Maintenance Charge for NiCad battery?

[lctorana]

(Attn: SilverFox!)

This is a question on the optimum rate for NiCad cells permantly on-charge, for alarm or standby use. The key assumption is that they are fully charged before connecting to the circuit.

I have searched and researched in forums, battery manufacturers and the writings of RC battery "experts" for the optimum maintenance charge rate for a Ni-Cad battery.

We know that we have:

• Fast Charge - (say) C2 rate
• Standard Charge - C10 rate
• Trickle Charge - C10 - C50 rate

but if the cell is already fully charged, but is to be ready for immediate use, you need to overcome the innate self-discharge of the cells. This is what I mean by "maintenance charge".

For NiMH, all sources agree on the C300 rate. No issue.

But for NiCad, the concensus seems to divide into three camps.

1. A constant current, just enough to balance out the self-discharge. Typically of the order of C1000 or even less.
2. A constant C50 rate. Effectively trickle-charging 24/7.
3. Charge at the "standard" C10 rate with an intelligent (-dV, temp sensor) charger for one hour every day.

Methods 2 and 3 seem needlessly wasteful of power.

But what would the cells, if they could talk, tell us they really prefer?

 

[dano]

Nicads don't like constant "maintenance" charging.

The "memory effect" nonsense that permeated and followed Nicads for years was partially due to overcharging; a vast majority with consumer grade "dummy" trickle chargers that would overcharge the cells, because these chargers never stop charging, and Nicad cells will dissipate the overcharging by heating up and the cells would become damaged (usually called a voltage depression). The electronics that were using these cells could not compensate for the immediate depression, and would shut-down.

To avoid constant trickle charge induced cell damage, the base case scenario would be a timed trickle charge that will (theoretically) compensate for the self discharge rate.

 

[lctorana]

To avoid constant trickle charge induced cell damage, the base case scenario would be a timed trickle charge that will (theoretically) compensate for the self discharge rate.

There's the beginnings of an answer there, but more information, please.

At what charge rate is this trickle charge? C10? C1000?

And what exactly is meant by "timed" in this context? How often and for how long should the charge run?

 

[Patriot]

As I understand it NiCads store well with minimal energy in them. Constantly trickle charging will wear them out faster than just sitting on the shelf empty.

 

[lctorana]

Yes, but the question is NOT about storage. It's about keeping them at 100% charge, for alarm or immediate standby use.

 

[VidPro]

them?, are Them in series or single Cell?

i would go with a lightweight 1/10th C -1/25th C pulse timing relative to 2-4times the actual initial self discharge, say like a 1 second blip every 20 seconds. doesnt matter what you do, or call it, your going to be maintaing a trickle charge on it of some sort, if you expect the alarm to work when it goes off years from now. the blip could also be informative, showing that it occured with a led indicator, and mabey doing a quick test and showing a green led that is only the battery, right before the blip for charging. if you get 2 lights, than you can assume the alarm would at least sound.

shelf or charger the voltage depression on Load will exist , and its relativly the same, i would far rather have a charger voltage depression, and a charged battery that is weak, than a shelf voltage depression and a discharged battery when it comes to an ALARM the battery will really need manuel, cycling to maintain it, dont mater what you do. if you have a brain cycle it it would be down when you needed it, assuming it also got dimwit savings time all wrong too , meaning computer cycling during the day for a night alarm.

it isnt much going to matter if it floats down a tiny bit, then is recharged, in a lump or never gets to lean down ever, untill the battery is cycled through the part that is going to be needed, the stuff that hasn't been cycled wont be cycled, untill it all is, it wont act to its fullest potential. the further is sinks down before charge still will not address the part that didnt sink down, so the only highly active part of it will be the part you allow to sink down, and self discharge doesnt seem to BE cycling, so letting it go down on its OWN, to recharge it, isnt really an improvement anyways.

really unless the alarm is going off regularly, which it is then useless anyways, the lead acid with a float is about as good a choice.

if the alarm sounds, then it is going to need a Fast recovery, not a slow one, so the faster charger that pulses short when the charge is complete for maintance , would be a logical choice, how would you maintain a voltage float, if you have a voltage depression although that would be a simple method of Brain control of the thing.

 

[Turbo Guy]

Why Nicad?

Pb would be a better choice IMO .

Trickle is C20-C/ 50 or so. A daily timed charge at C/10 would be better than constant trickle.
Measure the standby drain,set timer fr period required to replace lost mAh each 24 hours.

10 Ah battery bank. Charger 1A. 500 mA standby drain in 24 hours. set timer to power charger up 1/2 hr. per day. Check sysytem after a week and see how it is holding voltage,adjust as needed.

 

[lctorana]

them?, are Them in series or single Cell?

i would go with a lightweight 1/10th C -1/25th C pulse timing relative to 2-4times the actual initial self discharge, say like a 1 second blip every 20 seconds. doesnt matter what you do, or call it, your going to be maintaing a trickle charge on it of some sort, if you expect the alarm to work when it goes off years from now. the blip could also be informative, showing that it occured with a led indicator, and mabey doing a quick test and showing a green led that is only the battery, right before the blip for charging. if you get 2 lights, than you can assume the alarm would at least sound.

shelf or charger the voltage depression on Load will exist , and its relativly the same, i would far rather have a charger voltage depression, and a charged battery that is weak, than a shelf voltage depression and a discharged battery when it comes to an ALARM the battery will really need manuel, cycling to maintain it, dont mater what you do. if you have a brain cycle it it would be down when you needed it, assuming it also got dimwit savings time all wrong too

it isnt much going to matter if it floats down a tiny bit, then is recharged, in a lump or never gets to lean down ever, untill the battery is cycled through the part that is going to be needed, the stuff that hasn't been cycled wont be cycled, untill it all is, it wont act to its fullest potential. the further is sinks down before charge still will not address the part that didnt sink down, so the only highly active part of it will be the part you allow to sink down, and self discharge doesnt seem to BE cycling, so letting it go down on its OWN, to recharge it, isnt really an improvement anyways.

really unless the alarm is going off regularly, which it is then useless anyways, the lead acid with a float is about as good a choice.

if the alarm sounds, then it is going to need a Fast recovery, not a slow one, so the faster charger that pulses short when the charge is complete for maintance , would be a logical choice, how would you maintain a voltage float, if you have a voltage depression although that would be a simple method of Brain control of the thing.

Thanks VidPro,

Yes, 3 cells in series for 3.6V (or 4V or whatever they settle at on-charge).

I se your reasoning, but there's a lot of electronics, timers etc implied! Worling on a simpler idea.

Oh, and you're both right about lead-acid being a better choice for alarm use. I will not and have never disputed that.

But I am still interested in the best way to keep Nicads at 100% charge.

 

[VidPro]

"simplest" have a voltage max set, if the voltage drops it recharges faster due to the voltage differential, if it is at the voltage top the speed slows down, and it keeps it at that. end of charge rate to be kept well within spec, beginning of charge rate after say alarm sounding could be very high.
voltage is about 4.1-4.35v max , gee wonder where you can get one of those from
make sure the charge unit is capable of handling the draw from the alarm if they are simeltaneous connected, and end of charge rate is well below spec, if anything changes, although if anything changes it is more likly that the battery will get softer not harder at the end of charge untill it becomes ruined..
same problem still exists.

 

[SilverFox]

Hello Lctorana,

I have actually thought a lot about this, and was approached to come up with a concept for a critical application.

I am not aware of the final "tweaks," but here is the basic concept that I presented.

Unlike your situation, we started with discharged cells.

The default mode charges the battery at 0.08C for 24 hours. If there is an interruption in power, the default mode kicks in and resets everything.

After the initial charge, a timer starts. The timer is set for 720 hours. At the end of this time, a 5 hour charge is done at 0.08C.

This is continued over 5 cycles.

At the beginning of the sixth cycle, a 0.5 amp load is applied to the battery for 1 hour. If the voltage drops below 1.1 volts, the discharge is terminated and a warning indicator is activated advising to check the battery pack. If every thing goes smoothly, the discharge terminates after 1 hour, then the 24 hour 0.08C charge is activated and the whole thing starts all over again.

The final thing that I threw out was that after 5 - 8 years I would like to do a discharge down to 0.9 volts per cell, but the unit would have to be taken off line during this, so automating it has a lot of difficulties. I believe the working model did not have this incorporated into it. It was suggested that this could be taken care of by simply replacing the battery pack. It was also speculated that it may be unreasonable to count on 8 years of uninterrupted power.

As far as I know, this was built and put into extended testing a couple of years ago. I believe it is still performing well.

Tom

 

[lctorana]

Thanks SilverFox,

So what you're saying is that for NiCad, you don't actually recommend a maintenance charge at all; rather a monthly C12.5 (essentially full trickle rate) charge for 5 hours. A duty cycle of 1:144.

Only one question - about the 6th-cycle discharge - what was the battery capacity in your example? Just trying to put the 0.5A current in context.

Interesting stuff. Surprised the design concept doesn't crop up more often, SLA preference notwithstanding. I'll think about it some more.

 

[SilverFox]

Hello Lctorana,

In this case the battery pack was a 10 cell 1200 mAh NiCd pack with a nominal voltage of 12 volts.

One of the hard parts of this was being able to dissipate 6 watts of heat during the discharge and keep the package compact. The target temperature range was from 10 F to 100 F.

The cells for the battery pack were formed and matched on capacity and voltage under load prior to being built into a pack. The actual charge current ended up at 95 mA. I believe the maximum voltage is 18 volts.

NiCd cells self discharge during the month. The 5 hour charge is designed to replace most, if not all, of the capacity lost during self discharge. After 5 months of this, the 0.5 amp load is to verify that the pack is still vibrant, and then the 24 hour charge is to re-balance and fully charge the pack. The 0.08C charge rate allows for some loss of capacity in the cells while still coming close to a 0.1C charge rate.

What seems to be happening is that during the hot periods, the charge does not fully charge the pack, but during the cold periods they are fully charged each month. This variation is an effort to reduce the effects of voltage depression, as is the 0.5 amp discharge every 6 months.

Tom

 

[Mr Happy]

If there is scope for some form of intelligent charge control, why not charge on voltage? IMHO there is a tendency to get fixated on constant current charging with nickel chemistry, when variable current, constant voltage charging is often simpler to arrange.

Pick a voltage, like say 1.40 V per cell, that represents a 90% plus charge state, and set the maintenance charger on a daily or weekly timer that applies a 0.1C current until the voltage reaches the set point and then cuts off until the next time. Alternatively, apply a constant voltage of 1.40 V via a current limiting resistor and "float" the pack at that voltage. Once the pack reaches the set voltage the current will reduce to just what is needed to maintain the charge, essentially the C/1000 rate but accurately balanced against the self discharge.

 

[lctorana]

OK, after that last post from Mr Happy, I feel ready to present my own thoughts:

A 4.5V transformer, with half-wave rectification, NO filtering, directly feeding the 3-cell battery via a resistor.

Assuming the battery is floating at 4.0V, the diode will conduct for only 25% of the cycle time, allowing the battery to rest for the other 75%.

As the battery voltage rises, the pulse gets shorter and the peak current also drops. And vice-versa, so it should compensate for battery condition to a limited degree.

I have estimated the self-discharge of the cells from a few months of usage, and have chosen the resistor value so that the average current, over the full cycle, is just a little higher than the self-discharge rate, which in this case is C1000-ish.

The overall effect is that the battery receives a very short belt of current, up to a peak of about 7 times the chosen average current, every 20 milliseconds.

And all that done with just three components - transformer, diode and resistor.

What I am NOT doing is the regular discharge. But I'm tempted to go with my minimalist design, and see how kind or harsh it is on the batteries. But remember, the currents involved are tiny.

Further thoughts on this fascinating (to me) topic are most welcome.

Apologies for the length of this post.​

 

[SilverFox]

Hello Mr Happy,

In practice that may work, at least for the short term, but there is some issues with it in theory. The reason it may work is that the battery pack would be sized so that only part of its capacity is needed.

The problem with picking a target voltage is that as cells age, their internal resistance increases, as does the full charge voltage. In addition, all cells don't age at the same rate. You can end up with a battery pack that has good voltage, but when you go to use it, a cell quickly dies as does the performance of the pack. In the short term it may work, but there are long term issues.

Tom

 

[Don McKenna]

Years ago I talked to Jarvis Yeh from Sanyo Batteries at an RC exposition. He said that nicad batteries can be continuously maintenance charged at 0.02C (2%). That was back in 1988 or so. I have been continuously maintaing my batteries at 0.2C since then. I have some packs that are 11 years old and they still have not lost appreciable capacity or voltage.

To do this I made a charger from a 24V Radio Shack transformer and 4 diodes in a full wave bridge configuration. The charge cords are made up of a resistor in series with a light emitting diode (LED). The size of the resistor determines the current required for each individual battery pack.

I have had as many as 40 battery packs floating on this unit at one time.
It keeps my planes and radios ready to go at any given time.

I do cycle my packs once a year, in February to see if they are still within 80% of their rated capacity, and shuck out the packs that are below 80%.