Why do we need to apply a load to a battery when testing its voltage?
I've been testing battery capacity so far using a multimeter and just connecting the probes to the positive and negative side of the battery, which seems not to be correct.
Why is this? Why do we need to test batteries with a load and why can't we just simply check voltage with no load?
You need to put a load on the battery to see if it has any charge left.
Without a load, it may show an acceptable voltage, but when you actually try to use it the voltage drops because the battery is nearly dead.
So to see if a battery is really usable you must measure the voltage when the battery is connected to a load. Like this:
Dead Battery, no load, 1.4 Volts
Dead Battery, load of 100 Ohms, 1.0 Volts
Good Battery no load, 1.5 Volts
Good Battery, load of 100 Ohms, 1.4 Volts
Those numbers are just representative - do NOT use them to actually measure your batteries. Check the unloaded voltage of a good battery, then check the voltage of a good battery under a typical load. Use that typical load to test other batteries. That is to say, figure out the equivalent resistance for the load and use a resistor of that value in your test.
Quite a lot of battery chemistries will, if left alone, raise their terminal voltage. But there may be no capacity behind it and it will drop as soon as you try to use it. So a load is connected to the battery to verify that it is actually useful.
As typical Alkaline and other batteries go bad or get weak, they develop greater internal resistance. With no load or very little load you could say that there is a voltage divider formed by the internal resistance and the high resistance external "load". The high external resistance will show a high or full voltage drop. With a good external load, low resistance, the internal resistance of the battery will experience a greater voltage drop, meaning you'll see lower voltage externally.
To add a practical example...
Have you ever found an old flashlight (especially incandescent), turned it on and thought the batteries were good only to see it dim and possibly turn off several seconds later? Or possibly a motorized toy?
When the batteries aren't being asked to do any work, there may still be enough of the active chemicals to raise the voltage to nearly-new levels, but not enough to sustain that voltage under load. Once you've stopped using them, the voltage will gradually climb back up again.
An ideal multimeter in Voltage test mode draws no current, so you're not testing the battery's ability to sustain the voltage when doing work. No multimeter is ideal, but even a cheap one draws little current.
It doesn't work as well with LED's or other semiconductor-based objects because they have a relatively flat performance in the "good" voltage ranges and then just stop once they get below a certain threshold.
P.S. mA and mAh
Current (measured in Amps, Milliamps, etc.) is the measure of how fast electrons are passing through a circuit.
Capacity (measured in "Amp hours", "Milliamp hours", etc) is the measure of how long a power storage device can supply a certain current.
It is roughly true to say that a 1000mAh battery could output 1000mA for 1 hour or 1mA for 1000 hours. Reality is a bit more complicated, though. Batteries are much happier outputting 1mA than 1000mA (1 Amp). You'd have a hard time getting a 1000mA battery to output 10 Amps for 6 minutes without exploding.
POI - mAh, or similar, is a measure of charge, not energy. Multiply by the assumed or measure voltage to get energy.
