PDA

View Full Version : Testing individual cells



DaveAK
25 June 2011, 0950
Somebody make this in to a HowTo for me please. :D

Since it's pretty obvious I've knocked out at least a couple of the cells in my pack I've torn it down so that I can check each cell individually to assess the damage. What I need to know is a good DIY way of testing them without buying a proper tester like Travis uses, (although I'd love to have one of those!)

What I have available is:


a lab power supply that I can use as a CC/CV charger to charge the cells individually
a Turnigy watt meter so I can measure voltage and current simultaneously under load
some nichrome wire that I can use to place a load on a cell


(There's been a delay on the power supply, but it should be here by next weekend.)

So would somone be kind enough to outline a good testing procedure of charging, discharging, recording of data and calculating health of a cell. Can I use this stuff to measure things like IR and AH capacity? What am I going to get out of it at the end to enable me to compare one cell to another, or to say this cell is good and this cell is bad?

DaveAK
25 June 2011, 1013
Hmm. I just checked eBay and I can get a CBA-III analyzer for $163. That would normally be in the budget for next payday if I hadn't been hemorrhaging money lately. I'll keep it on my watch list, but still would like to know if there's a cheap and easy DIY method.

larryrose11
25 June 2011, 1228
Hey Dave,
does the Turnagy meter iterface to a computer?
If so, your in business!
Basicaly, I would compare battery internal resistance, and capacity, and do it under a healthy load. As for battery capacity, you could start with a full cell, and discharge the cell to the LVC. From the recorder data, you can caculate battery capacity. It all hinges on having a sampling rate about 10- 100 HZ.

larryrose11
25 June 2011, 1229
Oh, and calculating internal resistance:

http://www.buchmann.ca/Chap9-page2.asp

chef
25 June 2011, 1244
Internal resistance can be calculated from the ratio of voltages to currents measured at two different loads:
Ri = ΔV / ΔI

The deltas are absolute values. One of the measurements can be at no load (open circuit voltage & zero current), though I have a vague recollection that it's not as accurate as using two non-zero currents. Might want to record three measurements and compare (no load + two different loads).

Measuring charge (Ah) is harder. To draw constant current, I believe the test device needs to compensate for changes in voltage & internal resistance as the SOC drops. A good-enough measurement may be to use a CellLog 8S to log voltage over time with a fixed load and approximate constant current. You'd need an LVC to stop the test and prevent over-discharging, of course. Maybe hack a MiniBMS module to trigger a cutoff relay?

lugnut
25 June 2011, 1252
Somebody make this in to a HowTo for me please. :D

Since it's pretty obvious I've knocked out at least a couple of the cells in my pack I've torn it down so that I can check each cell individually to assess the damage. What I need to know is a good DIY way of testing them without buying a proper tester like Travis uses, (although I'd love to have one of those!)

What I have available is:


a lab power supply that I can use as a CC/CV charger to charge the cells individually
a Turnigy watt meter so I can measure voltage and current simultaneously under load
some nichrome wire that I can use to place a load on a cell


So would somone be kind enough to outline a good testing procedure of charging, discharging, recording of data and calculating health of a cell. Can I use this stuff to measure things like IR and AH capacity? What am I going to get out of it at the end to enable me to compare one cell to another, or to say this cell is good and this cell is bad?

So what I would do is use the power supply for charging a single cell by setting the voltage open circuit (leads disconnected). Then connect the cell and set the current control to where you want, say 1C. Then the power supply should charge in constant current at 1C until it hits the pre set voltage and then tapper current down and transition into constant voltage. Leave it there until current decreases to a small value (0.1 to 0.2 amps). Charge the same way for all tests. Let the cells rest for about the same time before each discharge test.

For discharge tests, to find the Ah, you want to have a constant current. This way, all you have to do is multiply the discharge current by the time it takes to get to the LV cut off value and you have the Ah for that rate. You can use the nichrome wire for a resistive load if it will handle your current. Adjust the length and parallel strands for resistance and capacity (so it doesn’t melt). But just a resistance for load will not get you a constant current over the discharge because the current will decrease as the cell voltage goes down or the resistance varies with heat.

So to get a constant current discharge, you can use your power supply CV setting. Put the power supply in series with the cell and resistor. Just measure the voltage across the cell. The current will be the same thru the cell, power supply and resistor. Have the voltage setting on the power supply set to about the same as the cell as the cell. Next adjust the resistor value so the full voltage cell draws a few more amps than you want to test at when connected to the cell only. Then connect the power supply, cell and resistor all in series and adjust the current setting for your desired load. Start your timer. As the cell voltage drops, the power supply will make up the difference and keep current constant. Just monitor the cell voltage. When it hits the LV point, stop the test and your timer. Amps times time = Ah.

Be careful and be sure to stop the test when the cell hits LV otherwise the power supply will keep current flowing and damage the cell.

For internal resistance measurement of the cell, read the voltage at two current loads on the cell, like 1C and 2C. The R_int = (V_2 – V_1) / (I_2 – I_1).

DaveAK
25 June 2011, 2004
Thanks for the suggestions guys!

@Larry - unfortunately the Turnigy doesn't log data or interface to the PC so I'll have to record manually.

DRZ400
25 June 2011, 2034
Single cell charge to 3.6v (single cell charger or power supply), discharge at .5C or 1C with your wire. Record how many minutes it takes for each cell to hit 2.5V....hopefully they will all do 60 minutes or better for 1 C. Compare and sort.

podolefsky
25 June 2011, 2108
Dave, this doesn't need to be that complicated. If you just want to test for bad cells, you can use a fixed resistance. The cells will go from about 3.4 down to 2.8V. With a 0.1 Ohm resistance, that's 34A down to 28A. Not that big a difference. Besides, the cells will spend about 80% of their time right at 3.2V, so it will be constant current most of the time.

At 0.5C, the cells will take about 2 hours to discharge. The bad cell(s) will hit LVC much sooner. You can just read off the voltage and current every 2-3 minutes and get a curve that will tell you what you need to know. (Like DRZ400 said.)

If you really want to get detailed, you can multiply the current at each reading by the time difference and that gives you Ah. You don't need a constant current if you record the current at each time interval.

If you got the 15 AWG nichrome, it's about 0.2 Ohm/ft. Each strand will handle up to 40A. It gets HOT - I put a fan on the wire just to keep it cooler. It won't melt under 40A, but it got hot enough that I didn't want to be next to it for long. Make sure whatever you mount it on is not flammable (seriously...I actually used a piece of wood, but it started smoking so I put aluminum foil on it to deflect the heat. That worked, but don't leave it unsupervised.) Make sure the wire is suspended, not touching anything but the mounts at the ends.

I was testing my whole pack at once, so I used a 10 ft section coiled up and strung between two bolts. My BMS gives voltages for all the cells, so I just read them off and put them in a spreadsheet.

If you want to test at 1C, use two strands in parallel, so that both carry less than 40A each.

podolefsky
25 June 2011, 2125
Also - calculating internal resistance (Ri) for a lithium cell isn't that straight forward. I mean, you can use the delta_V / delta_I formulas, but Ri will change with load, temperature, and degree of discharge. These things all feed back on each other - like high C-rate causes heating, which can make Ri go down, which increases capacity (to a point).

lugnut
25 June 2011, 2244
Dave, this doesn't need to be that complicated.

Yes, you're right. But if he has the power supply, once he does it and figures it out, from then on it's easier than taking reading and plugging it into a spread sheet. And all the published curves from the battery makers are constant current, so he can compare to that. I used to do constant resistance discharges, but once I figured out the constant current method, I never will go back :-)

lugnut
25 June 2011, 2258
Also - calculating internal resistance (Ri) for a lithium cell isn't that straight forward. I mean, you can use the delta_V / delta_I formulas, but Ri will change with load, temperature, and degree of discharge. These things all feed back on each other - like high C-rate causes heating, which can make Ri go down, which increases capacity (to a point).

The delta V/delta I method is very accurate. Just use a fully charged cell at room temperature and currents in the range of interest, like 1 to 3C. Internal resistance does not change much with load around the area of interest. It does vary with temp and SOC, so test at a full charge and room temp. He will get good results. I have found it is pretty consistent from one cell to another, or one battery to another of the same type and size, but often quite different from the manufacturer's spec. They tend to be way optimistic. But the test method is representative of what you get in the application, at least until it heats up and discharges.