well, nicads have special charging requirements, and if you feed exvess current you can buckle the plates.

The best chargers are called constant current chargers, and they charge at

These links will show you some details on how to do it

Thanks for the link...am looking at them now.

You are indeed correct about the current requirements, which is why I am asking for a solution as to how to go about controlling current relative to the situation I described above!

You could switch in a dummy load for each battery set you remove. Sized properly it would 'waste' the extra 1 amp --perhaps a light labeled "BATTERY GROUP REMOVED"

That option is a thought, regards the "dummy load"...I have considered it as a last option, as I wish to have a reasonable efficiency of power consumption. If I charge one bank at a time -instead all at once- I may switch to an uncharged bank (from a charged bank) instead and use the "dummy load" for the final battery bank. Thanks for that idea

I'd recommend a dedicated charger for each battery cluster. Could I ask, are these batteries memory-free? If would help to deep discharge them just to start fresh, assuming they're identically aged and used, you'd be back to zero. Check for leaks and feel them for heat to see if they're actually charging.

The batteries are a mixed bag of age and quality but have the same ratings. I will be discharging the batteries before each charge, with each individual battery having a diode or two to prevent zero voltage discharge. If I come up with a simple solution to controlling current, I am seriously considering charging each battery individually at the same time. Example would be charging a hundred batteries at once, with each battery having a thermistor attached to it, so that when heat is detected (charged state), an op-amp in conjunction with a transistor will switch that battery off from the circuit.

But that would require many op-amps

CO,

I understand precisely your goal. Look into the poly-switches I linked earlier. They are an elegant, cheap, self resetting fuse often used inside a rechargeable battery to prevent both accidental shorting causing battery or wire damage and preventing overcurrent charging. The link includes specifications for NiCad batteries. The poly-switch does the current limitation on each individual battery while the external charging supply provides a constant voltage for the paralleled battery bank.

yeah, cheers. If my concerns for resetting is addressed and tolerance (also mentioned by Del), I will be more than happy to go this angle.

I am more than pursuaded at the moment to have an Op-Amp linked to a thermistor that will trigger a signal transistor that in turn will trigger ON/OFF a voltage regulator connected to each battery cell.

I will be using a thermistor to register a charged state for each battery, with a shared thermister as a reference to ambient...the rig I have at the moment has precise tolerance to temperature shift and will turn off the voltage regulator that is charging the battery when a one degree shift in temperature occurs.

An interesting point I have found is that the temperature slightly decreases during charging for a NiCad cell.

My current test rig is only for one battery cell, I will be trying more cells in "parallel" so I can charge them at a nominal 5vdc for a 1300mah battery cell. Bear in mind that each battery is seperated from each cell by a voltage regulator that will feed 100mah, with the only common ground being that all the battery cells will be feeding off the 5volt 30amp rail for charging.

My only concern is that what will happen to the last battery in the "string" to be charged...will the regulator spit the dummy when exposed to 30amps and drawing only 100ma to charge the last remaining battery? I don't know, but as I am feeding off a 30amp supply for one battery that dishes out 5vdc, there is no problem due to overheating...so I am thinking that there will be no problem.

It seems the ATX power supply controls the current output on the 5vdc 30amp rail based on the load...which is why I am having no problem at the moment.

Are you any good with single-chip micros (PICs etc)? If so, a possible solution would be a tree of analog multiplexers feeding an on-chip ADC (or external comparator) to sample the thermistors, and a set of addressable latches to turn each cell supply on or off (via FETs).

i thought of using some programmable micros or even some standard I/O chips from printers...My programming skills are limited to BASIC (showing my age here) and the challenge here for me would be how to address the "sample" of a thermistor...latching is not a problem. I see myself going this angle as a natural progression, once everything has proven to work using the non-micro way, allowing the Logic to be easily written.

You are right about the individual characteristics of battery groups...even for each individual battery lashed together to form a group. "idependant current control with a voltage cut-off" is what I am at the moment considering, hence this post. I have thought of the option using something like a three pin voltage regulator rated for the current required to be consumed for a single group of batteries. Requires only a resistor to control control current.

Regards.

GA for good answer with Cct & description. !!!!!

I add that any 3-terminal Fixed voltage reg [5V mostly preferred eg 7805 ] R1

decides the current through Series pass. R1's PWR Ratings depends on the

R1 x I Sq = ?

Regards

Haagee,

He said: "The IGBT sources the 5VDC/30amp lead from the ATX power supply to charge the batteries."

That means the batteries have 5 Volts chopped and applied via an IGBT.

I dont know the Voltage and current rating of the cells he is charging.

The regulator circuit has to be able to operate from the 5Volt supply. 5 Volts wont operate a 7805 regulator because the input must be 2.5 volts higher than output.

An LM317 can operate ON 5 Volts between 1.25 Volts and 3.5 Volts. The input voltage to the LM317 must be at least 1.5v greater than the output voltage.

I dont know what it will do with chopped DC.

If the packs are less than 3.5 Volts at full charge he ran the 6 LM317 chargers directly from the ATX 5 Volt supply. Otherwise he can run them from 12 Volts if the ATX 12 Volts can output 6 amps.

Jon

I have a dual (actually a tripple) voltage input for controlling respective IC's. The PWM acts a push-pull to increase the voltage from 12vdc to 18vdc to drive the IGBT driver...I was actually a lousy half a volt short from 12volts (using power direct from ATX box) to run the driver

The IGBT is used to seperate the charging circuit from the control circuit. The batteries are charged in parallel for each group. This means that each group can be ideally charged at a peak voltage of 5vdc (as each battery is rated at 1300mah) with the current applied to each group of ten batteries at 1amp...which is roughly 100ma per battery.

You have raised the problem of voltage requirements for the regulators...a point to be very much aware of. Thanks. I have grabbed a fist full of MC7809CT's before I read the replies to this post and will take into account your concerns regards the driving of the voltage regulators and thier unique requirements.

C Oz,

If you use the 12 Volts of the ATX supply you can use the 1 Amp 7809s the same way that you would an LM317 for constant current.

You could eliminate the extra electronics if you don't need the pulsed DC too.

You said you are charging groups of ten in parallel. Some will hog more current and some will not get much current.

A 130 mA constant current charger for each one would be nice.

Another way would be to use a lamp in series with each one. Hot lamps are relatively constant current through a range of Voltage.

These are about a buck each. For the ATX supply a couple in series for 12 volts. One for 5 volts.

http://www.surplussales.com/Bulbs-Incan-Panel/Bulbs-2.html

GE Lamp 5v @ 0.115 amp

(LMP) 715 General Electric lamp. #715. 5v, 0.115 amp. T-1 wire terminals.

Just another idea.

Jon

you are spot on...I have been looking at this voltage (current) regulator scenario for the last few days. Simple and relatively cheap. This idea allows me to consider charging batteries individually, instead of lumping together into groups as a compromise.

Oz,

An electronic instrument factory where I worked long ago had a huge panel that was used to charge, discharge and recharge about a hundred individual C and D and Half D cell NiCads at a time to determine their characteristics so that they could color code them for matching sets when used in battery operated instruments.

Jon

Drooling here, I am aloud to drool on this forum?

Precisely what I need, please send the aforementioned unit, lol

Looks like you are on the road to making one.

It even had an LED indicator for each cell.

I don't know how they logged data. Perhaps they had Data Loggers sampling and spitting out reports.

A much easier solution though is to use one of the self resetting fuses. Check out the Tyco poly-switch. (Other companies make these, too.) All that you have to do is put one of these in series with the battery. I recommend mounting this to the battery so that if the battery heats up from charging a dead battery the poly-switch opens and permits charging of the other 4 batteries.

I briefly thought of this option using special fuses...you have rekindled an interest in this area. My only ignorant concern is that the fuse may reset during a charge or I may have to devise a method to allow for a fuse to be seperated from a circuit without resetting during a charge and to be able to be reset on the next seperate charge.

I will look into this more rather than reject an idea I have not researched. Thanks.

Posted my reply after the wrong response. See other reply.

Beware of those resetable fuses...they have very sloppy tolerances.
They would also need to be powered off to reset as self heating would keep them in the tripped state.
I tried using them for DC motor short/stall protection... worse than useless.
Del

I have looked into the spec sheets for those IC's. I have not progressed into using these as I am unable to source an economic quantity for a home-based consumer. You are right regards the problems of a shorted battery and the possible damages following. I need to monitor each group independant of other groups and control each group independant from other groups. Would you have any ideas that may point me into the right direction without purchasing charge regulator chips?

Constant voltage charger will have decreasing current when batteries are charged.

Constant current charger will have increasing voltage until batteries are charged.

You'll need separate charging and monitoring on each cell/bank to drop them out when they're fully charged. You can't charge them as a group.

Say you have 6 cells at 1.2V each. You can charge each cell at constant 1.2V. Or 7.2V when they're connected in series. If one of the cell is "dropped", charging voltage need to be dropped to 6V. Don't think its a easy circuit.

You can try to charge them in parallel but you're not supposed to.

For any charging circuit, you'll need to disconnect the batteries after charging is finished. You'll also need to disconnect batteries from load when charging.

You could have a common power source to supply voltage to individual charging and monitoring circuits which connect to 1 bank of batteries.

It maybe easier and cheaper to buy 6 peak charger for your batteries.

"Say you have 6 cells at 1.2V each. You can charge each cell at constant 1.2V. Or 7.2V when they're connected in series. If one of the cell is "dropped", charging voltage need to be dropped to 6V. Don't think its a easy circuit."

You are right here. I was thinking of charging each battery cell individually rather than grouping them. Voltage will be constant at 5vdc for an individual cell with current at 100mah for a 1300mah rated battery using the 5volt 30amp lead from the ATX power supply. The IGBT that switches this 5 volt supply either on or off is driven by a driver that monitors the charging circuit and cuts off the supply should a short or undervoltage occur. I also have other protection in place that will isolate the charging circuit from the control circuit.

If I charge each battery cell individually with an op-amp and thermistor (for monitoring charge state) with a simple switching transistor connected to voltage regulator (to charge each cell independant from other cells) I may have a solution. Op-amps I have plenty of, getting a thermistor for each battery is not a problem and not expensive.

I am liking the idea of charging battery cells on an individual basis but simultaneously, rather than going for easy option of grouping them that is frought with problems.

Thermo monitor is not good. Quick charge use peak detection. Slow charge just tickle the battery.

Try Google and see how to detect fully charge NiCd. If all your cells are 1.2V, buy AA quick charger and "reassemble" them. One charge 4 AA so you only need 2.

Peak detection is in my opinion, more relevant to NiMh batteries than NiCd's. I believe the signature for a charged state for a NiCd cell is the register of an increase in temperature. If I were to go the way of "quick charge", I would be using a pulse technique. At a 175ma - 200ma charge rate for a single 1200mah cell, the battery does significantly register an intial increase in temperature once charged state has been reached. The thermistor detects this state and shuts off the charging current to this cell via the Op-Aamp/switch transistor combo.

My intitial problem was presented in this post and I believe people have answered, confirmed and clarified some unclear points.

I used play with electric RC car and maintenance of the NiCd is religious. Non of the RC chargers use temp to detect full charge. They might use temp for failure shut off or reduce charge rate to maintain safe temp. Its because its too late when you detect temp raise. All of them use pulse and peak detect. To get most from a battery pack people will test and match cells to make a pack.

A few result from http://www.google.ca nicd+charge+temp:

http://www.powerstream.com/NiCd.htm

http://www.solorb.com/elect/tmpchrg/

http://www.stefanv.com/electronics/fastchg.htm

This maybe what you want:

http://www.angelfire.com/electronic/hayles/charge1.html