Radio College | Reading Room: Articles

Batteries Are Boring

by Chris Woolf

It is silly to pretend that batteries are anything but dull. The most exciting thing they aspire to is getting rather hot when abused. However they are central to a great deal of equipment and may represent a significant drain on budgets so its worth taking a look at recent developments and getting to understand them properly.

Language

Many users find the language of batteries as arcane as the chemistry but defining it makes discussion a great deal simpler. Though colloquially cells and batteries are interchangeable, in technical terms batteries are multiples of cells connected together. Cells have usable energy available measured in Watt hours (Wh) - the nominal capacity in Ah multiplied by the average terminal voltage during discharge. They can also be typified by their energy density in gravimetric or volumetric terms. This is the energy in Wh divided by either the weight or the volume of the cell. Where cell technologies have different characteristic voltages this is a useful method of comparison. However the energy density is to some extent dependent on current drain and can vary dramatically with temperature.

Although a cell is a storage system for electrical power it is imperfect and will self- discharge - the power leaking away over a period of time, again with the figure varying considerably with temperature. When in use the cell will have a discharge profile, a curve of voltage plotted against time. The profile may vary from substantially flat to steeply sloping and this can have an effect on the cell's apparent capacity - the amount of power that can be drawn from it in a particular application. The discharge profile is also associated with the cell's internal resistance which will govern the maximum current that can be drawn at any stage.

Primary cells are one-trip devices but secondary cells are reusable and have additional characteristics associated with their rechargeability. These include how far they can be discharged without damage - the discharge limit - and the maximum rate at which they can be safely charged. This is commonly classified in terms of C, the nominal capacity.

If this is, for instance, 1Ah then a C/10 slow charge will be 100mA flowing into the cell for 10 hours. A fast charge would be, perhaps, 2C - a 2A current flowing in for 0.5 hours. Cells also have a maximum overcharge rating - the charge they can continue to accept when fully charged without suffering damage - and an expected cycle life - the number of charge/recharge cycles that can be anticipated.

Power sources

The range of battery technologies available is bewildering so its sensible to limit this discussion to the ones that will be commonly met. The archetypal primary cell was the Zinc/Carbon (Zn/C) type. It is still available, as is the Zinc/Chloride (Zn/Cl) version which is able to supply higher current demands but, though they both have the apparent merit of cheapness, they are poor value and have little to offer. Their energy density is low, they cannot sustain high currents without a rest period and their self- discharge is so high (>15% / year) that it is rare to find a published figure for their nominal capacity.

They have been very largely eclipsed by another zinc technology, the Alkaline/Manganese (Alk/Mn) cell. This is able to supply continuous high current and has a greatly improved energy density. It is also much better value at about twice the Wh/$. It has the same nominal terminal volts (1.5V) and a similar, quite steep, discharge curve but a much better self-discharge figure (<5%/year). It behaves well at high temperatures but can only deliver 75% of power at 0 degrees Celsius and much less as it gets colder. Alk/Mn cells are potentially rechargeable but commercial secondary versions are not widely available and have severely restricted cycle- life as yet.

Silver Oxide (AgO) button cells are displacing the Mercuric Oxide types partly on environmental grounds. These 1.5v cells are designed for very long periods of low current use - calculators etc - where a flat discharge curve is needed. Zinc/Air (Zn/Air) button cells (also available as a PP3-style battery) share that characteristic too but are designed for a few weeks operating life. Their terminal voltage is 1.2v.

They need a ventilated battery compartment and can only deliver moderate currents (<20mA for the PP3) but have the advantage of extremely high energy density - 2- 300% better even than the Alk/Mn cell. They are not expensive in Wh/$ terms but need to be kept warm. Under heavy use at 0 degrees Celsius they will deliver barely 25% power.

Lithium technology is able to deliver very high energy density and operates below -20 degrees Celsius with little loss. Lithium Iron Disulphide (LiFeS2) is the only1.5V type. Energizer make an AA package which is a useful alternative to Alk/Mn under adverse conditions. Like all lithium batteries they are also very light - 33% less than the Alk/Mn cell. Lithium Manganese (LiMnO2) and Lithium Thionyl Chloride (LiSOCl2) are more typical in being 3.0v and 3.6v systems. This means that the AA/C/D-style packages are not interchangeable with zinc anode types but a PP3 LiMnO2 version (Ultralife) exists. This has an energy density 85% that of the Zn/Air type but can deliver a currrent up to 120mA.

Amongst secondary cells the Sealed Lead Acid (PbSO4) in "brick" or Cyclon styles is rapidly losing favour for portable use. It has a 2.0v terminal voltage but is heavy and environmentally unattractive. The simplicity of design with a rectangular shape and the cell's very low internal resistance (and hence ability to deliver high surge currents) has allowed it to continue for a while in some equipment. PbSO4 loses capacity in the cold and though it has quite low self-discharge it quickly deteriorates if left in a discharged state.

Nickel Cadmium (NiCd) is the most common rechargeable cell. Its nominal volts are 1.2V and its discharge curve is a flat S. It is available in a very wide variety of sizes and grades, some designed for trickle charging and others for fast. The self-discharge for sintered high-capacity types may be 15% / month. They can be stored charged or discharged but they are subject to some environmental restrictions for disposal because of the cadmium content. Many versions can accept quite high prolonged overcharge currents (C/10-20). They deliver about 80% power at 0 degrees Celsius and 50% at -20 degrees Celsius and the discharge limit is 0.9v. When being charged they need to be warmer than 10 degrees Celsius but cell temperatures above 30 degrees Celsius will prevent a full charge and may shorten cycle life. When fully charged the terminal volts peak and then drop slightly.

Nickel Metal Hydride (MiMH) is very closely related and is rapidly finding favour since it has about 40% higher capacity than an equivalent NiCd cell and no environmental restrictions. The main characteristics that differ are a tendency to higher self-discharge (~20% / month), a poorer tolerance to overcharge (C/50-300) and at full charge the terminal volts peak to a plateau - both factors that require consideration when charging.

Another secondary cell to consider is the Lithium Ion (Li+). This is becoming very popular for high-power and lightweight applications because of its excellent energy density (~70% better than NiCd). Its self-discharge is low at 5%/month and it delivers >80% charge at -20 degrees Celsius . The discharge curve has a greater slope than NiCd or NiMH and runs from 4.1v down to 3.0v. It is extremely fussy about its charging - it needs a very accurate constant voltage charge for the last period. While this is usually dealt with in the charger some Li+ batteries incorporate charge and discharge control electronics in the package. Li+ batteries and chargers tend to be dedicated to each other and the packages are intentionally designed to be incompatible with conventional types. An exception is the IDX NP-L40 which can replace NiCd NP-1 style batteries with significant benefits in energy density.

Connections

A battery's only point of contact with a circuit is its contacts. These carry high currents and need to be kept perfectly clean at all times. Resistances of a few hundred milliohms at each cell can cause significant voltage loss. In addition the dimensions of cylindrical cells have a wide tolerance and this can cause problems with spring contacts. Where possible spot- welded strap connections should be used to give greatly reduced contact resistance and improved reliability. Primary cells can be parallel- connected in some circumstances for greater current capacity but this is generally not advisable with secondary cells. Poor current sharing during both charging and discharging will soon cause serious cell imbalance and early failure. You should use a "fatter" cell size instead.

The Loopo Mixer

Although it is essential to have the theoretical data it is easier to explain the problems and choices to be made in a practical manner. The Loopo mixer is a hypothetical ENG mixer, but surprisingly similar to many on the market, with a battery compartment that can house 8 AA cells connected in series. It also has an external power input and a power-through output. The Loopo uses a not-very-special DC-DC converter for power that becomes unreliable with an input of 7.5v and has a meter that can register battery volts with a "green zone" from 8-12v. It draws roughly 200mA.

The first thing to realise is that the power that is advertised in the battery's spec sheet is not always available. An Alk/Mn cell's capacity is usually listed with a cut-off point at 0.8v. This would mean 6.4v for the Loopo but the converter only operates down to 7.5v so the real cut-off is 0.94v/cell - about 5% less than you thought. If one cell fails early, perhaps because you have mixed new and part-used ones, the problem occurs much earlier - perhaps 20% of the nominal capacity becomes unavailable.

If you put in 8 Alk/Mn cells on a summer's day and switch on you might get about 11 hours use. Zn/C cells would probably last little more than 2 hours and, to many people's surprise, NiCd only about 4 hours. The much more expensive LiFeS cell also appears to be poor value - only 15% better capacity than the Alk/Mn. However if it were a bitterly cold winter's day the Alk/Mn cell would deliver <75% rated power. The rest would be available later in the bar when you were getting warm again but in the snow it just wouldn't be there. However the LiFeS cell would almost unaffected - the difference is between the two is now 45%. If it gets a bit colder and you double the current by hanging a diversity radio mic system on the mixer supply the situation will change even more dramatically. The Alk/Mn battery will not last 1.5 hours while the LiFeS cells will deliver nearer to 6 - a difference of about 400%. Under these circumstances the expensive cell suddenly becomes very interesting.

Rechargeables

Many users will want the overall economy of rechargeable cells since their much lower running costs offset the initial capital outlay. NiCd is widely and cheaply available but NiMH is becoming attractive for its greater capacity. You can put 8 of these cells in the Loopo but you need to be slightly more careful now. With primary cells if you left the mixer switched on accidentally it would become unreliable at 7.5v but continue to draw current until near 0v. This would not matter - the cells were going to be chucked anyway. However at 8.0v the secondary cells would be discharged on average to 1.0v. With 8 cells there is a possibility that at least one might be as low as 0.9v. Below this voltage the cell's terminals may drop rapidly to 0v and then become reverse polarised by its neighbours. It may recover when recharged but its capacity will be permanently degraded. Equipment designed for use with rechargeable cells normally has protection against this built in but the Loopo (and a lot of other audio gear) does not. At 8 volts you must switch off or risk ruining your investment.

When running a mixer, radio mics, courtesy headphone amp and transcript cassette the extra load may become excessive for small cells. Under these conditions even a high capacity AA NiMH battery will struggle - it will supply about 10% less power than its rating would suggest. Many people will sensibly resort to an NP1-type battery plugged into the external power socket to counter this. If it is a 13.2v NP1B it will have 11 NiCd cells. When fresh off a fast charger it is quite likely to have a terminal voltage of nearly 19v - it could be even higher. You will have to check the Loopo's spec - and those of all the through-powered equipment - to see if that is allowable. For a lot of nominally 12v quipment 16v is the upper limit - letting the NP1B rest for half an hour may be sufficient to bring the volts down to the safe range. The over-discharge point for an 11-cell battery will be 11v (Sony risks 10.8v on Betacams). The battery meter indication at which you must turn off your mixer is now the same point at which your 8 internal NiCds/NiMHs start from. So much for the "green zone"marked on the meter dial.

Charging

Charging cells is essentially a matter of reversing the chemical reaction that produces power. Indeed most cells - even primary ones - can be recharged by forcing current into them. However the physical circumstances (temperature / pressure) under which the reaction can take place, the rate which it can proceed at, its limit point, and the number of other possible (unwanted) reactions that can occur all act as constraints. With primary cells there is a real risk of explosion and little chance of useful replaced charge. True secondary cells are those with cell chemistries and physical design that allow reliable charging over many cycles.

Slow chargers (~C/10) provide current from a constant voltage (PbSO4) or constant current (NiCd / NiMH) source. The rate is low enough that the preferred cell reaction dominates, little heat is produced and the degree of overcharge is unlikely to cause significant damage. In some designs a rather mean voltage specification for the supply allows the current to taper off near the end of charge. This tends to prevent overcharge damage but may elongate the period required for full charge. Part-used batteries should have something less than a full charge but it is hard to estimate how much.

Fast chargers charge at ~C/2 or higher and can only be used on cells designed to accept these rates. Here the problem is to force charge into the cell without overheating it or causing gas formation faster than the cell can reabsorb it, and terminating the charge exactly when all the reactants are converted. Simple techniques used include a timed charge which requires a full discharge first or a cut-off when the cell reaches a fixed temperature. Both fail to take account of ambient temperature and battery condition. Looking at the rate of rise of temperature (dT/dt) is more reliable but, like most temperature methods, requires additional battery contacts. A 2-contact system can use the alternative of monitoring the change in terminal voltage (dV/dt) at full charge - a curve which becomes better defined as the charge rate increases but which must be tracked very carefully in order to avoid extremely hot batteries. In practice many fast chargers tend to be dedicated to a single battery type and may combine several of these charge termination techniques for safety.

The most sophisticated chargers provide a complex charge profile. They run through various routines to sample the battery state by checking its voltage and its reaction to a charging pulse. This information, which allows the charger to recognise the number of cells and "understand" the battery, is stored and the battery is then charged appropriately using high current pulses. The average value of these is high enough to allow very fast charging (~4C). It is common to use a lower current for the last few % of charge - sometimes called a topping-off charge. In batteries of several cells one cell will always reach full charge before the others. By using a lower charge rate that does not exceed the permitted overcharge for that cell it is still possible to top up the others - cell balancing - and thus obtain full capacity. To overcome self-discharge it is normal to then leave a very low maintenance charge running (C/30-50 for NiCd, C/50-300 for NiMH).

Li+ batteries require very precise charging and usually have custom-designed chargers which give a constant current followed by very accurate constant voltage charge. Paradoxically this requirement for extremely tightly controlled charging makes them amongst the easiest cells to keep in good condition.

The smart-battery concept that is gaining commercial ground adds a serial data-link to the battery that allows it to converse with its charger and its load. The battery's charge state can not only be directly measured but also compared with previous stored data so that its capacity under constant and even intermittent load can be predicted. When charging is needed the battery itself can specify the charge profile it requires from the microcomputer-controlled current source that is really what a battery charger has become.

Memory Effect

So far I have avoided all mention of "memory", the best known, most talked about, least understood problem with rechargeable batteries. Myth and hypochondria abound and comprehensible facts are few: even battery and charger suppliers give conflicting and sometimes incorrect advice. The first point to make is that the loss of capacity noted in a three year old NiCd battery that has been heavily discharged many times, left to cook for a fortnight on an anonymous charger and has severely corroded contact straps from the last time it got splashed with seawater is not suffering from a memory effect. It is just dying.

Batteries have a finite cycle life (which can be massively reduced by abuse) governed by a mixture of chemical and physical changes including gassing, evaporation and crystal growth. These effects mostly lead to a gradual loss of capacity though certain crystal formations may cause cell shorting and a sudden death. They are inevitable and irreversible.

"Memory effect" should apply entirely to temporary changes. When NiCd or NiMH cells are repeatedly put through shallow charge/discharge cycles the naturally uneven electrode material will show a preference for activity in some patches and become relatively inactive in others. The resulting loss of capacity can be recovered by a discharge to 1.0v and a full recharge. Pulse charging will almost certainly prevent this happening in the first place. In other circumstances recharging partly depleted cells or prolonged overcharge can cause phase changes in the electrode material which give the cell a lower characteristic voltage. The cell voltage then develops a kink in its curve. Since the end-point for the discharge of a cell is governed by its terminal voltage the result of this kink is an apparent loss of capacity - it reaches 1.0v earlier than expected. Several cycles of discharge / recharge will return the cell to normal but it's life may be reduced.

These effects are a factor of Ni chemistry but there seems to be agreement that NiCd cells display them more obviously than NiMH. Many advertisements for NiMH claim no memory effect but the data sheets merely claim a reduced susceptibility. However in practical circumstances the loss of capacity due to poorly balanced and over-discharged cells of either type is far more significant - "memory effect" is very rarely a real issue. Li+ does not display any such behaviour and shallow discharging will tend to increase its cycle life.

Fast or slow?

It is painfully common to hear specious arguments about whether fast or slow charging is best. A crude fast charger that blasts charge into a battery with little regard for its temperature or state of charge will do considerable harm whereas a low-rate trickle-charger has less wrecking potential. However a well-designed intelligent fast charger will almost certainly give greater battery life and better working capacity than the 14-hour device. One injudicious over-discharge involving a polarity reversal may do more harm than any of them. The point is to consider the entire cycle rather than just the charge period. With very great care of every aspect of a good quality rechargeable battery's use a cycle life of 3000 is possible - with severe abuse you may be lucky to get 30.

This article was originally published in the British magazine "Sound Pro" in February 1998 and is used here with the author's permission. Chris Woolf is an experienced sound recordist who now works on a consulting basis for manufacturers.

Follow-Up Q&A With Chris Woolf

Alexandre Harrois writes:

I was glad to find this article giving an overview on batteries and especially on "memory effect": it is clear and simple. Yet there is one sentence I have troubles with. It is the last sentence of the last paragraph in the "Memory Effect" section: "Li+ does not display any such behaviour and shallow discharging will tend to increase its cycle life". Do small discharge/recharge cycles REALLY have THAT effect on the life expectancy of the battery

I read also on the web (http://www.mobile tec.com/html/faq.html, and http://www.ora.com/reference/dictionary/terms/B/Batteries.html) that Lithium-Ion batteries do have a life expectancy based on the number discharge/recharge cycles (these pages gives estimates ranging from 400 cycles to 800 cycles). In this case, do small cycles (i.e. recharge after little use) shorten the life of the battery or not? Is there any reliable experimental data on the subject or are we only working on theories?

Thanks, Alexandre Harrois (from France)

Chris Woolf responds:

Hi Alexandre,

I'm glad you found most of the article understandable and clear. The comment on shallow discharge of Li+ does make sense but was probably less well phrased than it should have been - I apologise.

Li+ life is usually specified as the number of 100% discharge cycles it can survive. Different manufacturers use different figures to decide when a cell has "died" - some reckon on 50% of original capacity. If a cell is rated at 1000 cycles of 100% discharge but you only discharge about 40% of its capacity each time you will probably get considerably more than a 1000 cycles out of it. This would not be the case with a Ni chemistry cell - it would probably give you less than its nominal cycle life. Hence it is reasonable to claim that a Li+ cell's life is enhanced (comparatively) by shallow discharging.

I got some of the detail from http://www.atip.org/ar-pub-95.html (pick Atip95-33r) which does mention some real experimental tests. The other articles you have noted are rather brief overviews and have compressed facts heavily. Life cycles are very dependent on design details, the type of use and the charge routine. The same goes for capacity. Taking a particular battery (NP1 as used with some video gear) the highest capacity one sold currently is a 50W NiMH one - the Li+ versions are 40W - so volumetrically you might deduce that Li+ was a poorer bet. But that is not the whole story. The problem is simply that very few cells or battery types are exactly equivalent - and hence comparisons are always more elastic than one would like.

Keep questioning everything - never simply believe!

Chris Woolf

Russ Myer writes:

Great, informative article on batteries. Did NOT find a specific answer to MY dilemma:

Can I leave my NMHydride AA batteries in the charger and let them trickle-charge continuously? Or should I just "top off" before needing the batteries (hard to know, sometimes). Your information will be MUCH appreciated. Thanks.

Chris Woolf responds:

Ah well, it all depends, says he tugging at his grey beard and nodding sagely. This is what makes batteries such a mystery!

Any NiCd or NiMH cell will have a maximum overcharge rating. This is the charge current that can be allowed to flow into the cell on a continuous basis without damaging it. For NiMH cells this used to be about C/20 but may be as high as C/10 with the newer designs. Trickle chargers are usually built to give a full charge in about 16 hours so the rate of charge is usually around C/10 - remember that cell charging is not 100% efficient so a C/10 rate won't actually charge a cell fully in 10 hours.

This means that there is an area of uncertainty with the answer. If your cells are new types and the charger is a gentle one that does not deliver more than C/10 you can probably leave them on charge. If the charger is beefier and the cells are older style or slow charge ones (which tend to be less robust) then they may suffer.

In fact it is kinder to the cells to keep the continuous charge rate well below the max rating. Some chargers offer (or can be modified to give) a low maintenance charge for NiMH cells which is ~C/30. This rate can be indefinite on almost any cell and simply counteracts self-discharge. A good guide to what is OK is cell temperature. Remember that a cell that is barely warm to the touch is <40†C - which is as much as it should be. Cooler is always better until you get down to near freezing.

If your charger cannot provide a ~C/30 rate then take the cells off charge and leave them somewhere cool until you need them (this keeps self-discharge well down). A top-off charge of maybe 1 hour at C/10 per week of storage should give full capacity again.

Sorry, its a long answer but its better for you to understand rather than be told - that way you can work out what to do with what you have. Have fun, and distrust anyone who tells you anything too rigid about batteries. They are complex chemistry and their behaviour is always a little imprecise and variable.

Chris Woolf