Benchmarking Batteries

[waoj]

I'm thinking about benchmarking some rechargeable batteries, to know which ones I want to order more of, and to be confident when one has degraded enough that I want to throw it away.

I could spend several hundred on a calibrated network connected load bank.

Or I could hack somehting together.  I think we all know what's coming....

I have a usb controlled relay board with 8 spdts.  I could line up several battery positions with that, to either charge or discharge.

I also don't want to wait around to look at and write with a pencil the voltage every minute.  But I'd like to produce discharge curve graphs.

I was thinking I need two meters.  one to measure amperage, and one for voltage.  but wait a second...

if My load is really dumb, just a resistive element for instance, do I need to measure current at all?  of can I infer it by knowing the unchanging resistance of the load, and the voltage at the load? v=i/r so i=rv.  right?

if I had one serial dmm. could I measure voltage, connect a known load, and then watch the voltage until it reaches a lower safety boundary, then disconnect load.  record voltage every second, calculate amps by multiplying ohms and volts.  then multiply amps and volts for that second to know how many watt-seconds were consumed during that one.. earth second?

Or do I really just need two meters?  the wiring is going to be very short, just 8 inches maybe 12 at the most and I'm not trying to damage my batteries, so I don't want an extremely heavy load for AAA's, AA's and 9v's  (Anyone want to suggest how many ohms I should use?)

I want to compare these two head to head first:
https://www.amazon.com/gp/product/B0BC8P2TS6
https://www.ebay.com/itm/204266648120

I want to know with confidence, if they're really giving me 1200mah or 1300mah like they say.  or how far off they are.

I have a couple of these 50w 25ohm resistors for load.. https://www.amazon.com/gp/product/B07SPC8CKK/ or maybe I'd use an incandescent bulb

[Mechatrommer]

the problem is you need to be there when minimum safe voltage is reached and manually disarm the discharge process... missing that, you have damaged battery.. so you need some mechanism to automatically do the cut-off, maybe you can snap some arduino to do this. for me, i just use my diy constant current load (with voltage cut-off feature) to automatically shut it off. and then charge the battery from the minimum voltage to fully charged. my NiteCore UM4 charger can show how many mAh in so i know what was mAh out. i dont care about plotting chart, otherwise there are too many batteries to chart. ymmv.

[RoGeorge]

For AA/AAA you only need a resistor in parallel with the battery, and an electro-mechanical clock, the dirt-cheap alarm clock or wall clock that runs from one AA battery.

Put the battery in the clock, and add an extra discharging resistor in parallel, calculated such that the resistor will discharge the battery in a matter of hours.  When the clock stops, you read the clock and find out how many hours it took.  The longer a battery lasts, the better. 

Another way, if you have a lab power supply, it can be used to discharge batteries and log current/voltage, like this:
https://www.eevblog.com/forum/projects/cc-cv-power-supply-as-a-contolled-load-for-batteries-(i-e-rigol-dp832)/

[waoj]

I can already control 8 SPDT relays via software/usb.  with two relays, I can have one switch between a charging source and a drain. and the other to act as a disconnect.  These 9v's don't even need the charging done via my hackjob of a system though.  they charge usbc, and i have a couple of those usbc meters to measure how much goes into them.  it's the out that is trickier

[Mechatrommer]

The problem with your 8spdt is to sense Vbat  to make cutoff,  does it have that..? Arduino uno you can snap few resistors and spdts or transistors for discharge and/or charge, and can do all Vbat Ibat sensing and logging to pc for chart plotting with arduino ide serial comm chart plotter. I cant think of anything simpler. Unless like what you said, using a dmm, a resistor, and manually eyeballing the reading and disarm when done.

[waoj]

Sorry, when I said "if I had one serial dmm. could I ...."  I meant "I do have one digital multimeter, with a serial port output"

That's what I was hypothesizing measuring voltage with.  It can't measure voltage and amps at the same time which is why I was asking if I could infer amps if I used a simple resistive load.

[Smokey]

Both of those 9V rechargeables look like they have a charge controller internally.  If they also have a battery management chip then they might internally disconnect when they get discharged past the lower limit (around 2.75V or so).  It's probably a boost converter off a single cell anyway, so I would expect they would have a discharge cut off too.  Does the output stay around 9V even as the thing discharges?  If so it's likely a boost converter.  That makes it hard to watch the discharge curve.  You could take one apart and measure the actual cell voltage during discharge.  If the cell voltage drops from around 2.75V then directly to 0V, then it's internally protected.

So you might not have to worry about stopping the discharge test manually.... maybe...

[Mechatrommer]

Sorry, when I said "if I had one serial dmm. could I ...."  I meant "I do have one digital multimeter, with a serial port output"

That's what I was hypothesizing measuring voltage with.  It can't measure voltage and amps at the same time which is why I was asking if I could infer amps if I used a simple resistive load.

it wont hurt trying. Basically yes, you can do that.

[waoj]

2.75 sounds pretty low to take a 9v cell.  I'm really not sure how low they go.  I suppose they probably are internally regulating the output.  They're lithium inside, and that doesn't evenly divide/multiply into 9v.  So really the curve is gonna be a flat line on these.  That's okay to me.  I'm interested in characterizing them as they would be used by devices they power.  So that means (to me?) at a certain number of ohm's load, how long will they run before they shut off.  My DMM would just be serving as a hi/lo sensor then, to timetamp when it started and stopped.  I could use an 18 ohm resistor to draw 0.5a at 9v, and if it put out 9v at 0.5a for 30 minutes,  Then i'd write interpret that result as 1500mah, right?

I do also have some nimh 9v's that I'm curious about which I'm pretty sure do not have active electronics.

What kind of amperage should I try to draw to not have my test take forever, but also not be too hard on a 9v? a AAA?

Reading https://www.powerstream.com/9V-Alkaline-tests.htm it looks like the mah I get out is gonna depend on how many ma I draw.  great...

[Smokey]

2.75 sounds pretty low to take a 9v cell....

If it's internally a single cell at 4.2V max and a boost converter to get it up to 9V, then it will always be 9V on the output until it shuts off. 
I was talking about 2.75V at the cell, not at the output.  That's where the last Li-ion cell I used internally shut off.

I had asked someone in another thread to take some scope screenshots of the output ripple of one of the 9V rechargeable and it certainly looked bad enough to be a boost converter output.

[SuzyC]

The most exact way to benchmark your batteries is using a Arduino or a 5V Vdd DIP PIC chip to do all the control and calculations and precisely monitor discharging your NiMH cells while reporting reporting current and charge level to a monitoring PC terminal program.
 

The whole circuit could be built on a breadboard. Your choice for the number of cells/batteries to be analyzed at once.
My analyzer incorporates a precise charging circuit as well to both charge and be an analyzer up to 4 round cells and two 9V batteries, each independently of each other and it displays both the progress and charge/discharge temperature, charge time, voltage, mAH  results on a 8-digit LED 7-seg display using a tickertape display mode to rapidly show so much information.

The real challenge here is to physically find a way to hold all the batteries and have easy insertion and good connections to your DIY analyzer circuit during the analyzing process. I built my version of this idea by removing the guts and repurposing a really cheap dumb 4 round cell charger to hold the batteries and make good connections to the cells/batteries to be analyzed.

To accurately measure the amount of charge you stuff into the battery, you use the MCU to integrate the discharging current over time.

This strategy involves using the PIC's or Arduino's ADC convertor peripherals.

For each AAA or AA battery use a 2.2 ohm 2W resistor for discharging. For a 9V, only a 220-ohm .5W discharge resistor is needed.

It is only necessary to use 1 ADC converter per discharging cell and only necessary to monitor voltage across each cell being discharged and then use ohm's law to exact the current and then integrate the  current/sec to get a precise measure of each battery's mAH charge capacity.  A 40-pin ADC PIC (18F45K22 for instance) chip can have 11 or more ADC inputs and so you can easily analyze up to 8 AA/AAA cells at a time and still have enough ADC's to measure the discharge of two 9V NiMH at a time as well.

An AAA/AA cell is fully discharged when it's voltage falls to .8V and then the discharging is stopped.
A 9V NiMH battery is fully discharged when its voltage falls to approx 7 to 7.5V.

You charge each of your cells by using a conventional consumer quality charger after first fully discharging each cell to be analyzed.

You then discharge each battery and measure the amount of discharge in mAH by integrating discharge current over time in your MCU DIY battery analyzer. You integrate current/sec to get total mAH capacity over discharging time.

The MCU would use one single UART port to send the discharge current, discharge current and mAH data to a monitoring PC using its serial port. The MCU does all the calculations and transfers the results continuously to a serial terminal program on  your PC.

A single momentary switch can tell the MCU to initiate discharge analysis of all the cells/batteries at once.
 
A single logic-level N-Chan power 18V min MOSFET with a low RdsON (~.01-ohm) has it's gate connected to a digital I/O pin in series with a 220-ohm resistor would turn on/off the charging for each battery. Connect the Drain of each FET to the neg side terminal of each 2.2 ohm battery/cell or a 220 ohm per 9V battery to be analyzed. Alternatively you could use relays to accomplish control, but your analyzer's cost and size might be overwhelming.

A 3 to 1 voltage divider would be used to feed an ADC input on the MCU for monitoring the 9V batteries.

Discharging reveals the true capacity of any AA/AA cell and 9V batteries through discharging after fully charging!

[Smokey]

...

An AAA/AA cell is fully discharged when it's voltage falls to .8V and then the discharging is stopped.
A 9V NiMH battery is fully discharged when its voltage falls to approx 7 to 7.5V.
...

But a Li-ion "9V" battery which is probably made from a 4.2V cell with a boost converter, which will stay 9V on the output the whole time until it dies.

[Mechatrommer]

But a Li-ion "9V" battery which is probably made from a 4.2V cell with a boost converter, which will stay 9V on the output the whole time until it dies.

thinking about it actually... doesnt matter since... mAh is "integrated" (area under the graph) btw: i have those 9V cells rechargable through 5V powerbank (with battery level led indicator if button is pressed)... never dismantle it who cares? i assume inside is 2 cells in series a little bit boosted, or 3 cells buck converted. once each cell reach 3V or below, buck converter goes into "saturation" and becomes unregulated, who knows? but i doubt cells of that small will be able to do 1000mAh or more. probably only 800mAh top for good brand and quality.

[Jwillis]

Here's a couple ways you can test your batteries under $100. 
The DL24EW 150W electronic load is pretty accurate and has some simple software for your PC that will log and graph the discharge process. Around $60 to $80
Theirs an older cheaper version that will do the same thing. Around $30 to $40
Or you can get a DT24TW battery monitor that also has PC software that will log and graph battery performance. Building a simple, stable electronic load is not difficult.
Either product has blue tooth that will link to a PC or smart phone.
Both can be set to stop discharging at a set voltage.
I hacked in a FT232RL USB to TTL Serial Adapter Module to a DT24TW because the PC being used has no blue tooth.
The software is easy to use. 

[BILLPOD]

I too benchmark my rechargeable batteries.   I measure the many parameters, including: capacity, charge time, discharge time, internal resistance, 'and much, much more', (as advertising people like to say). All the info I want is available on the MC3000 charger, which I connect to my laptop via USB.  It seems pricey, until you see all that it can do.   Many reviews out there, but here's a short one, as an introduction:
https://youtu.be/7BOZ9loG2eU

[waoj]

With an MC3000, could I modify the duty cycle and pulse frequency?  I'd like to experiment with that to get better results.

I'm sort of thinking I want a programmable power supply/load with a temperature probe.  I could put little niodidlium magnets on the +, -, and temp and keep the battery under test over a sink so if it leaks or burns its no big deal, but with data from the thermocouple, my code should be able to cut power before it burns.  I'd like to tweak the parameters of charging and measure results with the load.

[radiolistener]

If you use constant resistor load, you will need to measure Voltage and Amperage every second to get instantaneous Watts power for every second and then integrate it to get total Watt-hours. Then you can convert Watts-hours to Ampers-hours. This way is not easy if you're doing it manually. But it can be done on microcontroller.

The more simple way is to use constant current load. You can buy it on aliexpress and then all what you need it just to measure time when voltage drops from max charge voltage to min charge voltage and then multiply that time on used constant current and get Amper-hours