DIY Battery Analyzer

[EEBlackSheep]

I have a bunch of 12V batteries I want to test to see if they are usable or not. I got enough crap in the shop that I should be able to build some kind of DIY battery analyzer.

I understand that I need some kind of high wattage constant current load.  Then I measure the voltage drop over time.

I have one of those radio shack RS232 DVMs - I wrote some software and can use the meter as a data logger.

I am not sure how to make a simple high wattage constant current load - suggestions?

Also - how do I interpret the data?  I'll have a chart of voltage drop over time.  From this how do I determine the amp-hour capacity of the battery?

thanks!

[IanB]

High wattage constant current load: bulb or bulbs. Just put various bulbs in parallel to get the load you need. Not exactly constant current, but close enough for a usability test on a battery.

[markus_b]

I don't think you really need a constant current load. But you need to be able to switch the load off automatically, when reaching a cut-off voltage to prevent overdischarging the battery. What is the chemistry of the batteries ?

[EEBlackSheep]

Sure I can use a string of lightbulbds for a load - done a variation of that before.

How do I interpret the data?  Ultimately I would like to figure out their amp-hour capacity as compared to when new.

thanks

[markus_b]

A couple of lightbulbs will be the cheapest load. It is simple and shows that it is running. The only drawback may be that is is slightly non-linear, but you can fix that with calibration.

Initially I'd measure voltage and current over the voltage range you are interested in. You'll get a table, like 13.5V - 4A, 13V - 3.8A, etc. To calculate capacity you measure the voltage, take the corresponding current (from the table) and calculate the Ah consumed.

[Balaur]

Firstly, please implement a way to cut-off the discharge process when the battery is depleted. All common types of batteries (Ni*, Li*, Pb) will not take kindly a complete discharge (i.e. down to 0V) and some damage to the cells is probable in that case.

As for what depleted means, it depends on the chemistry of the battery. Most of the time, you could compare the cell/battery voltage vs a known minimal threshold value.

Examples:
- A 12 V Lead battery: 11.7V (car battery) - 10.5V (the usual IEC value of 1.75V/cell)
- A single LiPo cell: 3.0V
- NiMH: 1V

... and so on.

The useful battery capacity is dependent on the method used for the charge/discharge.
The IEC standard specifies something like: full charge, rest, 0.2C discharge, rest, full charge

As you see, there is the question of 0.2C discharge, which greatly suggests the use of a constant current load. In this case, the available battery capacity (in A*h) is just a question of multiplying the discharge current by the discharge time (i.e. how much time it took to go down to the threshold voltage.

If you don't want to use a constant current load, then you should either
-  log the discharge current through your light bulbs load and integrate current over time
or
- approximately measure the current drawn by the load, measure the time it takes to discharge and hope for the best by a simple Capacity = Current * Time, with maybe some approximating (min, max) factors. This method supposes that the current doesn't change drastically when the battery voltage goes from max/nominal to min; which may be a reasonable assumption or not.

Cheers,
Dan

[Short Circuit]

For simple testing of the batteries, it does not really matter what load you use. Just needs to drain a significant amount of power from the battery.

If your were to test batteries for a specific application, you'd typically want try to use a load that is somewhat similar to the load of the application.
Could be resistive, constant current or constant power. Also the load would match the application; not very usefull to test with a 1C discharge if the
application is an order of magnitude higher or lower, since effective battery capacity is very dependent on discharge current.

Attached PDF datasheet for Yuasa NP maintenance free lead-acid gel batteries provides some detailed information on battery behaviour under various loads.
Other brands and models will have different characteristics, but you can expect them to behave roughly similar.

[markus_b]

Just re-reading the thread from scratch I think you just need a relay, operable with your PC and a 12V light bulb as load. You extend your measurement program to switch the relay off at a certain threshold and to measure the time until the switch off happens.

From the duration, the load, the start and end voltage you can calculate the battery capacity closely enough, certainly with 20% or so accuracy.