Discussion in 'Electronic Basics' started by pimpom, May 9, 2009.

1. ### pimpomGuest

I know this is an old topic and I've done a search, but couldn't find a
satisfactory answer to the question: What would be a reasonable charging
current for a NiCad battery (in terms of its mAh rating) under the following
conditions?

1. Battery loading will be infrequent but unpredictable - could be once a
day or a week or more.
2. Unattended charging at or near constant current for 6-10 hours a day,
everyday.
3. Charging circuit needs to be simple - no dv or temperature detection.
4. Battery service life need not be as long as that with optimum charging,
but should not be drastically shortened.
5. Does not matter if battery power is occasionally unavailable due to
insufficient charging time between (infrequent) loads, but should otherwise
be ready to provide power at any time.

2. ### whit3rdGuest

So, charge at the ten-hour rate from a current source. 800 mAh
battery
(kind of a low-end AA) will want 80 mA for ten hours, roughly,
so connect a wall-wart with enough volts into a resistor in series
with
the battery, or an unregulated supply with an LM317 + resistor
regulated
current source. The resistor should be sized to R= 1.25V / Icharge,
and suitable to dissipate 1.25 x Icharge watts.

3. ### pimpomGuest

Problem is, it will be permanently connected to the charging circuit and
will be charged 6-10 hrs a day regardless of whether it is discharged or not
in between charges. Under those conditions, I don't think a C10 charge will
be very healthy for the battery.
I'm ok with designing the charging circuit to any degree of precision as far
as current regulation is concerned. I just want to avoid a drastic reduction
in useful battery life by constantly overcharging it, while keeping the
circuit simple by omitting a complex full-charge detection.

Perhaps this is a good time to explain the intended application. I want to
make an emergency light using white LEDs that will turn on automatically in
the event of a power failure. Nothing new in that. But this unit will be
connected in parallel with a normal house light that's usually kept on every
night from about 5 or 6 pm to 1-3 am.

The emergency light will sense the state of the light switch even in the
absence of mains power and turn on only if the switch is in the 'on'
position - a fully automatic fit-and-forget operation. I've designed and
tested the circuit but am not sure what level of trickle charging would be a
good compromise.

4. ### Phil AllisonGuest

"pimpom"

** Most NiCd makers suggest a C/50 rate for constant trickle charging.

But why use NiCds at all??

NiMH cells have lower self discharge rates and less issues like the
formation of "dendrites" that plague NiCd cells left on trickle charge.

....... Phil

5. ### Guest

I was going to suggest a different battery type also. Why NiCds? If
you don't mind the weight there are also lead acid gel cells. I
haven't done much (read anything) with NiMH. But I thought I read
that they didn't want to be trickle charged. They wanted pulses... I
guess you can make a trickle pulser.

George Herold

6. ### Phil AllisonGuest

<
"Phil Allison"
"pimpom"
I was going to suggest a different battery type also. Why NiCds? If
you don't mind the weight there are also lead acid gel cells. I
haven't done much (read anything) with NiMH. But I thought I read
that they didn't want to be trickle charged. They wanted pulses... I
guess you can make a trickle pulser.

** Trickle pulsing is probably a very neat idea - funny I have never seen
it done other than in the standby-by mode of a pulse type fast Ni-Cd
charger.

The OP could set up a 555 timer, in astable mode, to produce a 1 second
pulse each 30 seconds and have a CCS deliver say 500 mA to the battery
during that pulse. Averages out as 16mA or C/50.

OTOH - emergency systems nearly always use Gell Cells and with good
results. Constant voltage ( current limited) charging is the go with them
and is simple to implement

..... Phil

7. ### pimpomGuest

I understand that NiMH cells have fewer issues than NiCds. The main reason
for using NiCds is that they're much cheaper and much more easily available
than NiMHs where I live.

8. ### pimpomGuest

charging are all technically sound ideas. But ATM, I'm committed to the idea
of using one of those compact 3.6V NiCd packs that come with cheap Chinese
import gadgets such as a US\$2 rechargeable flashlight.

My circuit uses 1 low-power transistor, one-half of LM393, 3x1N4007, 1 zener
diode, 8 resistors and 2 caps on a 1"x2" pcb, and I want to keep it no more
complex than that.

9. ### pimpomGuest

A charging timer would be superfluous here as the mains line to which it
will be attached is regularly turned on every night during waking hours and
turned off for the rest of the day - a pretty regular cycle. The uncertain
factor is the discharge frequency which is neither regular nor predictable.
It may go for days without being called upon to deliver power. You cite C/10
and someone else cited C/50. That's quite a big difference.

10. ### pimpomGuest

Hmmm. It looks as if there's no definitive answer. Would you say that
something like C/20 would be a reasonable compromise?

11. ### Phil AllisonGuest

"RogerN"
** That is bullshit.

** Such a silly scheme will result in very long charging times AND the
cells will never reach full charge.

There is also the risk of charged cells heating and hence their voltage
dropping and hence the charge rate increasing on and on to destruction.

....... Phil

12. ### Phil AllisonGuest

"RogerNutcase"

** So says a brain dead FUCKING LIAR.

** Much less, in most cases.

SO YOU LIE

** That is no answer at all to my complaints - so you have lost the debate

Clearly, you are know nothing damn fool.

** None of which is even slightly relevant to the original question.

** Not very interested where you got all your wrong ideas from.

But maker's mostly anonymously authored data sheets, pseudo technical
wankers posing as "experts " for scumbag import operations & ridiculous
software routines embedded in uCs by autistic code scribbling morons are an
* inexhaustible supply * of rechargeable battery bollocks.

Piss off.

...... Phil

13. ### Guest

Pimpom, Try googling "battery university". There is a bunch of stuff

George Herold

14. ### ehsjrGuest

Simple trickle charger with taper:

Use an LM317 circuit to set the current to C/10. Add a TL431 circuit
on the output, set to the full charge voltage, and designed to steal
current such that the current available to the NiCds drops to ~C/50
or less when the pack reaches full charge voltage.

A partial schematic:

Green
+ ---[LM317]---[D1]---+---[LED]---[R]---+
| |
+-------+------[R]--------+
| | |
[R] | |
| [R] |
Red [LED] | |
| | |
+-------+ [NiCd]
| |
[TL431] |
| |
Gnd ----------+-------------------------+

When green is on and red off, current goes only to battery.
When red comes on, current is "stolen" reducing the charge
rate to ~ C/50 with properly chosen resistors. You don't
need the LEDS, they're just nice indicators. I use 1.43
volts per cell as the full charge voltage. Compute the
parallel R's for ~1.8 volt drop when (C/10 - Iled) flows.

Ed

15. ### pimpomGuest

Thanks for the follow-ups. I've visited the links you provided and those
guys seem to know what they are talking about. Even if their information is
not 100% accurate, they give a good overview of the issues involved.

16. ### Don KlipsteinGuest

I would check into dendrites. Overcharging of NiCd cells, especially
long term constant trickle charging, contributes to those. Once a
dendrite punctures the separator, the cell has a foot in the grave.

It comes down to what amount of trickle charging while fully charged
does not cause enough dendrite growth to make cells die too much earlier
than they would be declared dead for some other reason.

Another thing - charging when already fully charged makes the cells
warmer and that makes them age faster. My experience so far is that C/10
warms cells a little noticeably, and that C/4 warms cells enough for
reasonably reliable detection of full charge via either direct detection
of warming (temperature sensor) or indirect detection of warming (voltage
decreases after increasing).
It appears to me that C/20 *usually* does not warm cells enough after
they are fully charged to accelerate their aging much, though I would
worry about the dendrite issue if the cells are cooking 24/7/365 at C/20.

Should you find it practical to reliably detect full charge and then
switch to a lower "trickle charge" current, such as with usage of one of
those "smart charger" ICs that detects reversal of increasing voltage that
results from C/4 or faster charge causing cells to heat up after achieving
"completion of charging", it appears to me good to have such a "smart
charger" reduce the charging current to something close to or slightly
outpacing self-discharge of "topped-off" cells.
I suspect that C/100 and C/200 will work well, C/200 may keep cells only
80-90% "full", though I also suspect that C/50 may be, maybe "usually is",
"not that big a deal" for NiCd cells being cooked at such rate 24/7/365.

NiMH appears to me to have less tolerance than NiCd for trickle charging
24/7/365 at rate exceeding self-discharge that occurs at "full charge".
Not that I ventured major testing efforts, but some NiMH cell datasheets
and some web-search-able items along those lines makes it appear to me
that a NiMH cell has ususlly zero to sometimes very minimal,
likely-well-tolerated overcharging when trickle-charged at C/200 while
fully charged. (NiCd appears to me to "usually not die much earlier than
life expectancy" if trickle-charged 24/7/365 at C/100, fair-to-good-chance
at C/50, some-to-fair chance even at C/20.)

- Don Klipstein ()

17. ### Phil AllisonGuest

"pimpom"
** No way for you to conclude that unless YOU are a expert yourself.

It is easy to find links that COMPLETELY contradict the ones given.

Google " NiCd trickle charge dendrites " for some sobering info.

** No way for you to conclude that unless YOU are a expert yourself.

Which of course you ain't.

....... Phil

18. ### Guest

Thanks for that Don, I was buying a pair of RC battery powered cars
for my son and I a few years ago and after a bit of research concluded
that NiCd's were more robust and less susceptable to abuse than the
NiMH. Even though the NiMH were lighter and held more charge, I
bought the NiCd's. I also read that the NiMH only store about 70% of
the charge that is pumped through them. Whereas the NiCd's keep 90 to
95% of the charge. I wonder if anyone has tested this?

George Herold

19. ### Phil AllisonGuest

<>

Thanks for that Don, I was buying a pair of RC battery powered cars
for my son and I a few years ago and after a bit of research concluded
that NiCd's were more robust and less susceptable to abuse than the
NiMH. Even though the NiMH were lighter and held more charge, I
bought the NiCd's.

** Fool.

the charge that is pumped through them. Whereas the NiCd's keep 90 to
95% of the charge. I wonder if anyone has tested this?

** Even if true ( and it is not) how would it matter ??

...... Phil

20. ### Guest

Why? Cost per A-hr was less for NiCds. Also thought my 9 year old
son could leave them plugged in for a day with out destroying them.

Oh, have you measured this? It certainly makes no difference for an
RC car powered from my wall socket, but I heard that Hybrid cars also
use NiMH, If so that's certainly a factor in all the efficiency
talk.