to 10 amps will be fine. Get one with meters
that show volts and amps. There are thousands of these on eBay
for $50 and less.
I prefer to get an older used name-brand one than a new junky
one. It will be more accurate, last longer, and you'll find
*lots* of uses once you have it.
- A load "resistor", which can be anything that draws 1-10 amps
at 3.3 volts. I like light bulbs, because they draw a roughly
constant current -- a couple 12v car headlights are fine. They
barely light at 3.3v, but that makes no difference.
Use your power supply to power your load resistor, and write
down the current it draws. This will be your "discharge" current.
- A SPDT relay, with contacts that can switch the load and charger
current, and a coil that drops out at about 2.5v (the desired
end-of-discharge voltage).
A common 12v automotive relay will work. Use your adjustable
power supply to test its pull-on and drop-out voltages. It will
probably pull in at 6v-9v, and drop out at 1v-2v. Add cheap
common diodes (1N4001 etc.) in series with the coil to raise
its dropout voltage (0.6v per diode) to get to 2.5v.
- A timer. The cheapest and easiest is an analog clock (the kind
with hands) that's powered with a single 1.5v AA cell.
- A reed switch. It's a little glass tube with a wire on each end.
They are sold at Radio Shack, and any burglar alarm sales and
service store. This is the thing you attach to a door or window,
with a magnet on the moving half. When a magnet (or magnetic field)
is near, the switch closes. When there is no magnetic field, the
switch opens.
Wiring:
- Connect the cell positive to the relay's common contact.
- Connect the normally-closed relay contact to power supply positive.
- Connect the normally-open relay contact to the load resistor.
- Connect the negative of the power supply, load resistor, and cell
all together.
- Connect the coil (and however many diodes you need to reach 2.5v)
across the load resistor.
- Connect the reed switch in series with the AA cell in the clock.
For example, remove the AA cell, and tape one wire from
the reed switch to the AA cell's positive end. Hold the
other wire from the reed switch against the positive contact
in the clock's battery holder. Now insert the AA cell.
The tape on its positive end should hold the wires in place,
but keep them from shorting to each other. Test to see that
the clock only runs when a magnet is held on the reed switch.
- Wrap one of the wires to your load resistor around the reed switch
a couple times. This creates a magnetic field when current
is flowing to the load, and so starts the clock.
Operation:
- Set the power supply for 3.6v (or whatever you want your "fully
charged" voltage to be). The cell will charge at as much current
as the power supply can provide. Any value is safe for the cell.
(It's the power supply that may have trouble providing its maximum
current for many hours -- a cheap one may burn out!)
- Let the cell charge until the current falls to less than 1% of the
cell's amphour capacity (i.e. under 1a for a 100ah cell, or 0.1a
for a 10ah cell, etc.) This is not critical, and may take a day or
more. For example, a 2 amp power supply will take 50 hours to fully
charge a dead 100ah cell! So keep checking back once or twice a day,
looking for a current under 1%.
- When the cell is fully charged, manually push the relay contact
closed with your finger. If you can't see the contact (i.e. the
relay case has a cover), either remove the cover, or momentarily
touch a 9v transistor radio battery across its coil to make it
pull in.
- When it pulls in, the relay
a) disconnects the power supply (i.e. stops charging).
b) powers its own coil, so it *stays* pulled in.
c) connects the load.
d) starts the clock.
- Set the clock's hands to noon. The clock will run as long as
the relay is pulled in and the load is being powered. Thus, it
will show the elapsed time.
- Check on it at least once every 12 hours. If it is off, write
down the elapsed time. The cell's amphour capacity is the elapsed
time multiplied by the load current. For example: Your load draws
3 amps. The clock started at noon and stopped at 10:00 (10 hours).
Then the cell delivered 3a x 10h = 30 amphours.
If the clock is still running when you check, write down the
elapsed time so far, and set the clock back to noon. With a big
cell and a small load current, it could take more than 12 hours
to fully discharge it. For example, it takes 3a x 20 hours to
discharge a 60ah cell. So you might check and reset the clock
3 times (reset to noon at 11 hours, then found it off with 9
more hours when you came back in another 10 hours). So the cell
is 3a x (11h + 9h) = 60ah.
This setup automatically turns off the load before the cell gets too
deeply discharged, so it prevents damage if you forget to check it.
It also automatically recharges the cell after a discharge test, and
won't hurt the cell if you forget and leave it on too long (3.6v won't
hurt a LiFe cell even if left for days).
It's slow, but cheap! And in the end, you'll know what you really have,
instead of having to rely on the claims of battery salesmen.
A public-opinion poll is no substitute for thought. *Warren Buffet*
I like this quote. :-)
--
Scientists investigate that which already is. Engineers create that
which has never been. -- Albert Einstein
--
Lee Hart -- See my Xmas projects at www.sunrise-ev.com/projects.htm
_______________________________________________
UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
http://lists.evdl.org/listinfo.cgi/ev-evdl.org
For EV drag racing discussion, please use NEDRA
(http://groups.yahoo.com/group/NEDRA)
