Trickle Charge Level

C/10 is usually taken as a safe rate for continuous charging. When you overcharge a NiCd cell, it starts producing hydrogen gas at the cathode, rather than deposting finely divided metallic nickel, and this hydrogen diffuses to the anode where it reacts with cadmium oxide (or whatever) to reform water. and realease heat.

If you produce too much hydrogen, the pressure of the gas builds up to a point where the safely valve on the battery pops, and you lose electrolyte.

You can detect the transition to overcharging by monitoring the temeprature of the cell, either directly, with some kind of temperature sensor, or by monitoring the output voltage, which drops appreciably as the cell warms up.

------------------------- Bill Sloman, Nijmegen

I am developing a circuit with a built-in NiCad charger. Rather than spend the extra board space for a high-tech fast charge circuit, I am opting for the slow trickle charge circuit composed of a diode and a resistor. My question is, what is considered a safe level of charge to be left on a NiCad indefinitely in typical room temperature ambient conditions? The battery pack is 300 mAH, 3.6v.

-Robert Scott Ypsilanti, Michigan

It depends.... As an "always on current" you can safely use C/20 which would be 15 mA. Don't forget the current the accu is delivering!

HTH Wolfgang

I highly doubt that you can reliably detect a cell temperature rise with a charge current C/10 (will be 30mA at a 300mAh cell). Also a minus-delta-U detection, as you proposed, is _completely_ useless at charge rates less that C/2, and will only lead to overcharged cells and thus shorted lifetime. Please refrain from telling people this stories. Thanks.

Wolfgang

As i understand it, all rechargeable batteries safely accept C/20 or less.

The method you use depends on your voltage source, and to some extent affects the charge rate.

If the charging voltage source is constant 5v, you're perfect. (The farther away from 5v, the less perfect). The resistor/diode will convert a constant V to a taper charger - heavy rate when the battery is at 3.0, light rate when it's fully charged (1.43 volts per cell). Play with the numbers. For example: When your source is 5V, the diode drops it to ~4.3. Put a 10 ohm resistor in series. With the pack discharged (at 3.0) you'll get 130 mA (C/2.3). When the pack is fully charged (3*1.43 = 4.29) you'll have 1mA and your pack can stay on that rate forever. If the source is say +12, the taper is way less pronounced.

If the voltage source varies all over the place, but is < 35 volts worst case, or is over 7.5 volts, you could use a TL317CLPR (which is a TO92 100mA version of the more familiar 1.5A TO220 LM317) and one resistor to make a constant current regulator in the same board space as a diode and resistor. I don't think the diode is needed with the 317, but you would have to verify that. But if the voltage is noisy, you'll likely need to add a .1 uf cap Vin to gnd, so it might take more board space. You could also use a TO92 78L05 to make a constant +5, then a diode and resistor to get the benefit of the taper charge. (And maybe the diode is not needed with the

I don't know *the* indefinite rate. I use taper charging on NiCds that may be left on indefinitely, and I reduce the rate way down low.

Ed

Insulate the cell well enough, and you could certainly reliably detect a cell temperature rise at a C/10 current, though there'd be no point in doing it. I've published a papsr on a millidegree temperature controller, and probably have a clearer idea of the limits on temperature detection than you do.

Minus delta-U detection (if that is what you call it) is a technique used in practice. It strikes me as horribly risky - what happens if you put a fully charged cell in a cold charger? - but it is a known technique.

And don't waste your time asking user-group posters to refrain from posting stuff you don't agree with - the group doesn't work that way.

You'd should have tried to make me look ridiculous by quoting some numbers for the temperature rises involved.

If the cell is just sitting in air, the thermal resistance to ambient is going to be of the order of 30C/W, so the temperature rise is going to be about 1C - easy enough to measure with a thermistor on the cell and a second reference thermistor a couple of inches away to compensate for changes in ambient temperature.

I'd be temepted see how a small Peltier junction would work as thermocouple, but I've not seen it done (since it would be much too expensive to be practical in most applications).

I'd say the egg was on your face. Why don't *you* refrain from casting unjustified and undocumented aspersions?

--------------- Bill Sloman, Nijmegen

Good chargers usually have timers as well. A document by Sanyo on NiMH chargers recommends using three terminators: noticing a small decrease in voltage, a marked temperature increase, or timeout. the decrease in voltage is the usual scheme, and some chargers forgo temperature change detection. For some NiMH batteries, they recommend 2C charging, since the voltage drop won't be noticeable otherwise. On the other hand, the temperature spike when the cell is charged is quite noticeable.

Looking at the Sanyo Cadnica battery specification, they recommend a trickle charge rate from C/50 to C/20, so the OP might want to use a constant current from 6 to 15mA.

I checked it this morning. The diode is necessary to prevent current from the battery "backflowing" to ground through the power source impedance if power is interrupted.

Ed

The safe trickle charge current for NiCad varies with construction. There are a number of types (and the 'normal' charge permitted reflects this as well).

For instance, Panasonic 'high temperature' cells cannot accept rapid or fast charge (or heavy discharge either), and may be trickle charged at C/30 or less continuously. I found this safe level empirically. Charging one of these continously at C/10 was indeed detrimental to it's operation.

Different cells with different internal construction (the Sanyo sintered cell for instance) can accept all the charge rates I mentioned and provide pulse outputs of up to 5C for

it's caled "1/10C", it means one tenth of the "hour" capacity.

so, 30mA

going lower may improve the life-span of the battery, going higher will be dertimental.

Bye. Jasen

That's getting close- some of the ultra fast chargers use a temperature controlled fan to keep the battery temperature from rising to the point of significant capacity loss- and of course, the higher the current, the more reliable the neg delta v method- but it looks like they only get bold like this with single cells and not packs.

Any overcharge will build up some pressure in the cell - it is a linear function of of the overcharging current. The hydrogen pressure, as such, has no effect on the cell voltage, though it rises linearly with current, just like the heat dissipation, and thus the tmperature, which does affect the cell voltage.

So this is a real problem for the minus delta-U detection technique (minus delta-V detection in countries where we talk about voltage rather than SpannUng).

These are not schemes for beginners, but they are used - and discussed in the manufacturers application notes.

Since I go on to estimate roughly how much it is, this comment is redundant. Read the whole post before you start compsoing your response.

So describe a cheaper solution that would work .....

You'd have to pay close attention to the way the cell and the charging circuit fitted together, and work out what the thermal gradients around the assembly were going to look like. Not a job for a beginner, but entirely practicable, if not perhaps all that practical.

The peltier junction is just a thermocouple - though what you buy are lots of thermocouple junctions assembled in series and parallel - and it reacts to the temperature difference between one side of the junction and the other. So you only need one junction to do relative temperature measurments, rather than two thermistors (which measure absolute temperature.

You didn't document what you meant by "too small" temperature rises - as I've done for you - and, since the temperature rise is emminently measurable, the objection was unjustified. You object that the measuring circuit is too expensive, but since C/10 is a tolerable trickle rate for most cells, it is also redundant. Real fast chargers deliver higher currents and have the benefit of producing larger temperature rises.

I wasn't talking about theory, but about standard commercial practice in fast charge circuits. You were talking about practice for dumb newbies, which may be appropriate to sci.electronics.basics, but isn't really appropriate to sci.electronics.design.

------------------- Bill Sloman, Nijmegen

Well, in theory yes. But you should take into account, that outside temperatures _changes_ while charging 14h or so. How will you reliably detect a temp change of 1 degree?

Thats not really a problem, if (and only if) the charge current is high enough to build up pressure in the cell. The charger will shut off after 1 to 5 minutes of charging. I have often used this technique and had _never_ problems with it (of course, again, with sufficiently high charge currents).

Only thing I found is no-go: putting warm/hot cells in the charger again. So I always only charge cells/packs which have cooled down to abient.

Well, sorry, no offence meant. I just wanted to make sure the OP dies not try to do your idea.

..which is really not much.

Oh yes. One can do that. I see no sense in detecting overtemperature with 2 thermistors. Too expensive for such a simple task. But even then, while outside temp changes one might have trouble to cleanly detect this 1 degree you quoted in an everyday usage.

Well, I don't have worked with peltiers....what do you think would be gained using a peltier junction instead of a thermistor?

Now you are make looking me bad. See what I mean? :-)

Because they aren't. I don't say you wrong. You told about theory, which is okay. I told about practice. Really.

Regards, Wolfgang

OK. Nobody has put that fact in doubt :-)

Well, yes. IMHO a users manual should state that fact. But that is only relevant to a charger which uses more than C/10 as charge current. We are not talking about fast chargers here.

Well, yes again :-) Since the OP is asking a question he could easily read in the accu cells datasheet, I assumed he is not very familiar with this kind of stuff.

Nitpick mode eh?

OK, here is it: just use the current stated for trickle charge in the accu cells datasheet. We are _not_ talking about fast charges here...see OP.

Right, but the OP _is_ a beginner, thats the point.

OK, but you have not described the advantage of using a peltier junction.

Why is it? Thats what the OP was asking about!

Of course... but again, we are not talking about fast chargers here.

here it is again....you simply don't want to see the point.

But this is EOT now for me.

Have a nice day, Wolfgang