NOTE: This guide serves as documentation for the tests that were conducted on the cells, not as the best way to do these tests in the future. The best way to get controlled cell data is to measure cells using an Source Measurement Unit which supports charging and discharging so we can control the amount of rest time in between charge and discharge, as I believe this had a large effect on some of the results. Automated test procedures can also be developed which will greatly reduce testing time.

In order to select the best cells for MSXIV, lots of testing must be completed to properly characterize the cells. We want to use the safest, highest energy cells in order to maximize the capacity of our battery pack. Several new cells have come out since MSXII and are looking promising based on online reviews, so we will be doing a comprehensive test of the best cells on the market.

We have ordered 5 each of 7 different types of cells

Panasonic/Sanyo GA
LG MJ1
LG M36
Samsung 35E
Sony VC7
LG M50
Samsung 50E

The following parameters must be measured in the least amount of time, with the highest accuracy possible:

Internal Resistance
Capacity
Temperature rise under different discharge currents – Test current justification: 1A will be a typical load for a small pack cell, 3A is high load for a small pack cell, 5A is worst case scenario for small pack cell

Cell current testing note: All the current ratings are current per cell, DC = discharge, Cells are discharged based on relative size and associated rough capacity (21700 cells discharge I*5/3.5, where I is the current that a similar test is used on for 18650 cells)

Equipment being used:

BK Precision 8600 Electronic Load, controlled via PC with LabView and Battery Test software used to log voltage and current. Capacity will be extracted from this data. SoC will also be extracted in future measurements.
Agilent 34410A Multimeter with RTD thermistor (Digikey: 223-1790-ND) used to log temperature connected to PC via ethernet cable. The multimeter runs its own web server and is able to be controlled via this interface.

Preparing all the cells:

Keeping track of the results of all the cells is required. In order to do so, we will mark every cell of each type with a number 1-5.

Use a permanent marker to mark the cells with a numbers 1-5 for each type of cell

Next up to connect the cells to the testing equipment with low-resistance connections.

We will be using strips of MSXII’s busbars to connect the cells in the 4P configurations.
For the single cell initial tests, we will use a removable system - clamping the cells in so we don’t ruin the potential to spot weld cells later (spot welding twice to a cell does not work well, as once the original tab is removed, the top is no longer flat, making it harder to get a good spot weld). This will be done using dimpled copper strips, and some kind of clamping mechanism to hold it in place.
For the multi-cell tests, we will use extra copper busbars from MSXII and spot weld 4 cells together using nickel strips soldered to these copper plates.

Measuring Internal Resistance:

The internal resistance of a cell depends on it SoC (State of Charge). We will measure all cells when they are fully charged, as this is an easy SoC to attain – just charge the cells.

Charge the cells using either a DC power supply set to 4.2V, 1.0A, and monitor it until the current reaches under 100mA - OR - Use the iCharger 306B on a 1S LiPo setting at 1A.
Use the iCharger 306B to measure internal resistance of the cells – it is one of the preconfigured tests.
Update the result in the spreadsheet

A resistance of the probes used to connect to the battery must also be measured so that we can measure the actual internal resistance of the cells, and not including the resistance of the cables.

Clamp the probes together without a cell in between.
Use the multimeter to measure the resistance and update it in the spreadsheet.

Setting Up BK Precision Battery Test Software

The electronic load will draw a constant current and will measure the voltage and current, and will log it.

Download, extract, and run the Battery_software.zip from the downloads link on this page: https://www.bkprecision.com/products/dc-electronic-loads/8600-150-w-programmable-dc-electronic-load.html
Download and install the National Instruments Visa Driver from here: http://www.ni.com/download/ni-visa-18.0/7597/en/. You will need to create an account.

Starting and Setting Up the Discharge Test

Start the Battery Test program and click on “Discharge Test”.

To connect to the electronic load, connect the load with a USB cable, and select the appropriate module (it should be the only one there). Also, set the baud rate to 4800.

For the test conditions down the left side apply the following settings:

Setting	Value
Discharge Condition	Discharge Current
Discharge Current	1, 3, or 5 Amps multiplied by the number of cells in parallel, for 21700 cells, multiply the 1,3, or 5 Amps by (5/3.5) in order to test equivalent C rates.
Stop Condition	Cut-Off Voltage
Voltage	Look on cell’s datasheet for End Voltage of Cut Off Voltage (Usually 2.5V, Samsung 35Es are 2.65V)

There are other ways to discharge test a cell, but for capacity measurement, we are going with this method.

To start the test, click the green button in the bottom right. Once started, stay in the bay and monitor the temperature of the cell. If it gets too hot (>80C), stop the test by pressing stop in the bottom left.

Once the test is completed, right click on the upper graph, and click Export – Copy to Clipboard. Then paste it in to a text file or into Excel on a new sheet, and add the test parameters – cell type, cell parallel count, and discharge current (or you can just export it to Excel, but this will require that a separate file is made for every test).

Setting Up Thermistor on Cells

The connection between the thermistor and the cells is critical in getting correct, as this will influence the temperature readings the most. If the RTD temperature sensor is loosely attached, it will measure the temperature of the air rather than the temperature of the cell.

In order to attach the RTD to the multimeter, solder a piece of wire (roughly 5”) to each leg of the thermistor and cover it with heat shrink to avoid shorting the legs and getting bad temperature readings.

To attach the multimeter to the cells, we will use some Kapton tape – because it is thermally resistant and will maintain its stickiness when heating up.

For single cell configurations, place the RTD with the flat side against the cell, roughly half way up the cell. Wrap it with Kapton tape to keep it in place and assure good connection.
For multi-cell configurations, do the same procedure as for the single cell, but do this before attaching the cells together. When attaching the cells, keep the RTD on the inside of the cell group where it has this least airflow – this will be a scenario with low airflow on the cells. If they get too hot in this location, additional cooling will be required.

Setting Up Temperature Logging

We will be using the Agilent 34410A with an RTD thermistor (Digikey: 223-1790-ND). The multimeter has built-in datalogging, and we will be using that to log the temperature. It also runs its own web server that can control it (when connected through ethernet).

To set up the multimeter, go to the 2-wire thermistor measurement (SHIFT, TEMP then CONFIG), set it up to 2W RTD, 0.00385 standard for platinum RTDs, a 100ohm setting for this RTD, and default setting for the rest. To run datalogging, select the datalog button, choose SETUP – leave all the defaults values (logging at 1s intervals, start delay small (140 usec default). For EVENTS/TIME, choose a duration, and set it to 2x the time that is should take to discharge the cells. (A 3500mAh cell discharged at 3A should take just over 1h to discharge). Make it 6 hours if you’re unsure – we can always stop the log early, or delete readings.

Once the discharge test on the DC Load has stopped, stop the datalog by going to DATALOG -> STOP, and make sure to SAVE the readings. The multimeter can only save one set of readings at a time (so it seems), so be sure to get the data off of it before starting the next test.

Get the data from the multimeter by connecting over Ethernet and copy the data to Excel or to a text file. See this video for details: https://www.youtube.com/watch?v=8MZHP6gKcOU. The video makes it look extremely easy, but it really isn't (at least nowadays).

After a few hours of trying different configurations, I got it to work using Internet Explorer, having Java RE 7u79 installed (the web applet uses java), and bypassing all the security features that prevent an old version of java from running (https://www.math.ucsd.edu/~crypto/fixjava.html). Since updating the firmware, it now works with Java 8.

The above is a method that I have found that works, although there might be easier methods. Keysight creates a software called BenchVue that advertises that it is able to connect to and control the Agilent 34410A multimeter. I have not tested it yet, but it is costly software.

A Note for 4P Tests

The cells get connected to the constant current load via the 12AWG cables and M3 nuts and bolts.

VERY IMPORTANT - Charge procedure for the tested cells:

Charge 4P pack at 4.2V, with 4A constant current
When current falls to 300mA, remove from charger
Wait 30 minutes for voltage to settle, then start discharge test

Cell wrapping: For all 4P tests, the module is to be wrapped in blue tape in order to remove the cooling effects of circulating air. The blue tape should be attached to the black holders, and should not be touching the cells directly.

Testing a Cell

How to test a cell:

Duplicate the test data sheet on the excel file
Set up thermistor on cells
Connect cell to DC load
Record the Cell Start Voltage, Cell Type, and Cell Number to the excel sheet
Set up the Discharge Test
Set up Temperature Logging
Start discharge and temperature log tests
Do not leave the bay until the test is finished
When the discharge test is finished, stop the temperature logging
Copy the data from the DC Load to the excel sheet
Connect to the multimeter and copy the data from the logs to the excel sheet

Clean up the test area