An Electrics Primer: Building the Perfect Nickel-Chemistry Charger Part I

If you just can't seem to find the commercial charger that fits your needs, or just prefer to DIY, then there is something you can do: read this series of articles and build your own PERFECT smart charger!

Article Thumbnail


Recently I needed a battery charger, so I started shopping. I looked at dozens of commercial chargers, and while each had their pros and cons, I didn’t find one that I really liked. So I did the only sensible thing - I set out to build my own. (Ok, I’ll leave “sensible” open to interpretation).

It turns out that building your own smart charger isn’t as difficult as it sounds. No doubt in response to the increase in battery powered devices, the big integrated circuit chip makers have created special “battery management” microprocessors whose little digital brains incorporate all of the best features we desire in a good, no, a perfect charger. What’s more, these chips don’t require ROM programming or computer interfaces or anything else spectacular and uncommon. Usually a few resistors are all that are required to get it working, and then a ton of LEDs and such to make it pretty.

The perfect charger means something different to everyone. For me, the perfect charger should be fast, capable of charging a pack in less than an hour. It should also have the ability to charge slow, should the necessity arise. It should charge everything from a loose pair of AA’s for my wife’s camera to a 8-cell transmitter pack. It should bristle with every type of connector for every type of pack I own. Most of all, it should be smart, capable of charging my batteries right to capacity and keeping them there until I’m ready to use them.

This article is part one (of two to who knows how many) that I’ve put together for E-Zone on the topic of building your own smart battery charger. In it I’ll discuss batteries and chargers, as this knowledge is indispensable when it comes to making a charger out of fistfuls of electronic ingredients. Forgive me if I spend time stating the obvious, but my hope is that anyone with the desire to build a charger can understand and do so, even if they’re not experienced with electronics. Actual charger construction will be the focus of following articles in this epic series.

Know you batteries, know your chargers

I consider myself a pretty smart guy, but when I first started looking into batteries and chargers I must admit, things were a bit mystifying. I mean, what just happened to just having AA’s, AAA’s, C’s and D’s, from Energizer or Duracell? So, after some serious research with the help of Google, I unveiled the mysteries of the rechargeable world. Lucky for you, I decided to combine that knowledge and share it right here on E-Zone. But enough ado, let the de-mystifying begin.

NiCds, NiMHs, and LiPolys (oh my!)

Nickel Cadmium (NiCd), Nickel Metal Hydride (NiMH), and Lithium Ion Polymer (LiPoly) batteries are the three most popular battery chemistries we put in our planes. Each has their advantages, but before picking on individual differences let’s look at common features among all cells.

Cell sizes: We all know that batteries come in some package size. The standards are AA and AAA, though others tend to crop up, like N-sized, which are roughly AAA-sized in diameter, but half as long as a AA. There are even a whole host of fractional sizes, a phenomenon that can only be blamed on the multitude of cordless phones we own. Table 1 has a few of the more common sizes, and has an exhaustive list.

Table 1: Common cell properties
Name Diameter (mm) Length (mm) Weight NiCd/NiMH (g)Com. Capacity (mAh)
AAA 10.5 44.5 10/13 300-800
AA 14.2 50 21/27 1000-2000
N 11.5 28 9/11 200-300
2/3AA 28.7 30 12/15 250-350

Nominal voltage: The next thing to know is that each type of battery has a nominal cell voltage that is completely determined by its chemistry. Good old alkaline batteries, like Duracell and Energizer, are 1.5 Volts. NiCd and NiMH cells each have 1.2V nominal, though as we’ll discuss later, this changes as the cell charges or discharges. LiPolys boast 3.7V per cell. By combining multiple cells in series we add the voltages, yielding a battery with a voltage total voltage of (nominal voltage) x (number of cells). Hence a 4.8V flight pack is four, 1.2V NiCd or NiMH cells. (Oh, and from now on I’ll try and say “cell” if I specifically mean a single cell and “battery” to denote any number of cells in series. I’ll also say “nominal voltage” if I mean the general cell voltage, and “terminal voltage” if I mean the voltage across the cell under specific conditions, such as fully charged, fully discharged, or under load).

Charge capacity: Every cell has a specific charge capacity, most often rated in milliAmp-hours (mAh). The milliAmp-hour is a unit of charge (1mAh is actually 3.6 Coulombs), and makes determining charge or discharge time really simple. Take, for example, a 700mAh battery. Such a battery can supply 700mA of current for one hour, 350mA for two hours, or 2.8A for fifteen minutes. Come charging time, it would likewise take a 1.4A current supply about half an hour to charge the same battery to full capacity. Remember that capacity is a property of each individual cell, so putting two 1000mAh NiMH cells in series creates a single, 2.4V, 1000 mAh battery (NOT a 2000mAh battery!). Should you mix together a 700mAh cell with a 1000mAh cell (a bad idea), the effective capacity of the pair becomes 700mAh - the smallest rated capacity - and the 1000mAh cell will never fully discharge. While a wide range of capacities exist, table 1 includes typical capacities for common cell sizes.

To summarize the three things we’ve discussed so far (which is also what you’ll see when you go to buy a battery), the battery basics are type of battery (chemistry), nominal output voltage, and capacity. The voltage increases with additional cells and the overall battery capacity is that of the smallest cell capacity, though we should always use cells of alike capacity.

There are a few other characteristics of batteries that don’t tend to make the label, but that we should understand. One is the internal, parasitic resistance or output impedence that exists in all power supplies, including batteries. A perfect battery (fig. 1a) should be able to supply arbitrary current to any load while maintaining its voltage. That is to say, you should be able to hook a voltmeter up to a perfect 1.2V cell, attach a variable resistor R ( Rload in figure 1) to its terminals, and see that the cell terminal voltage is always 1.2V regardless of the resistor setting, and that the current can be found by (1.2V)/R. In reality, batteries look more like figure 1b, and the terminal voltage tends to baulk when we demand high currents of them. The terminal voltage of the real battery is Vterminal = 1.2V R/(R + r) and current through the load resistor is (1.2V)/R + r. Another way of saying this is that the current starts to decrease from its ideal value when the load resistance R is less than or equal to the internal resistance r. As fate would have it, this is pretty much the case in electric flight.

There is one thing to which I must admit now that I’ve mentioned cell capacity and internal resistance, that being inefficiency. Sure, a battery stores charge, but what it’s really storing is energy (we could even loosely say that the energy in a cell is QV/2, where Q is the charge and V the terminal voltage). That internal resistance is a source of energy loss, and some of the power that could be delivered to the load will be burned off as heat because of it. This is why we don’t tend to get the labeled capacity out of a cell, especially at discharge rates greater than 1C. For example, instead of a 1000mAh cell sourcing 4A of current for fifteen minutes, we may only get ten or twelve minutes out of it.

When we “match” cells in making a matched battery pack, our primary goal is to make sure that the internal resistance of all cells are alike to within a few percent. Doing so ensures that all of the cells will charge/discharge at the same rate, no one cell gets any hotter than another, and we avoid an ill-fated cell reversal (to be discussed shortly).

If you’d like to match your own cells in this manner, you need only find out the internal resistance of each cell individually. To measure the internal resistance (fig. 2) of a cell you need a voltmeter, Ohmmeter, some alligator clips, and a few low-resistance power resistors (around 1 to 5 Ohms) or a few inches of fine gauge tungsten or steel wire (tungsten and steel wires tend to have a resistance of about 1 Ohm/inch for 30AWG) to play the role of load resistance. First hook the voltmeter to the cell and record its unloaded voltage. Use the ohmmeter to measure the resistance of a small length of the wire or a power resistor, record the value, and then remove the ohmmeter and attach the battery to the same length of wire. Now loaded, the voltage across the cell should drop slightly - record this loaded voltage. Knowing the unloaded voltage, loaded voltage, and resistance of the wire, internal resistance can be found from the equation:

r = R (VUnloaded/VLoaded - 1)

Repeat this experiment a few times for each cell, using a different resistance R each time, and take the average to find a more accurate r. Oh, and don’t leave the cell attached to the load wire longer than you need to record the voltage, as you’re draining the cell fast which will skew your results.

The last battery characteristic worth knowing is the allowed charge rate. We’ll discuss the charge rate ad naseum later, but suffice it to say that a charge rate of 2C to 4C is pretty common. Anything less that 1C (i.e. C/2., C/4, or less) is a crummy cell that will take a long time to charge, and the so-called “fast charge” batteries that claim charge rates of 10C are pretty impressive (10C is a charge time of six minutes).

Now that we’ve discussed general features, let’s look at battery specifics.

NiCd batteries up close and personal

I’m sure that Nickel Cadmium batteries hold a special place in all of our hearts, because they were the first rechargeable cell. Historically speaking, as a replacement for Alkalines they were abysmal, since their terminal voltage is only 1.3V at full charge and decreases slowly with discharge, often manifesting itself as what appear to be a short charge life. But we modern flyers know better (now), and have designed our receiver and motor systems to account for this, making NiCds a fiscally superior way to keep our planes aloft.

The actual chemistry involved in NiCads is really quite simple. The cell is composed of a nickel oxide (NiO2) cathode, a cadmium (Cd) anode, and potasium hydroxide (KOH) electrolyte (isn’t it ironic that a rechargeable battery’s ‘acid’ is really a base?). During discharge the cadmium oxides, becoming Cd(OH)2, and the nickle reduces, become pure nickel metal. When charged, the current forced through the cell converts the Cadmium back to its metallic form and the nickel to its oxide form, restoring the available energy. The anode is actually a mesh coated with microcrystalline cadmium. When the cadmium is not returned in its microcrystalline form we have what is often erroneously called the “memory effect” (to be discussed in the “When bad things happen to good batteries” section).

On a more practical level, NiCds can be recharged roughly 1000 times, have a shelf life of five years, and have a gravimetric energy density (that’s fancy speak for how much charge the cell holds divided by how heavy it is) of 45 to 80 Watts/kilogram. These cells self discharge at a rate of about 1% per day, which is why a battery is mysteriously dead after a month of neglect.

NiCds are best suited for the high-current demands needed in electric flight (and subsequent fast charge). They have the smallest internal resistance of any rechargeable, typically less than 0.1 Ohm. They are often hailed at being able to output current up to 20 times their capacity (so a 750mAh battery could source 20A for about three minutes), and for the same reasons lend themselves to high-rate fast charging.

Looking NiMH in the face

Nickel Metal Hydride batteries are newer to the rechargeable scene, and have gained popularity for one very good reason: increased energy density. With a gravimetric energy density of 60-120 W/kg, NiMH batteries have roughly a 50% increase in capacity over NiCds. They also lack toxic Cadmium, though some argue that there is so little cadmium in NiCds that the difference is unimportant, especially considering that it is recommended that both NiMH and NiCds be properly recycled. While there are those who claim NiMH can exhibit a “memory effect,” they had to try very, very hard to prove it.

NiMH batteries are not without their problems, though I expect time will cure them. While the chemistry is very similar, the major difference in construction is the replacement of the cadmium anode with a metal hydride mesh capable of absorbing a high amount of hydrogen. This new anode is very susceptible to damage from overcharging and has roughly half the expected recharge cycles at 500. These batteries have a shelf life of five years and self discharge at approximately 1-2% per day. Internal resistance is greater than NiCds at around 0.25 Ohms, and discharge and charging rates are also reduced, often to about a maximum of five times capacity in both cases.


LiPoly batteries have gained significant popularity because they have the highest energy density of any small, rechargeable battery, typically 100-130W/kg, and they can be found in just about every cell phone and laptop computer out there. Their chemistry is completely different than the Nickle-based cells. Every LiPoly cell is 3.6V nominal, so you don’t need many cells to get a reasonably high voltage. Unfortunately their chemistry makes them incompatible with most NiCd/NiMH chargers (their terminal voltage rises rapidly, and the charging current must be slowly reduced until they are full). LiPolys don’t come in the standard cylindrical sizes (I assume this is because of the incompatible voltage, unique charging techniques, and the fact that every cell phone is shaped differently!).

When bad things happen to good batteries

Most of the time bad things happen to batteries only when they are overcharged or overdepleted. Overcharging is the application of fast-charge currents after the battery is full, so trickle-charge currents don’t count. Overdepletion is when the cells are drained below an accepted level. For NiCd and NiMH batteries, this is 0.9V per cell (that is, a cell with a terminal voltage of 0.9V is a dead battery). There are a number of bad things that can happen to batteries that (thankfully) don’t involve them bursting open and spilling their electrolyte everywhere. Here’s a short list of what to avoid:

Cell reversal

Cell reversal is when the terminal polarity reverses or falls to zero and the battery will not hold charge. This usually happens in one of two conditions. The more common is when a sufficiently large voltage is applied backwards across the terminals (i.e. a positive voltage to the negative terminal and a negative voltage to the positive). This is the case when one cell in a pack has a lower voltage than its neighbors: the cells on either side actually force the terminals to switch polarity. Less common is when the cell must discharge at a very high rate, similar to a short circuit, and has the same symptoms. A reversed cell will usually register 0V across its terminals, and can sometimes be resuscitated by charging it very slowly.


Overdepleting a cell creates an imbalance of chemical ratios, throwing off its chemistry making it difficult to recharge. A NiCd or NiMH battery with a terminal voltage of 0.9V is considered depleted. Typically an overdepleted can be restored by first charging it slowly, by a rate lower than C/4, until 0.9V terminal is reached, when it can be fast charged. Overdepleted cells are also more prone to reversal.


Overcharging NiMH batteries at a fast-charge rate is a really, really bad idea. Their anodes become pretty badly damaged, and a lot of unwanted chemicals are produced that cannot be removed from the battery. NiCd batteries are also damaged, but not to the extent of NiMH. Once a cell is fully charged, it starts to produce more heat, which is also bad. A NiCd can typically be trickle-charged at a rate of C/10 or less without damage, while NiMH trickle charge rates are lower, usually C/25 to C/50.


Overheating is a generally bad thing. It is usually the product of unwanted chemistry and allows other unwanted chemistry to occur, and it could damage some of the battery innards. If a cell gets too hot it may actually rupture, spilling electrolyte all over your model.

The NiCd “memory effect”

I only call the so-called “memory effect” by that name because that’s what everyone else calls it, even though the name is completely erroneous. As the story goes, you charge and (partially) discharge, charge and (partially) discharge, charge and (partially) discharge, and all of the sudden it seems like your battery doesn’t have the longevity it used to have. Almost like it has started to remember the partial discharge, and thinks that charge state is the new definition of empty. Well, here’s news for you: batteries don’t have memories, and they can’t think.

The “memory effect” is really just the result of the microcrystalline cadmium at the anode clumping together. Perhaps you’ve run a battery until you thought it was completely drained and then let it sit for an hour or so, to find that it found a little extra kick during its rest. What has happened is, after the last charge, some of the cadmium formed together in large clumps at the anode. Once all of the microcrystalline cadmium (and the surfaces of the clumps) was converted during discharge, these clumps of un-hydroxiginated metal hold the remaining battery charge, but their surface area is insufficient to maintain the desired current. The results are manifest in the symptoms we assume are a dead battery (low current, waning terminal voltage, and high internal resistance), but they are actually just a loss of efficiency. Waiting an hour or so allows some of these clumps to break apart, as they are wont to do given the low cadmium concentration in the battery’s depleted stated, and hence the brief second wind.

Conditioning is what is necessary to combat this effect. The role of conditioning is to place a very small demand on the battery, only drawing tens of milliAmps or so, until the cell terminal voltage reaches 0.9V. It is important that the load be light such that the current is small such that the cadmium will dissolve on roughly the same time scale (thus not giving a false, loaded voltage reading). The 0.9V terminal voltage is what is popularly considered a depleted cell, and going below this could lead to cell reversal. Now simple charge the battery back to maximum capacity and it is as good as new!

I’ve read many possible reasons that cause this effect. The two most believable are overcharging and the one unavoidable, the passage of time. Overcharging a battery causes some unwanted chemistry, such as the release of oxygen and hydrogen in the regions around the anode along with plenty of heat, and some spurious reactions can lead to the clumping effect. As for time, well, repeated charging and discharging will take its toll. If you’re a regular NiCd user, expect that you’ll want to condition your batteries every thirty cycles or so.

On a side note, I’ve read a few explanations of how this phenomenon was dubbed the “memory effect.” The most plausible was a NASA spin-off. Apparently early satellites, which undergo very regular charging and discharging cycles due to their orbits, started to exhibit bizarre battery behavior while on the dark side of the earth. Try as they might, the scientists couldn’t reproduce the effect in the lab, and the name was born (and then incorrectly applied to our problem).

All of the chargers under the sun

Charging is a relatively simple concept: you force current back through a battery until its chemistry has been fully restored, then turn off the current. The trick is deciding what determines if the battery is ‘full,’ and then implementing a million little features to make the charger user-friendly (read: lights, readout displays, options, options, and more options).

The heart of every battery charger is a power supply. The better supplies are constant-current supplies, but standard, constant-voltage supplies are not uncommon. Sometimes these supplies have some sort of additional electronics that monitor the state of the batteries under charge and make decisions on how to charge. We call these “smart” chargers, but there are many different things that can be monitored, so smart can mean many different things. We’ll talk about power supplies in general, as it is important to building your own charger, but first let’s do things out of order a bit and look at the features one would see when shopping for commercial chargers.

Characteristics of Chargers

The two most basic characteristics of a charger are its output current and output voltage. The output current is the most important rating of all and rated in terms of current (Amps or milliAmps). A good charger should be able to sustain a maximum 1 Amp or more of current, while cheap, crummy chargers tend to only produce a few hundred milliAmps.

The amount of charging current use in is related to the cell in terms of a charging rate:

Charge rate = Battery capacity (mAh) / Charge time (hr)

We’ll often see this as a charge rate described by something like “C/2,” ”1C,” or “3C.” A C/2 rate indicates that a battery would be charged in two hours, 1C in one hour, while 3C is a 1/3 hour (twenty minute) charge time (remember that the number in the denominator is the charge time, so 3C = C/(1/3)!). The charging rate is really a unit of time, not a rate (which would be 1/Time, but perhaps I’m being too picky).

The output voltage is less important, as long as it is greater than the fully charged terminal voltage of the battery. That is to say, a charger with a max output of 5V can’t charge a 7.2V battery pack (in fact, it would be hard pressed to charge a 4.8V pack), since each cell may reach as much as 2V terminal during charging.

Charging is only a 60-80% efficient process, so the actual charge time will have to make up the efficiency overhead. Some people calculate this efficiency into the charge rate, so an 80% efficient charging of a 1000mAh battery at 1A of current would actually take 1.2 hours, not just one hour. These folks like to increase their charging current to keep the charge time constant (i.e. they would use 1.2A for 1 hour to maintain the charge rate of 1C). I do not subscribe to this theory because I don’t like to play with the charge current (increasing it further reduces the charging efficiency anyway, so you end up in an geometric cycle of re-calculations!), so just expect that, for instance, a C/2 charge rate will actually take a little more than two hours to complete.

One other thing to realize about charging is that it always increases the terminal voltage. If a full battery has a nominal terminal voltage of 1.3V, then we can expect this to increase to around 1.4V if we force 100mA through it, or to reach an even higher voltage at higher currents. The terminal voltage during charge depends both on charging current and charge state (i.e. how full the battery is), so there’s a little voltage overhead involved. A cell may get as high as 1.8V or 2.0V during fast charge when it’s nearly full, which is why I earlier said a 5V charger can’t effectively charge a 4.8V pack (the four cells in a 4.8V pack would need 8V or more).

Charger IQ: What makes a charger “smart?”

A charger is made “smart” when it can switch from a high current, fast-charge mode to a lower current trickle-charge when the battery is fully charged. How the charger determines when the battery is full is the critical phenomenon. Common methods include peak-voltage detection, negative-voltage slope detection, temperature detection, or some combination thereof. This is one of my pet peeves about commercial chargers. A manufacturer will use “smart charger” in place of actually describing the method used. When ‘smart’ means maximum voltage detection, which (as we’re about to discuss) is not very smart at all.

The classic AC/DC adapter (dumb)

There are a myriad of chargers available, and we all own at least a few, so it wouldn’t hurt to understand how they all work. The simplest, cheapest, and “dumbest” charger of all is a plain old wall-cube AC/DC adapter (fig 2). This charger is merely a power supply that outputs a fixed voltage and a few hundred milliAmps of current as long as the battery is plugged in. These are the worst chargers of all because they require that WE monitor the charging time closely to prevent overcharging. To make matters worse, we must know how depleted the pack is before we start charging to do so! If you’re the type that likes to watch grass grow, you can occasionally test to see if you batteries are warming up as a rough test to see if they’re charged. I just cut the grass when it gets tall, so I avoid these lousy chargers like the black death.

Timed chargers (mostly dumb)

A step up from the plain wall-cube is the timed charger, which will output a fast charge current at some voltage for a known amount of time (say, a few hours) and then switch to the lower trickle charge current. (Technically these are smart chargers under my definition, but they’re really only smart in the Scarecrow from the Wizard of Oz sense, in that he could walk and talk but didn’t really have a brain). These very popular chargers are usually described as “4-hour” or “8-hour” chargers, and often have one or two LEDs that indicate if the timed cycle is finished. Though an improvement over simple AC adapters, these chargers are also undesirable. For starters, their timed fast charge rate applies to only one battery capacity. For instance, a 4-hour Duracell charger I own is meant for 1800mAh batteries (a C/4 charge rate, or 450mA), so a 2200 mAh battery will only be 82% charged after four hours, while a 1100mAh battery will be badly overcharged. They are meant to only charge completely discharged batteries, and in the event that their charge cycle is interrupted or the battery is not completely discharged, battery will be overcharged.

Maximum voltage chargers (moderately dumb)

An ideal maximum voltage charger would fast-charge a battery until the battery terminal voltage reached some pre-set level, usually around 1.3V x number of cells, and then switch to trickle charge. Usually they don’t work quite this precisely, and their fast charge current slowly decreases as the battery charges, due to the increase in battery terminal voltage, with the minimum current occurring when the battery attains the aforementioned pre-set level. These chargers walk a fine line between smart and dumb, but I’ve categorized them as dumb because I’m a tough sell. For starters, this method only works for rather low charging currents - a few hundred milliAmps at most. Any more current and the battery’s terminal voltage will reach the pre-set limit before it’s actually charged. Since the current decreases from this already low start point, these can be very slow chargers. Their only real elegance is that the require very few components to build, and they are unlikely to overcharge a battery.

Peak voltage detection chargers (smart)

The most popular charger that can be considered truly “smart” is the peak detecting variety. These are sometimes called Delta-V/Delta-t chargers, meaning it watches for a change in the slope of voltage vs. time (i.e. it looks for a peak!). Just as the battery voltage slowly decreases while discharging, it slowly increases while charging. When a NiCd battery is fully charged, the voltage drops appreciably (around .05V per cell), creating a definite peak in the voltage. When a NiMH battery reaches about 95% capacity, its voltage plateaus and then decreases ever so slightly. The peak-detecting charger notices when the terminal voltage levels off and switches from fast-charge to trickle charge, as shown in figure 3a.

A peak-detecting charger can charge any battery, regardless of is state of depletion, to maximum capacity in a minimum amount of time. No more sitting around, setting the egg timer, waiting for the cells to start warming up to unplug the charger! The also work universally with NiCds and NiMH batteries.

Negative-slope detection chargers (smart)

The deference between negative-slope detection and peak detection is rather slight, almost semantic. A negative-slope detector actually looks for the decrease in battery terminal voltage, so it will charge a bit longer than a peak-detecting charger. Usually this is only a difference of a few minutes. Justification of the existence of the negative-slope detection is to get that final few percent of charge in NiCds. Since the decrease in slope of a NiMH battery is so small, negative-slope detection is not recommended. Figure 3b shows a negative slope detecting charger at work.

Temperature detection (moderately smart)

The other common type of smart charger is a temperature sensing variety. The temperature of a battery also increases as it charges, usually reaching 35-40 degrees Centigrade when full and continuing to increase if overcharged. The temperature is usually measured with a simple negative temperature coefficient (NTC) thermistor. These devices are little resistors that experience a drop in resistance if warmed, or an increase in resistance if cooled, make for fast and cheap thermometers, and are perfect for circuits which generally require some sort of transducer (i.e. a device that turns temperature into voltage - in this case, a change in temperature becomes a change in resistance and, under constant current bias, a change in voltage).

Temperature detection is far less popular than peak detection because it depends on the charge rate and the weather, and at least one of those things is difficult to control! When charging at a high rate (>2C), more heat is created by the internal battery resistance, potentially causing premature fast charge termination. On a warm day the batteries cool less effectively with the same result. There are many other reasons that make temperature detection less than an exact science and unfit as the primary means of smart charging, but it is still very useful as a fail-safe to keep batteries from overheating during the fast charge cycle.

The best charger of all combines many of these. For instance, assume that peak detection is the primary means for determining when a battery is full. Adding a timer function can prevent overcharging in the event that the peak-detection electronics misses the peak. Also adding a temperature monitoring circuit can stop charging if the batteries become excessively hot.

Now you’re probably wondering, how does one accomplish peak or temperature detection, or even simple timing? Well, there are probably analog circuits that can perform all of these to some extent, but all three processes benefit greatly from digital circuitry, specifically by dedicated battery management microprocessor chips. Now don’t fret when I say microprocessor - we won’t be doing any programming or anything! Microprocessors can record the battery voltage or temperature at regular intervals and compare them. I doubt I’ll ever turn a true microprocessor into a battery charger - it’s too much work! But chips such as the Maxim MAX712 and MAX713 or the BenchMarq bq2002 and bq2004 from Texas Instruments. These little wonders make designing a simple smart charger almost trivial!

Power supplies and battery chargers: kissing cousins

As I mentioned before, the heart of a battery charger is a power supply: the electronics that add smart features are just icing on the cake. Often the battery management chips I’ve mention can be configured to basically control a power supply. If you go on to read my following articles and want to build your own charger, then the choice of power supply is critical.

The real power supply question is do you buy a cheap adapter for $5, a full-fledged regulated supply for $50, or go ahead and put together your own, seeing as we’re in the design and construction business? I tend to lean toward the cheap ideal, as I’m poor and power supply design can be a complicated business (and coming from a guy who insists on building his own charger, that’s saying a lot!).

The supply ratings needed are pretty simple to figure. The output current should be equal to the highest fast-charge current desired. For instance, if you want to charge a 1000mAh pack at a 2C rate, you’ll need a two amp supply. The minimum supply voltage is determined by the highest cell count you want to charge, figuring about 2V per cell and an additional 2V overhead. Thus a 8-cell transmitter pack would require 18V minimum, but it doesn’t hurt if this number is a little higher.

We ultimately seek a constant-current supply, as it is the best solution to battery charging. These can most precisely maintain a fixed charging rate without interfering with peak detection schemes. The charging current is usually set by monitoring a sensing resistor, as seen in figure 4. The control circuitry, which is usually a battery management chip, maintains the voltage across the sensing resistor to some preset value by increasing or decreasing the charging current. The voltage across the battery terminals is whatever naturally occurs, determined by the current being forced through it. Thus we have unfettered implementation of peak detection at a well-established charging current. Of course, constant voltage supplies are much more common, so we convert one of these into a constant-current supply with the addition of a single transistor. I’ll be using this method in my designs.

Before launching into the different types of supplies, I may as well admit that I’m a fan of the switching supplies, which I discuss last (for climactic drama, of course).

Unregulated and linear regulator supplies

There are three types of power supplies: unregulated supplies, linear supplies (regulated), and switch-mode power supplies (regulated). I’ve mentioned unregulated power supplies before in this article - we called them el-stupido wall cubes or AC/DC adapters (figure 2). These devices have beauty in simplicity and high efficiency. Considering they’re made from only a handful of components, can be found for a range of voltages and cost less than $5, they tend to be the first choice for most manufacturers of rechargeable electronics. All of their sophistication lies solely in the transformer. The transformer must be able to step the voltage down from 110V mains voltage to something slightly greater than that of the battery and provide the desired fast-charge current.

This probably isn’t the best place for a discourse on transformers, so I’ll keep it simple: the higher the current and voltage you want, the bigger and heavier the transformer becomes. This is typified by a transformer’s “VA” rating, which is short for Volts x Amps, both RMS values (I know, why didn’t we call it a VI rating and describe it in Watts?). Say you want a 24V output at 4A (RMS) of load current - that’s a VA of 96. Most of the transformers found in wall cubes are small, weighing less than a pound, and have a VA of around 30, and rarely more than 50. The RMS output voltage of these supplies also drops with increasing load current and the peak-to-peak ripple increases, both undesirable qualities, as seen in figure 5.

Since it is difficult to find high-current transformers above about 15V, making unregulated power supplies unattractive for this particular application. I’ll give these their due attention when designing a charger, but realize that they’re not going to output more than 1A at around 20V.

The next type of supply, the linear regulating power supply, starts with an unregulated power supply and then uses some semiconductor device to clip off the unwanted ripple, yielding a nice, clean, regulated output voltage that is relatively independent of load current. These chips are cheap (about $3), are trivial to implement, and require at most a few external components. Their biggest drawback is their terrible efficiency. All of the power that is “clipped off” is simply thrown away, discarded as heat. This means that you need a big heat sink to dissipate all of that waste, and it is in your best interest to find a transformer that steps the voltage down to nearly the desired output to minimize the heat produced.

I’m not a fan of linear supplies for the same reason I don’t like unregulated supplies - they need a big transformer to produce the necessary current. What’s more, most battery management IC’s are made to work with unregulated supplies, so making a unregulated supply a linear is unnecessary. My one exception to this rule is using a linear regulator chip to provide the power for the other IC’s in the charger: that is, pick the supply needed to charge the batteries, which may be 20V or more, and then add a 5V linear regulator for all of the peripheral 5V electronics.

Switch mode power supplies (SMPS)

The final supply is the switch-mode power supply, or “switcher.” Switching supplies are of a very different construction than the previous two discussed. For starters, they generally rectify power right off of the mains, creating about 80V DC! Then they have a big switching transistor and some driving/feedback circuitry that chops this DC into pulses at around 30kHz. The pulses drive a rather small transformer (by comparison to the linear types), and that output is fed into an inductor that stores energy. These are sometimes called DC/DC converters, because they take DC input and provide DC output at a different, regulated voltage. The supplies can produce very large currents at almost any voltage (up to many hundreds of Watts), and they are about 80% efficient, which is a huge improvement over linear supplies.

If you are really serious about a charger with high currents and plenty of voltage overhead, then you should look into buying a switcher. These are the supplies found in PC’s, though their 12V rail won’t be enough voltage for anything more than a five-cell battery pack. The best solution is an external “wall plug” or “desktop” switcher, like those used for laptop computers, some printers, or flat-panel monitors. I’ve found a few cost-effective choices on the web, such as a $6 supply dual output supply rated at 5V, 1.5A and 24V, 3A from All Electronics ( or a 24.1V, 2A model from the Electronic Goldmine (

As far as I’m concerned, an inexpensive switching power supply is the way to go. Their drawbacks include electrical noise, which shouldn’t bother a charger made to work with unregulated supplies, and sometimes minimum load currents for high-power supplies (the 5V rail on a PC supply is like this, requiring a minimum load of 100mA to avoid self-destructing).

The battery management based supply

Some truly exotic charger designs put the battery management chip right in the heart of a switching power supply, creating the fabled “switch mode charger,” such as the one provided by Maxim (fig. 6). I suppose that if I were a manufacturer, I might consider an option like this because it combines two processes (the charger design and supply design) into one. But I’m not a manufacturer, and whatever gains there are to be had in beauty, efficiency, and cost are quickly lost in complexity of design.


To sum up, here are the basic steps to take in dreaming up your perfect charger:

Figure out what kind of batteries you’re likely to charge, including their capacity and total voltage.

  • Decide what kind of smart features you want to include, and which of these will be the primary feature and which will be secondary.
  • Determine how fast you want your fast-charging to be, and find a power supply that can meet this criteria.

Once you’re armed with this, you’re ready for my next article: building the perfect charger with the MAX712.

Thread Tools
Feb 02, 2005, 11:11 PM
I found the discussion of various battery chemistries inadequate, inaccurate and far to generalized. Specific cell types of each chemistry need to be addressed. While it may only be a typo, the units used for energy density were W/Kg this is actually power density. Similar units for energy density are W-H/Kg. The representation of NiMh cells at 60-120 W/Kg (sic) vs. LiPoly cells at 100-130 W/Kg (sic) is not consistent with the marked superiority of LiPoly cells. I could not find a NiMh cell that exceeds 85 W-H/Kg. I found a number of LiPoly cells that exceeded 160 W-H/Kg. This is consistent with the often stated generalization that LiPoly technology has twice the energy density of NiMh and NiMh technology has twice the energy density of NiCads. It is my belief that the readers would be better served by either an omission of this section of the article or specific data on more specific cells. It should have been noted that the energy density decreases with increasing delivered power and cells with higher power capability often have lower energy density.

I don’t understand why so much of the article was used to address the NiCad memory effect. This is mostly irrelevant to even users of NiCads and of no value to users of the other chemistries which will dominate in the future.

There was a bunch of discussion of power supplies without any mention of the use of a 12 volt battery to power our battery chargers. I often wondered if the writer flies electric powered airplanes.

Overall I found the article a hard slog to find few gems.

Feb 03, 2005, 03:35 AM
Registered User
I found the article very interesting and learned alot. Looking forward to part2 when it comes out.

Feb 03, 2005, 11:49 PM
Thread OP

From the author

Wow - what can I say! The article has been up on the site for only a few days and already has nearly 1000 hits and even some feedback (even if it's not all glowing...). Thanks to all the E-zone editors (especially AnnMarie) for their good work. Anyway, I wanted to introduce myself and let the readership know that I'll check the thread periodically for questions and such.

I'm a physics Ph.D student living in the Cleveland area, working in the field of quantum computing (don't be impressed - sometimes I think I'm a glorified plumber/refrigerator repair man). I'm excruciatingly poor, so if you're the frugal type, you'll find I tend to substitute just about anything I can find for actually using the right part! I wouldn't call myself an electronics expert, but I have a lot of experience in the field and have yet to find a subject that I failed to understand, when I really want to. My motivation for the battery charger was just as I stated in the article - I feel I can better better satisfy my demands by making one than buying many. That and it's really cold right now, and I'd rather stay inside!

Since the question has been raised, I do "fly electric," though not for show or competetion: I do it because I find it fun. That said, I will admit that I enjoy a challenge and would just as soon try and make a cereal box fly as assemble a kit. I have never (nor will ever) hauled a lead acid battery out to the field, and as such my designs will reflect that. I would rather charge half a dozen packs, crash half a dozen times, and call it a day. My wife pretty much nailed that down when she saw the charger prototype (the topic of pending article 2) with the comment, "It's not exactly portable, is it?"

As for the next article, I hope to have it done soon. The prototype is a mess of wires and breadboarding, and I want to better understand the brain of the beast (the MAX712 chip) before going public with the whole thing. So far it is built from an old PC power supply, will charge 2 or 4 cells at a selectable current of .25, .5, 1, 2, or 4 A. With any luck it'll be done soon and my house will not be burned to the ground!
Feb 04, 2005, 12:23 AM
Thread OP

Correction: Energy density

As dtknowles pointed out, the units of gravimetric energy density are indeed Watt x Hours/Kilogram, not Watts/Kilogram as I it appears in the article. I'm such a stickler for units, and it is to my eternal shame (as a physicist) that I admit this mistake. In my defense, I had originally written the units in abbreviated form (WHr/kg), and during a revision decided that there was a greater pedagogical value in writing them out long-form, where I made the mistake. The rest, as they say, is copy-and-paste history.

As for the values for particular cells, I will not apologize for the ones I listed. I knew those values would be obsolete the moment they were published. What's more, they're just guidelines. For a premium you can find cells that far exceed anything I listed. As for LiPolys, it is always the case that the demanded for high discharge rates accompanies a reduction in energy density and visa-verse (viz. the difference in a LiPoly found in a cordless drill vs. a cell phone). As soon as a company creates a 2400mAh battery with a 10C discharge rate and 200+WHr/kg energy density, I'll buy their stock (and while I'm at it, I'll have my cake and eat it too).

Since my training says never make a statement without proof (and I never shy away when I have science on my side), here are a few references that corroborate my statements:

..for starters...
Feb 04, 2005, 10:16 PM

On the units thing, I blame the editor. They are there to protect us from our simple or glaring errors. I also only meant it in a constructive way. I have a lot of respect for Jim and AnneMarie but I would like them to try a little harder to keep the Ezone the excellent publication it is.

On the relative energy density of NiMh cells vs. LiPoly cells. I looked at your text before and your references now and they are inconsistant with my "real world" "RC model" experiences. The numbers I quoted in my response were based on 1C discharge retailers ad data, not really the best source for data but maybe as valid as your references. I am now going to use a set of data from a recent article in QuietFlyer and note this is not the best available LiPoly cell just one for which I have recent (easily available) credible test data. When tested at 10C discharge, 3.5 volts avg., 1800 mah delivered, 62 g for 102 WHr/Hg for Kokam 2000. In the same article a GP2000 NiMh cell (which is probably the best cell of its kind) was tested at 10C discharge, 1.12 volts avg., 1950 mah delivered, 36 g for 61 WHr/Kg. I might argue that a AA sized NiMh cell might have done a little better and the GP2000 can handle much higher currents but these numbers are pretty representative of real RC use.

Feb 05, 2005, 03:07 AM
rpage53's Avatar
As much as I respect Isidor Buchmann's work, his lithium article is years out of date. We've seen more than a 4X improvement in discharge rates of "LiPo" since 2001 when that was written.
However, its important to note that what we call LiPo he would probably call LiIon. A true LiPo battery has a plastic electrolyte without all the organic electrolyte that causes our cells to burn. So even in 2001 he did have LiIon at 160 Wh/Kg

Which is all sideways to an interesting article about Ni batteries. I'm looking forward to part 2.

Feb 05, 2005, 10:09 AM
Originally Posted by jeremynd01
----Selective snips-----

T -I would like to start this out by asking everyone to read this like it was a friendly conversation at the flying field.

J-"I'm a physics Ph.D student living in the Cleveland area, working in the field of quantum computing (don't be impressed - sometimes I think I'm a glorified plumber/refrigerator repair man).

T-Do you work with cryogenics, LN2 or LHe, I did some cryo work at GRC and Plumbrook. I could use a good plumber/rocket scientist sometimes.

J-I'm excruciatingly poor, so if you're the frugal type, you'll find I tend to substitute just about anything I can find for actually using the right part!

T-Been there done that, it was a lot of fun, low expectations, high benefit/cost.

T-You ever look into Thermo/Acoustic power conversion. Once I was working on a pulse tube cryocooler with a thermo/acoustic driver. While I understood the concept some of the issues that affected the efficiency were beyond my grasp. I do believe this field has great promise if more bright people work on it.

J-My motivation for the battery charger was just as I stated in the article - I feel I can better better satisfy my demands by making one than buying many. That and it's really cold right now, and I'd rather stay inside!

T-I understand, I am designing a charger that will charge a large number of single LiPoly cells, I don't know of anyone making a LiPoly charger with 6 outputs. This is to prevent unbalanced packs and to allow the cells to be reconfigured for different packs, ie. buy 6ea. 2000 mah 10c cells and you could have 6s1p 2000 mah pack good for 20 amps/420 watts or 3s2p 4000 mah pack good for 40 amps/420 watts or 2 ea. 3s1p 2000 mah packs good for 20 amps/210 watts or 3 ea. 2s1p packs, etc.

J -I do "fly electric," though not for show or competetion: I do it because I find it fun.

T -The same here, I fly almost every Sunday, "religiously" just for fun and fellowship, good friends at my current field.

J -I enjoy a challenge and would just as soon try and make a cereal box fly as assemble a kit.

T -I am not sure I understood this correctly. As an approach to clarification, I believe that ARF's are assembled and Kits are built. Or do you mean you would rather design and build you own plane even if it performs poorly than build someone elses design.

J -I have never (nor will ever) hauled a lead acid battery out to the field, and as such my designs will reflect that.

T -I assume this means that you walk to your flying field. I drive to my flying field and the lead acid battery I use is the one in my Ford Ranger. I do charge my three lithium packs before I go and sometimes I put a charge in a couple NiMh packs but run two chargers of my truck battery at the field. I can get lots of flights this way and I have never had a problem with my truck battery at the field and if I did one of my buds would jump me. As an aside I have been looking for a big cheap solar panel, that would be a way cool way to run a charger. Maybe one of those flexible rollup panels as a truck bed cover, still too expensive today. Actually sometimes I walk to the park to fly and back before LiPoly packs I would put my charger, transmitter, a 10 cell 2400 mah NiCad pack, a couple 6 cell 270 mah NiCad packs, spare props and rubber bands and a few tools in a 5 gallon bucket in one hand and Tigermoth or SlowStik in the other. Run the charger off the big NiCad pack. I could get quite a few flights in that way.

J -I would rather charge half a dozen packs, crash half a dozen times, and call it a day.

T -Half a dozen packs for each of my planes would be a fortune. My 10 cell GP3300 pack cost more than $60 and I assembled it myself. I usually take two to four planes to the field. If the plane uses cheap packs I may have as many as three packs for more expensive packs I am lucky it I have two unless I can use them in more than one plane. I certainly would call it a day long before my sixth crash. Days like that I would be better of hanging out talking, I already have too many planes that need repairs

J -As for the next article, I hope to have it done soon. The prototype is a mess of wires and breadboarding, and I want to better understand the brain of the beast (the MAX712 chip) before going public with the whole thing.

T -I will be certain to read it, and I might offer some comments but I still respect you efforts and while you might get some compensation for your article I expect it is minimal and as such I consider your writing a contribution to our hobby.

J -So far it is built from an old PC power supply,

T -I use a surplus (free via neighbors trash) PC power supply to power my chargers in the shop.

J -(my charger) will charge 2 or 4 cells at a selectable current of .25, .5, 1, 2, or 4 A.

T -I am curious about your planes if they fly with 2 or 4 cell packs. That to me sounds like Indoor stuff and if so why the 2 and 4 amp current selection.

J -With any luck it'll be done soon and my house will not be burned to the ground!

T -"Laugh, Laugh, HO, HO. I am sure you are not serious, you are way too smart to not be careful and use appropriate precautions
Feb 05, 2005, 12:49 PM
Registered User
peterangus's Avatar

I was relieved to see your correction regarding the units of "energy density". But there's more. Your words "current up to 20 times their capacity" should be "current up to 20 times the one-hour current"

Back to "energy density". I know the terminology is established, but that does not make it sensible. How much more-meaningfull to say "energy-to-weight ratio".

The units "watt hours/Kg" are perfectly correct, but they are not convenient. The units "watt minutes/gram" are better for two reasons. Our run-times are in minutes, and the numbers are easier.
Feb 05, 2005, 03:33 PM
Originally Posted by peterangus

-----stuff snipped

The units "watt hours/Kg" are perfectly correct, but they are not convenient. The units "watt minutes/gram" are better for two reasons. Our run-times are in minutes, and the numbers are easier.

I with you on this although for bigger packs Kg is OK. I tried a few years ago to get some people on this forum to consider talking about pack capacity in terms of amp-minutes, a 2000 mah pack would be a 120 amp-min pack (am) could deliver 20 amps for 6 minutes. I will in the future I will refer to energy density in wm/g.

Jun 25, 2005, 08:35 AM
Registered User
John Gallagher's Avatar
Did we get part 2 yet?
Mar 27, 2006, 05:00 PM
When will the part2 be released??

Quick Reply

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Article Building the Perfect Nickel-Chemistry Charger Part II jeremynd01 Electric Plane Talk 6 Aug 24, 2017 02:56 PM
An Electrics Question. ifxmnfly Beginner Training Area (Aircraft-Electric) 3 Aug 10, 2004 10:03 PM
Building the Perfect Beginner Plane rutat Beginner Training Area (Aircraft-Electric) 3 May 07, 2004 01:28 PM
Building the perfect HOTLINER PARKFLYER (sport sailplane) soholingo Electric Sailplanes 55 Sep 17, 2003 07:12 PM
Building the perfect Tigermoth joemi Parkflyers 1 Nov 16, 2001 05:55 PM