Trickle charger advice

Un bel giorno Alexander Avtanski digitò:

There are two currents of thought. Some say that if you use trickle charging currents between C/20 and C/40 you are in the clear; most manufacturers declare this kind of trickle currents in their datasheets, for example:

In this case the capacity is 150 mAh and the trickle current is 4.2 mA, i.e. about C/36.

Some others (few) say that a continuous charge of C/20 could lead to a long-term degrade of the battery, and advise to use a much lower trickle current, between C/100 and C/500. I don't have enough long-term data to categorically disprove these claims, but they don't look very convincing to me.

Suggestion is to pulse (boost) charge it, not continuously (cook) charge it.

My understanding is that long term damage is caused by overheating. I wouldn't go over C/40 for trickle charge. As self discharge depends on temperature I probably wouldn't go much less than this either.

We service stuff we made 10 to 15 years ago which uses C/40. I change the back up battery if it's been installed for 5 years or more. I've yet to see premature battery failure.

Hello again and thanks for the answers.

So, there is no major issue charging with the diode/resistor charger. Anything below C/40 should be pretty much safe and wouldn't damage the battery, at least not too soon and not catastrophically.

Pulse charging also sounds very attractive and may turn out to be better. It is also easy for me to implement, since I have a PIC microcontroller already in the device and I can ask it to control the pulses for me, so almost no extra parts. What is stopping me from using pulse charging, however, is that it seems to be a bit more dangerous (what if the microcontroller hangs and leaves the charger in a high current state, or if something else goes wrong?). The resistor charger looks foolproof to me, and if there is nothing I miss I would go with it, even if it might not be the best for the battery - so what, a 9V NiMH is not that expensive, big deal if I have to change it once every few years.

Any other considerations?

Thanks,

- Alex

If it self-discharges at 30% per month, you should be able to keep it happy at a charge rate of 100% per month. That's equivalent to C/744. So something in the C/500 range should be fine.

John

Just add temperature monitoring and you are almost doing a real charger. A 11 cents themoresistor will work with any uC for temperature. I am about to buy a thousands 100K themoresistors.

Some of the manufacturer data sheets I've read for NiMH cells say that this chemistry's efficiency of charge acceptance is very poor at these low charge rates. Almost all of the energy tricked in at such low rates is wasted (ultimately as heat, which probably isn't an issue at such low currents) and very little ends up rebuilding the electrochemistry. A trickle at C/500 might not be enough to overcome self-discharge.

One data sheet suggested that if maintenance charging is to be performed, it's preferable to use a low-duty-cycle pulse charging method at a higher current level. As an example (from memory, not from the data sheet) - try a one-second pulse of current at the C/5 to C/10 rate, once a minute or so.

As long as you monitor the voltage and temperature, you can pulse charge it at C or even higher. Just use long enough delays between pulses.

but I

ion

ouldn't see

lematic

Well, I said NiMH just because that is what I saw in the local RadioShack the last time I looked. It doesn't have to be NiMH, if you think NiCd might turn out to be better.

What do you think?

- Alex

but I

see

problematic

NiMH is friendlier to the environment. Cadmium in the Cd is extremely toxic. If you already have the microcontroller, why not try a little harder in using NiMH?

Admittedly NiCd is a different chemistry from the long-term charging POV, but I have used C/100 as a charge rate for the 9V backup batteries in reticulation controllers for over a decade without any issues. For your NiMH I couldn't see any reason to go higher than C/100 unless their charge acceptance is problematic at these rates.

So be it, then - NiMH.

About the charging mode: what worries me with the pulse charging is the remote chance that the PIC might go haywire - power surges, interference, who knows. I run it at 20MHz and I noticed that sometimes I get resets if I touch the wrong pin with my hand (might be something loose in my prototype board). And if it does fail, it might just blow the battery - something I want to avoid at all cost, because I plan to be using this as a clock at home too. It will stay powered on all the time and unmonitored potentially for days. The simpler charger looks to me completely foolproof - I cannot imagine a way for it to fail catastrophically. What do you think?

On the other hand, if I was doing a dedicated battery charger, sure, I would go with pulse charging and a PIC with simpler program, that is dedicated only to monitoring the battery. Also, such a device would not stay unattended for long periods of time.

I don't know... The pulse charging still looks tempting - it would take me minimal effort to implement it too. Hmmm...

- Alex

There really is no such thing as a 9V NiMH since the cells are nominally 1.25V.

The self discharge nowadays is not as bad as 30% per month. That is the number when they first came out. The chemistry is different these days. If you used Sanyo ENELOOPS, the self discharge wouldn't be an issue at all.

I really don't know if dendrites are an issue anymore, which is the reason people didn't keep NiMH or even Nicads under constant charge.

Personally, since you have the grid, this sounds like an application for a gel cell. They like constant charging. Further, they self regulate their current, so all you really need to do if provide a constant float voltage. Charge the battery through a Schotky diode, and you have your power back up switch solved.

If you can handle the size and weight, lead-acid batteries are very well suited for standby charging. You can get them in fully sealed versions, so there's no worry about acid splashing about.

I have never seen them smaller than a few Ah, though, so they may not be an acceptable replacement for a 9V battery.

see

problematic

One interesting fact about NiCd, they love being deep cycled. It is the only battery chemistry that does. Old NiCd battery reconditioners take advantage of this property. If you do not want serious deep cycle (preferably between two battery packs) use a lead - acid gel cell, they love float duty.

The smallest I've seen was the size of 2 AA cells, a single cell, it was used to power a walkman.