PPP - Rechargeables

Rechargeables - The unknown species ...

.
.
After releasing Battery Control, I got alot of e-mails concerning the right way to handle rechargeables. So I decided to post a general information page about rechargeables. As a freelancing hardware developer, I designed a special charger for a customer a few years ago. Since then rechargeables became nearly a kind of obsession for me. The prototyp of the charger from then is still in my personal use and I still modify and add different charging algorithms (it's controlled by a 68000). Some of my many experiences I made you'll find below.
.

.Introduction

.Charging rechargeables

.Some more theory

.Poison for your rechargeables

.
Introduction
.
Basically when I talk about rechargeables, I talk about NiCads and NiMHs. Their behaviour is very similar regarding the charge and discharge process. In general one could say, a NiMH is just a better NiCad. One application is still a NiCad center of gravity: High current applications, like in car modell-making. NiMHs have a better capacity/weight ratio, so given a certain size, NiMHs just have a higher capacity. Another major advantage is their (nearly) lack of a memory effect and a lower self discharge rate. Of course there are significant differences in the chemical process inside a NiCad and a NiMH cell, but as long as you don't want to build your own battery, we don't have to worry about that.
One word about renewals: They are no rechargeables from my point of view. They are merely modified Alkalines - so called primary cells - and bascially designed for one time use only. It's still unclear whether some normal Alkaline brands on the market are not also 'renewable'. Anyway, the handling difficulties, the relativ expensive chargers, the short lifetime and last not least the price versus a 'true' rechargeable makes them a bad choice. The only reason for renewables would be their higher cell voltage, but especially the Palm doesn't need that. It's working at a voltage range of 2-2.6V as well as at 3V. In fact, due to the low overall consumption, the Palm is predestined to work on rechargeables.
And before you ask: Lithium rechargeables are also a total different story. They have an excellent capacity/ weight ratio and they are damn bitchy with their handling. Charging them without some sort of 'intelligence' is not possible (well, it is ... one time), they're temperature sensitive and they are very expensive. Last not least, a single Lithium cell has a nominal voltage of 3V. So rechargeable AAA's will be very unlikely.
.

.
Charging rechargeables
.
Now, the first thing you have to do, is to charge the cells properly. Basically that's very easy: Take a *constant* current source of 1/10C (C is the nominal capacity of your cell) and apply it for 14 hours at room temperature, if you know the cell is empty. Please note that a resistor is *not* delivering a constant current, because the voltage of the cell doesn't stay constant. A really neat trick is the following:
.

If you're familiar with a soldering iron and you play around with electronics from time to time, you will probably know the fixed voltage regulators of the type 78XX, where XX stands for the voltage they'll output. They cost not even a dollar and you'll get them easily. It's very little known, that they make also an excellent constant current source. The input pin (1) remains standard, but instead of using the output pin (3), we draw the current from the normal ground pin (2). NOTE: The whole circuit has no ground connection. It's used like a resistor, the negative lead is directly connected to the battery.

There are some things to consider: Your input voltage *must* be much higher than your battery voltage. First you have to compensate the nominal regulation voltage, in the case of a 7805 5Volt (you can also use a 7809 or 7812, but then you need 9 or 12V only to overcome that). Then, the internal circuit 'eats' about 2V of semi-conductor loss and you need about 2V/cell. While charging the voltage of a cell goes far above 1.2V. Let's say you want to charge two AAA cells for your Palm, you need 5+2+(2*2)=11V at least. So a cheap 12V plug-in power-supply is perfect. The resistor for 50mA should be around 110 ohm *AND* it must be able to dissipate at least 1Watt. You have only a 15V power supply? No problem, the current remains constant until 35V input voltage (then you blow the 78XX), *BUT* the power dissipation raises. So everything gets a bit hotter :) and you might need to cool the 7805 and take a more powerful resistor. The same applies, if you need more current. Making the resistor smaller increases the current, taking a higher value decreases it. NiCad users with 250mAh rechargeables should try it with 220ohm. A last note: Pin 2 is electrically connected to the heat sink of the 7805, so take care when you fix it to some metal parts! And don't leave the batteries connected to the charger when it's not supplied. Depending on your input voltage power supply, the batteries might be discharged again via the reverse path.
.

.
Some more theory
.
The above way to charge a battery it for sure the cheapest and easiest, but it also has alot of draw-backs. The first is the charging time. If you keep two sets and use them only with your Palm which runs easily days on a set, it's of course no problem when the charging takes 14 hours. But using the batteries with a cell phone for example is another story. The first thought of course is, to raise the charging current, i.e. to 1C. Then the charging process would take theoretically only 1 hour. But the risk of overcharging is much higher and the rechargeables have to be able to convert the electrical energy to chemical energy fast enough. During the charging process, oxygen is produced in the cell at the positive electrode. As long as the charge current is low enough, the oxygen is absorbed again at the negative electrode. If the oxygen can not be absorbed fully, the pressure in the cell raises, the temperature raises and in the end ..., well all cells have a security valve to release the pressure gracefully. But you can trash the cell afterwards. The same happens when the cell heavy overcharged or the temperature is too low.
An 'end-of-charge' recognition would help. A relative bad technique is, to check the absolute voltage of the cell. It's true, the cell voltage increases with increasing charge, but different cells can have different end voltages. You had to consider also the influence of the charge current, which raises the cell voltage, but the amount depends on the current.
Fortunately NiCads and NiMHs show a nice effect when they get full, that is, the voltage decreases again after reaching a maximum. Chargers using that effect are called delta-peak chargers. On the graphic below you can see a typical voltage curve (the blue one) during a charge process of two NiMH cells:
.

BTW, this delta-peak is the major difference of NiCads and NiMHs. The 'peak' of NiMHs is much lower than the one of NiCads. So a charger suitable for NiMHs must have a better A/D converter to detect that peak. Using a bad NiCad charger for NiMHs can be fatal, though most chargers have an additional time limit to avoid overcharging due to an overseen peak.
.
What you see above, is the result of a very sophisticated process. One condition for a true voltage graph is that the charge current is switched of for a short moment (15mSec) while recording the voltage. That produces comparable graphs even when using different charge currents. To be able to use a current up to 4C (15 minutes charging time), another trick is necessary. The whole charging process is devided in 1 sec. portions. In every second, there are 3 phases: Phase 1 takes 980mSec - so most of the time - and the charge current is applied. Phase 2 takes 5mSec only and applies a relativ high *discharge* pulse (2.5C). Finally Phase 3 takes 15mSec and meassures the battery voltage in idle mode. The trick is phase 2. The discharge pulse helps immense to reabsorb the oxygen produced during the high charge current phase. Additionally it 'reconstructs' even older cells by 'smashing' crystals that grow too big during normal use. These crystals decrease the reaction surface of the chemicals inside the cell, such decreasing the ability to convert energy back and forth. The loss of charge with that pulse is minimal due to it's short duration.
Together with that so called 'reflex-pulse' the charge efficieny increases also enormously. That means, when you charge the batteries 65 minutes with 1C they're full, and only 5 minutes of energy are lost in heat, pressure etc., all things a rechargeable doesn't like very much.
In the graph above you can see also 3 discharge curves, which show the result of an refreshing process of old cells that were used with 'normal' chargers for half a year. Not only the overall capacity increases after 3 cycles, but also the voltage remains higher over the complete discharge time.
.
Ok, when you read all of the above and you do your own electronic experiments, you could build your own intelligent charger. You need a programmable current source for charging, a programmable current sink for the reflex pulse, an A/D converter to meassure the cell voltage and a processor to bring that all together :). But for the little AAA's for the Palm, it would be a bit of an overkill. My advice is to buy a good charger or, if you want to build your own one, look for ready available charging processors. Maxim and Linear Technologie for example offer a wide range for all kind of applications and battery types.
.

.
Poison for your rechargeables
.

Never mix the cells you use. When you buy i.e. 4 new cells as two sets for your Palm, mark them from the beginning on and use them always as same pairs. If one cell is half full and the other one is full and you use them together, the half full cell will be 'reversed' by the other one when it's empty. That is one sure way to kill it. Of course charging cells with different status is as bad as well. While one is still not full, the other one is already overcharged. In fact, good manufacturers that offer battery packs, even match the cells. If you own a volt-meter you can do that also: Just meassure all cells immediately after you bought them (before first time charging) and pair always the two with the most similar voltage.
Never charge a hot cell or charge them long enough to get them hot. It's normal, that the batteries are getting warm, up to 40 degrees C (~100 F). But above that you should definitely shut down something. Especially NiMHs are much more sensitive.
Don't discharge the cells below 1V/cell for longer. Get me right, it's no problem to finish your work for 10 minutes on your Palm working on 1.9V, but constantly discharging them below 1V decreases their lifetime.
Don't recharge NiCads too often when they're not empty. It is no problem from time to time, but doing that constantly produces the memory effect, and the cells won't give you their whole capacity. NiMHs show nearly no memory effect. I recharge them whenever I think I need a new set because I'm not home the next few days. Assumption is, you have a charger that doesn't work with a timed charging process only. Also, an optimal treated NiMH can give you easily 1000 cycles and more (hey renewable users, do you read that ... ? :) ) and they cost about $3.50 / cell. If you don't treat them absolutely optimal and have to change them after 700 cycles ... well, who cares ...

.
.