Ni-Cd battery measurement

[oriol]

Hi everyone, I'd like to measure the capacity of a Ni-Cd battery, I thought on a LM3914 but the voltage variation is so short that I don't know if it could measure it. (I've attached the datasheet with the curves of charge and discharge of the batteries).
Any idea to do that

 

[audioguru]

the voltage barely changes as the battery is discharged. The voltage changes much more with changes in load current. the datasheet does not show the voltage change that is caused by temperature change.

I think you need to measure the load current and multiply it with a timer.

I had my Ni-Cad powered flashlight in the dark at the top of a high ladder. Then suddenly the flashlight dimmed to nothing with no warning. After a minute of resting it lighted for a couple of seconds.

 

[Willbe]

I'd like to measure the capacity of a Ni-Cd battery

Usually it's measured at a 20 hour rate. You discharge the cell at a constant current until the voltage reaches the cutoff value. Since the voltage is pretty constant you could just use a resistor. 1.25v at 7700/20 = 3.2 ohms @ 1w.
Whatever current it takes for this to happen in 20 hours is your "C/20" rate.
The C rate for this battery is 7700 mAh, and you can see from the graphs how the capacity varies with the discharge rate.

 

[Sceadwian]

Flooded lead acids are measured at a 20 hour rate typically, I don't know that any other batteries have such a standard hour rated listed for their capacity, as far as I know no such standard exists for measuring the capacity of a battery, even the flooded lead acid 'typical' values could be different for different makers.

 

[oriol]

Thanks for the replies. Firs I thought that with the voltage could be enough since alkaline batteries for example have voltage variations while they are discharging. In my case, Ni-Cd batteries are very linear with voltage discharging so it's difficult to do it with something like a voltmeter or voltage divisors, although I have a pack of 10 batteries like the datasheet one so I have 10 more times the voltage of 1 cell.
I would do that 20 hour test but I don't understand what happens if the batterie is not fully discharged when it reaches the 20 hours. And what do I do with that C/20? I would like to put a meter with some leds to have a reference of the battery level.
the idea of measuring the load current I don't know if is correct, as I would have to measure the maximum current the batterie gives to check if it gives the 100% or less, and I don't know if there are something to measure 7A autonomously, because it would be mounted inside an AUV (autonomous underwater vehicle).
So, the load current measuring could not be done ins't it? And if I have a C/20 what should I do?

 

[Sceadwian]

Load current measuring could be done, but you'd have to apply the discharge curve model to the current as higher currents will discharge the capacity faster. I'm not sure how but the effect of current on capacity is somehow linked to the cell resistance. No idea where to begin to calculate that.

 

[Willbe]

I would do that 20 hour test but I don't understand what happens if the batterie is not fully discharged when it reaches the 20 hours.
Then you should have used heavier current! That's a paradox with this method.
But, by plotting many curves you can approximate some current that will fully discharge the battery in 20 hours, or whatever rate you are interested in.


And what do I do with that C/20? I would like to put a meter with some leds to have a reference of the battery level.

How I did it for the Space Shuttle in the late 70s was to convert amps to volts using a shunt, and then volts to frequency using an V-F convertor, and then a counter. The count was then proportional to amp-hours used by the battery and so you then knew, in principle, how much charge remained.

Since batteries vary, the best you can do for a non-rechargeable battery is to have some level of confidence that you have discharged it fully, after testing several batteries.
A rechargeable battery can be tested several times for capacity and so you would be more certain of the charge level.
You also might want to research how closely terminal voltage correlates to charge remaining. There is a better correlation to terminal voltage if you load the battery first with a known, constant load, probably using a constant current sink.
If you model a battery as an ideal voltage source in series with an impedance, then probably both the voltage decreases and impedance increases, depending on your state of charge. Modeling electrochemical systems is messy, at least for me.

It ultimately depends on the desired level of accuracy and the desired level of confidence in your results.