Reports of the demise of this Web site are greatly exaggerated! We at sheldonbrown.com thank Harris Cyclery for its support over the years. Harris Cyclery has closed, but we keep going. Keep visiting the site for new and updated articles, and news about possible new affilations.


Lead-Acid or NiCd Batteries?
find us on FB
by Marty Goodman
converted to HTML by --Sheldon Brown
Spoke Divider

Note: this article is outdated. Lithium-ion batteries now rule for bicycle lighting.
BUT -- the advice here applies to older batteries, and other applications.
Sealed lead-acid batteries are still common for use in applications where weight is not an issue.

Which type of rechargeable battery best suits your night cycling needs? Folks often ask me about which battery to get with their bicycle lighting systems: Lead Acid or NiCd (nickel-cadmium) batteries. This is intended as a "canned" first response to such questions.

The two most common technologies of rechargeable battery used in bicycle headlight systems are Sealed Lead Acid ("SLA") and Nickel-Cadmium ("NiCd"). Low- to medium-price ($70 to $150) systems will feature SLA batteries, and medium to high priced systems ($140 to $400) will feature the NiCd batteries.

The Short Version of My Advice:

If you do a great deal of night riding, such as many nightly commutes per week, you should get a NiCd battery-based system. If you do only occasional night rides, you'll likely do fine with a SLA-based system. But in either case, you will probably be very wise to invest in a third-party battery charger, and not use the battery charger supplied with the system you buy.

The Long Version of My Advice:

Battery Characteristics

Costs/Benefits

Lead Acid batteries are about 2 to 4 times less expensive at time of purchase than are NiCd batteries.

However, NiCds, if properly cared for (this is a key operative qualification!) can be recharged 3 to 5 times as many times before they wear out as can be SLA batteries.

Cost of a high quality third-party charger for either system is roughly the same ($45 to $90). Note that SLA batteries require a different charger from that required by NiCd batteries.

Overall, NiCd batteries are at least as inexpensive, and probably actually somewhat less expensive a source of power than SLA batteries if you are using them frequently, over the course of their total life.

However, if you are using the battery infrequently, for, say, 20 rides per year, then the more expensive NiCd will probably die due to its shelf life's expiring before you use all its available charges.

NiCd batteries are, overall, about 30% lighter for a given amount of power capacity than SLA batteries. A significant, but not utterly overwhelming difference.

Storage Characteristics

SLA batteries retain nearly their full charge for two months or more just sitting on the shelf, unattached to a charger. NiCd batteries lose about 1% of their charge per day when sitting on the shelf, due to internal "self-discharge".

Discharge Characteristics

NiCd batteries have a flatter voltage vs time curve during discharge than do SLA batteries. This means your lights will remain relatively more constantly bright during the entire useful discharge life of the battery with a given lighting system than would be the case for a SLA battery of comparable amp hour capacity and voltage.

[This applies to systems with incandescent bulbs. LED lights include power regulation to hold the input to the LED constant -- until the battery can no longer supply the amount of power needed -- John Allen]

This is both good and bad: It is good that they stay bright, but this same characteristic means that they give little or no warning when they are about to poop out.

Don't Fully Discharge Either Type

Both NiCd and SLA batteries can be severely damaged by being deeply discharged to down below 75% of their rated voltage. With either system one must never run the battery "into the ground", letting ones lights go from yellow to orange to dim orange. Turn your lights off when they get noticeably yellow, else you risk permanantly damaging your battery.

Many ignorant folks claim NiCd batteries are subject to "charge memory". This is false. As used by cyclists for night lighting applications, there is no "charge memory" problem with NiCd batteries. Period. (See below for details.)

Some manufacturers who supply SLA batteries with their lighting systems (such as VistaLite with its VL4xx systems) choose the Hawker Industries (formerly called "Gates") Cyclon type SLA batteries. This particular make and model of SLA battery is significantly superior to ALL other SLA batteries. If you are replacing a SLA battery in your existing lighting system, get a Hawker Industries Cyclon battery pack (available in 2.5 amp hour and 5.0 amp hour six volt modules). These offer greater usable battery capacity for a given amp hour rating, are able to withstand deep discharge somewhat better than ordinary SLA batteries, and they last thru more recharge cycles than ordinary SLA batteries. Interestingly, the retail price for a Hawker Cyclon SLA battery is not all that different from the price of a similar ordinary SLA battery. Power Sonic (headquarters in Redwood City, CA) sells Hawker Cyclon batteries. Locally in Berkeley, Al Lashers can order and sell these batteries.

Spoke Divider

Spoke Divider

Charger Issues

With the exception of the Nite Rider Digital Pro 6 and Xcell Pro (formerly called NiteHawk) lighting systems, virtually all bicycle lighting systems on the market supply inexcuseably cheap, often quite destructive to the battery type of chargers. The problem is that with most supplied chargers, they charge the battery rather slowly (require 10 or more hours to provide a full charge) then then keep jamming current into the battery after it's full, heating it up and ultimately destroying it. many cyclists have destroyed their $140 replacement cost NiCd water-bottle battery by leaving it hooked up to the charger for some days or weeks.

While not a "smart" charging system, the NiteRider Xcell Pro and Digital Pro 6 systems do have a reasonably safe "set it and forget" charging system, though only when used with their supplied battery. Their system charges the battery at a modest rate for 10 hours, then a timer switches over to a 3 times slower charging rate for maintanance of the battery. Their system is not a "smart charger" in that it does not in any way sense actual battery condition.

It's relatively easy to make a cheap but safe "set it and forget it" charger for SLA batteries. All that's needed is a voltage regulator, which adds about $1.00 to $2.00 at time of production, or at most $10.00 at time or retail sale, to the cost existing bicycling lighting systems. Amazingly, few if any commercial systems provide this, and instead provide an unregulated DC power supply, which has the potential (liklihood, actually) of destroying the SLA battery if left attached. NiteRider MAY provide such a safe SLA charger in its new product, the "Trail Rat." I've not yet had a chance to analyze this new product on my lab bench.

To more quickly charge SLA or NiCd batteries (full charge in 2 to 4 hours), one needs a "smart charger". Such a charger senses battery condition during charging, pours current into the battery as long as the battery needs it, senses when the battery is full, and then cuts back to a much reduced current flow (or pulses of current at intervals) to keep the battery filled without harming it.

This kind of a very smart charger is a nice convenience with a SLA battery, but not necessary in that a simple voltage-regulated "trickle" charger will usually do the job fine for cyclists. Smart SLA chargers can be purchased from Power Sonic, at a cost of about $50 to $80 for chargers appropriate to existing bicycle lighting systems. You have to add your own cable, of course, to attach the charger to your particular system. Tinkerers should note that a proper trickle charger for SLA batteries is a regulated power supply set to 6.90 to 6.95 volts for a "6 volt" SLA battery, and to 13.8 to 13.9 volts for a "12 volt" SLA battery.

NiCd batteries really benefit from a proper smart charger. Unfortuately, one has to press into service chargers made for other purposes if one wants a smart charger for one's bicycle lighting system. Or make one oneself from scratch. I've done both, successfully. Certain DeWalt and Black & Decker power tool chargers can be converted into very effective smart chargers for bicycle lighting system batteries. The DW9106 and DW9104 in particular are good choices. Some camcorder and cell phone 6 volt NiCd battery chargers may be suitable as smart chargers for 6 volt NiCd bicycling batteries. I've built from scratch two smart chargers for my battery systems using a Maxim MAX 713 smart charger controller chip. Both work very well. Some have used the more modern 2002 NiCd smart charger controller chip made by Maxim, Benchmarq, and Unitrode. Contact me for details if interested.

Spoke Divider

Spoke Divider

The Myth of NiCd "Memory"

Discussion of the MYTH of "charge memory" in NiCd batteries
(from a discussion on the BikeCurrent mailing list)

Date: January 12, 1997 13:59:24 +1300
From: Steve Christall (steve@visioncomp.co.nz)
Subject: Re: Charge Memory post (again! arrgh!)

From GE Tech Notes ....

"Among the many users of batteries in both the industrial and consumer sectors, the idea of a memory phenomenon in nickel-cadmium batteries has been widely misused and understood. The term 'memory' has become a catch-all 'buzzword' that is used to describe a raft of application problems, being most often confused with simple voltage depression.

To the well informed, however, 'memory' is a term applied to a specific phenomenon encountered very infrequently in field applications. Specifically, the term 'memory' came from an aerospace nickel-cadmium application in which the cells were repeatedly discharged to 25% of available capacity (plus or minus 1%) by exacting computer control, then recharged to 100% capacity WITHOUT OVERCHARGE [emphasis in the original]. This long term, repetitive cycle regime, with no provisions for overcharge, resulted in a loss of capacity beyond the 25% discharge point. Hence the birth of a "memory" phenomenon, whereby nickel-cadmium batteries purportedly lose capacity if repeatedly discharged to a specific level of capacity.

"The 'memory' phenomenon observed in this original aerospace application was eliminated by simply reprogramming the computer to allow for overcharging. [Note that no mention is made of adding an intentional *discharge* to clear the problem - RLM] In fact, 'memory' is always a completely reversible condition; even in those rare cases where 'memory' cannot be avoided, it can easily be erased. Unfortunately, the idea of memory-related loss of capacity has been with us since. Realistically, however, ' memory' cannot exist if any one of the following conditions holds:

  1. Batteries achieve full overcharge.
  2. Discharge is not exactly the same each cycle - plus or minus 2-3%
  3. Discharge is to less than 1.0 volt per cell.

"Remember, the existence of any ONE of these conditions eliminates the possibility of 'memory'. GE has not verified true 'memory' in any field application with the single exception of the satellite application noted above. Lack of empirical evidence notwithstanding, 'memory' is still blamed regularly for poor battery performance that is caused by a number of simple, correctable application problems."

End of quote ...

Basically memory (loss of capacity) due to discharge is a myth.

Reduction of your NiCads / NiMH capacity due to overcharging (heating) and, cell reversal in voltage-depressed battery packs kill your batteries.


Peter Ludwig adds:

January 14, 1997 12:09:44 +0002
From: Peter Ludwig (peter.ludwig@bfi-bbrz.or.at)
Subject: Re: Charge Memory post (again! arrgh!)

Steve wrote about the memory effect:

"Specifically, the term 'memory' came from an aerospace nickel-cadmium application in which the cells were repeatedly discharged to 25% of available capacity (plus or minus 1%) by exacting computer control, then recharged to 100% capacity WITHOUT OVERCHARGE." [emphasis in the original]

To be correct this appeared in a satellite which cycles around the earth, so charging & discharging is very precisely the same in every cycle. Temperature is very constant through each cycle, and there are absolutely no mechanical shocks or vibrations and and and so on.

"This long term, repetitive cycle regime, with no provisions for overcharge, resulted in a loss of capacity beyond the 25% discharge point."

Also to be correct: The memory effect causes a lower voltage from that point than otherwise expected. This resulted in an early cut off by the software. The cells where by no means empty at this stage, but as we know, when a NiCd cell's voltage starts to drop, there is almost no remaining capacity. So the real 'loss' was caused by cutoff.

Anyway, there is AFAIK no other confirmed case where the memory effect had any influence in an field application.


Marty Goodman in Oct of 1997 adds:

I believe what Peter was trying to say in the last big paragraph is that normally, in the absence of a development of a "voltage knee", when a NiCd cell's voltage starts to drop significantly, it's going to die very quickly. And this is why control software in those satellites was written to shut them down when the voltage dropped below such and such. However, if a voltage knee has developed (as was the case) you have a situation where the voltage goes rather low relatively quickly, but it will stay at that low level with power drain for a long time... the total power output (amps time volts = watts) you can get from a cell with a voltage knee is not decreased, just the voltage at which it will put out that power. Thus, the control software was, sort of, needlessly and improperly shutting down the satellite (because it didn't "know" about voltage knees, and thought that these were normally discharging batteries that were about to be drained so deeply that they might hurt themselves permanantly due to cell reversal).

But the bottom line is that "charge memory" does not exist for virtually all intents and purposes of applications of NiCd batteries here on earth. In those extremely unusual cases where the phenomenon that ignorant folks call "charge memory" occurs, this phenomenon does not involve loss of total power capacity of the battery, but rather involves the battery delivering its full power capacity at a somewhat lower voltage during most of the discharge cycle. Hence, the phenomenom is more correctly termed "development of a voltage knee in the discharge voltage vs time curve" than "charge memory".

But for uses of bicycle lighting sytems, cell phones, camcorders, and laptop computers "charge memory" does not exist, despite what ignorant salepeople and writers of ads for these products may tell you.

Spoke Divider

Spoke Divider

WHY has the myth of "Charge Memory" in NiCd batteries persisted so strongly, given how false the concept is?

It is true that NiCd (and SLA type, too) batteries do in some situations benefit from being "cycled" (discharged nearly completely, then recharged). If a NiCd battery is injured by virtue of having been shorted out and discharged down to zero volts and kept there a little while, it MAY experience some degree of recovery if it's "cycled". Similarly, NiCd (and SLA) batteries that have been stored for a long time on the shelf and lost some of the their capacity may gain some or all of their capacity back after being cycled two to four times. Thus, there ARE reasons in some cases to "cycle" a battery down to a moderate degree of discharge, then charge it up, in order to "condition" it. But this is the case only in special situations, not routinely after each ordinary discharge. And it's not a function of "charge memory" (which, again, does not exist) but rather of other more complex issues.

NOTE that such "cycling" has to be done very carefully!!! NiCd batteries likely will be harmed by being discharged all the way down to zero volts. Proper cycling involves using a device that discharges the battery only down to a certain voltage (typically 1.0 volts per cell, or there abouts, for a NiCd battery pack, which means down only to 10 volts for a nomonial 12 volt NiCd pack).

Limiting the voltage you discharge a SLA battery down to during cycling is, if anything, even more important (a limit of about 1.8 to 1.9 volts per cell works well with moderate discharge rates). For SLA batteries are even more vulnerable than NiCds to being harmed if brough all the way down to zero volts: doing so will most likely destroy the SLA battery right then and there. So, if you try to "cycle" a battery, use equipment designed to do this properly and safely... do not just hook some load to the battery and walk away, and come back hours later or the next day!

NiCd batteries can be (I believe most often are) injured by overcharging. They can self desctruct due to growth of internal shorts ("dendrites") while left on the shelf. They can be harmed by repeated too deep discharge to more than 85% of their rated capacity (some others alledge this is the most common cause of premature death of NiCd batteries). The point is that there are lots of bad things that commonly happen to NiCd batteries. And what has happened is that, over the years, whenever someone has a NiCd battery that has died or lost most of its capacity, folks have tended to say "Ah! This battery is suffering from "charge memory", when in fact (a) "charge memory" does not exist and (b) the battery was destroyed by one of the above mentioned situations, usually one that would be preventable one knew the real cause.

It is, I believe, for the above reasons that the myth of "charge memory" has persisted. This phantom, false explanation has been invoked any time anyone has had a problem with a NiCd battery. Sadly, this false "knowledge" has kept folks from recognizing the real causes of problems with NiCd systems, and doing things about them.

More Articles by Marty Goodman

Spoke Divider

Spoke Divider

Articles by Sheldon Brown and Others


Reports of the demise of this Web site are greatly exaggerated! We at sheldonbrown.com thank Harris Cyclery for its support over the years. Harris Cyclery has closed, but we keep going. Keep visiting the site for new and updated articles, and news about possible new affilations.

Copyright © 1997 Marty Goodman

Harris Cyclery Home Page

If you would like to make a link or bookmark to this page, the URL is:
https://www.sheldonbrown.com/article.html
Last Updated: by Harriet Fell