Load testing SLA Batteries?

I would use the one in the link posted below as probably the best and cheapest alternative to a very expensive industrial tester. You can adjust the load manually and combine it with a digital voltmeter and stop watch to test the 12v 12ah batteries using manufacturer's specification data as a baseline comparison. The tester listed above probably only does 50amps(>4C) or 100amps(>8C) and may damage batteries and significantly shorten their life if you are not very careful.

http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=91129

How much load should be used to test B&B EB12-12 Batteries

I would stay within the 180A for 5 sec spec. Maybe go 100A for 5-10 sec to be safe as the standard 30 sec at this current may be too much for that battery.
---
Avatar taken from http://www.electricmotorbike.org/
Anyone got one they might want to sell?
My KZ750 Project: here

To be safe I would do the following:
Standard 30 sec test: 50 amps
20 sec test: 75 amps
10 sec test: 100 amps
5 sec test: 150 amps

It depends on the duration of the test.

---
Avatar taken from http://www.electricmotorbike.org/
Anyone got one they might want to sell?
My KZ750 Project: here

It appear that one or more of the cells in your 4 pack has flip ( reserve charged) their impedance is much higher as before and will not permit the charger to recharge ,or if charged , the cells impedance limit the current in the discharge state.I recommend that you check each 4 pack with a impedance meter Like HP4328A , hopefully only one of the 4 pack has a high impedance cell.

Been reading this web site and other for several years , it is amazing how seldown I see the world IMPEDANCE ever mention discussed or inquired into about .

First thing that I do when checking any battery pack ( or cell) regardless of chemistries is check its impedance with a HP4328A milliohmmeter or similar AC 1Khz measureming meter , the reading can be anywhere from 100 microhms for a high capacity cell ( larger than 100 A/H) to as high as 50 milliohms for a good " AA " cell . From my experience having 2 Lepton-E scooter with Evercel 48 volt 30 A/H pack ( 28 cells in series @ 1.65 volts ) any cell with higher than 5 milliohms is on its way out and have to be monitered .
A good 30 A/H nickel-Zinc cell will be between 1.6 to 2.5 milliohm ( depend on its state of charge )

Load testing batteries is one way to test the state of health , working with batteries for over 45 years , found that impedance measurement is a far quicker , safer and better indicator of the cell ability to either deliver or accept current .

The lower the impedance the less heat created during charge and discharge.
HEAT it the primary reason for battery failure either because it dried out or the seperator broke down , alway goes back to heat ( alway impedance related ) Have been working with Panasonic "D" cell 6.5 A/H having impedance below 2 milliohms .have taken out over 300 Amps for over 20 second out of ONE cell ( direct short, those same cells where used in the Toyota first Prius ).
To sum it up . IMPEDANCE MEASUREMENT SHOULD BE THE SECOND SPECS AFTER VOLTAGE , LAST A/H .

A 48 volt 200 A/H pack is capable of delivering almost 10,000 W/H but without a impedance # it is useless to determine whether it could even move a 1/10 H/P motor .

There are 10-14 A/H " F " cells sold by numerous firm with ridiculous claims , theire impedance is such that you will dissipating most of the pack energy in the cell when placing any load higher than fraction of C rate .

The best way to determine what is the maximum power that a pack is capable of delivering , is
measure the internal impedance of the pack , example : 48 volts nickel-zinc that I ise in the scooter measure about 50 -65 milliohms across the whole string + wire etc .Multiply the cell impedance by two , in this case 110 - 130 milliohms ( Maximun power transfer is when source and load have match impedance )
Take your 48 volt divide by .110 -.130 , max current that you will ever be able to drawn is about 400 Amps or
about 2000 watts ( 48 X 400 that assuming all the cells are good shape .

Maybe this long winded explanation of impedance importance will help to determine which is the most suitable power pack for EV .If you are looking to take the energy out of pack over several hours (5-20) or even days then impedance
figure become far less important .
Be glad to answer any question by e-mail or Phone (203)627-8426 .
Presently doing work with Lithium Iron phophaste cell LifePO4 such as A123,valence and other , they appear to have low initial impedance and are very safe very light , price is the biggest issue ..

Had another thought , the warranty talk about 60 % capacity , I would not suprised if you measure the A/H over a 5 , 10 , or even 20 hours load that it will be within specs.
What the warrantee does not cover and address : is the drastic change of internal impedance ( higher of course)this most likely came about series a bunch of mismatched batteries pack ( one being lower and driven into reverse mode which cook the cell )

I is easy to make your own test setup. I don't think there is much utility to testing a 20AH battery at 100 amps. The testers shown at more for testing car batteries for adequate cranking amperage. For EV batteries, I'd test it at c/2 to C/4 or so - so for the 20AH batteries, about 5 to 10 amps.

I found that an #1157 automobile bulb draws almost exactly 2.5 amps with both filaments burning, so rig up 4 in parallel and you have a 10 amp test load. You should verify it with an ammeter to be sure. Hook up the bulbs and take battery voltage readings at appropriate time intervals until the voltage is down to the minimum values shown on the discharge curves on the battery's data sheet. Graph voltage versus time. (Or the log of time since that is what the battery data sheets usually have on them - probably just to make the curves look flatter) The time to reach the minimum voltage, times the amperage is the amp hours, and the curve you plot should look similar to the appropriate curve on the data sheet.

I've got a spreadsheet to do the calculations and graphing if anyone wants it.

`PJD ,
We would love to have a copy for our EV Batteries Collaborative Hand Books
Just send it to us at board_moderator [at] visforvoltage.org, and thank you.

Regards,
Board Moderator

Moderators are your dedicated volunteer V Team members who help keep your V Forums running smoothly and provide you with Forum Support.

Gman:

but I'm not really sure it's the batteries any longer.

I'm almost sure it is the batteries. Open circuit voltage is not a very good indication of battery health. Put a load on each one individually while measuring the voltage. This is a cheap way to measure DC resistance. Record how much the voltage drops on each one, and if you can't find a bad one than try and put more load on.

HCT, thanks for the explanation on impedance.

HCT:

HP4328A milliohmmeter or similar AC 1Khz measureming meter

Kind of expensive though: ebay one

HCT:

The best way to determine what is the maximum power that a pack is capable of delivering , is
measure the internal impedance of the pack , example : 48 volts nickel-zinc that I ise in the scooter measure about 50 -65 milliohms across the whole string + wire etc .Multiply the cell impedance by two , in this case 110 - 130 milliohms ( Maximun power transfer is when source and load have match impedance )
Take your 48 volt divide by .110 -.130 , max current that you will ever be able to drawn is about 400 Amps or
about 2000 watts ( 48 X 400 that assuming all the cells are good shape .

To my understanding, you can model any battery as a zero resistance power source in series with a resistor. That resistor determines how much the voltage will drop at a given current. You will get peak power (power = voltage * current) when voltage has dropped to 1/2 the open circuit voltage.

Your 48v NiZn batteries with .065 ohms of resistance would yield the following:
To get the voltage to drop to 24v you would need to draw 24/.065 = 369.23 amps. Because power is a function of voltage and current, at this load the battery would deliver 24v * 369.23 = 8861.52 watts.

---
Avatar taken from http://www.electricmotorbike.org/
Anyone got one they might want to sell?
My KZ750 Project: here

Andrew , you are correct in your math , my error was in rounding numbers. I came up with 2000 Watts should have been 20,000W . The most efficient transfer of energy is when both source and load have equal impedance , then I should have come with 10,000 Watt about 12 % higher than your calculated figures .
My first measurement after voltage is impedance , with those two , I am able to determine how much power I can take out , for how long will be determine by A/H figures .

Andrew ! very impressed ! that you have such good grasp of impedance and energy transfer, the concept take a while to sink in .

A additional concept that I have been using is comparing price in term of W/H per $ , not volt , A/H or whatever ,but pure a Watt?Hour which is energy .
Battery cost varies from a low of $ .18 W/H to as high as $ 5.00 W/H for the exotic chemistries with BMS , almost 30 to 1 range this is in commercial and industrial area , in military and airosspace application multiply that by another 10 or 300 to 1 range .
The evercel batteries nickel-zinc at about $.50 W/H is very cost competitive with a nice weight advantage compare to lead acid .The LifePO4 appear to be a very strong competitor when weight , life , safety and price are taken into account at about $ 1.00 W/H , of course BMS will add about 20-25% more to most chemistry and prevent catastrophic dead .