battery analysis

[SilverFox]

Hello Evan,

Here is the section on testing for internal resistance from the manual:

"Internal Resistance Testing

You can get a good idea of the quality or condition of a pack by measuring the voltage drop at different test current levels. This will give you a measure of the packs internal resistance. High internal resistance is a sign of a poor pack, one that gets hot and is inefficient.

Charge the pack, and set the CBA for example to 11 amps, assuming the battery is safe at that level. Run the test for 30 seconds and stop the test and record the voltage at the end. Run the test again at perhaps 1 amp for 30 seconds and record the voltage again. If the battery has .5 volts less at 11 amps than it does at 1 amp use Ohms law: R=E/I. R being resistance in Ohms, E is voltage in volts and I current in amps. The internal resistance of the pack, in this example, would be: R = .5/10 or .05 ohms, note that we used the difference in current between the two tests of 10 amps. The lower the resistance the better."

You have to start a new test for each measurement, but can stop the test at any time. Saving the data to file allows you to overlay the results.

Tom

 

[evan9162]

Ah, so it's not automatic.

One thing they don't mention is that you'll have to take into account the resistance of connectors and wire as well - so you'd have to devise a method of doing that so you can subtract that amount from what you calculate from the different current levels.

Thanks for the info...

 

[Tweek]

It seems my Excel knowledge is lacking in the line-graph making area. I can import the .csv (comma delimited text) file the software generates just fine, I just can't get it to create something useable. Maybe I need to twiddle with it some more.

As far as low current operation, the sense resistor might be a factor, or DAC resoultion perhaps. But the leads and connectors they've chosen (13GA high strand count/powerpole connectors) are pretty much not a factor at all below 1A currents. I doubt the MOSFET's resistance too, because all the power to run the unit is derived from the USB port, the battery under test doesn't contribute to anything but the heat load.

Now under high currents, the plugs/wires, MOSFET, sense resistor, circuit traces, everything becomes a factor in determining maximum load current. Sofar, the most current I've been able to pull (stable load) with just 1 cell has been 23A, using a 3000mAh sub-c cell. I've got some of the Aerobatt 4AH 1/2D cells Ginseng made his last batch of UMC's out of, I may try 2 or 3 in series, to see what they're capable of... /ubbthreads/images/graemlins/smile.gif

Chris

 

[SilverFox]

Hello Chris,

The thing you are missing in the exported file is the amp rate the test was done at. The export includes time in seconds and the cell voltage. To get amp hours, you need to include the discharge rate.

I had another chat with the fine people at West Mountain Radio and they may visit us here. The most requested missing feature so far is the ability to test below the 0.5 amp rate. I believe the problem is accuracy in measuring the discharge rate when a low rate is selected.

I also asked them to include the discharge rate in the export file in their next revision.

There is a rumor that they may offer an upgrade that will be capable of lower test rates for a reasonable fee. You will have to send your unit in for the upgrade, if it is offered.

There is also a rumor of a "Professional" model. It would have a lot more features and capabilities. Cost estimates are around 4X the price of this unit.

Tom

 

[wptski]

Got my CBA yesterday and must make up some leads tonight!

Either there is a glitch in the CD-ROM or I missed something because it onlys loads the drivers! I had to go into the Program folder and find the autorun.exe to load the application itself.

Anyone else run into this too?

 

[SilverFox]

Hello Bill,

Mine loaded everything without any problems.

Tom

 

[wptski]

After reading the manual, it states that after connecting a battery, it shows the resting voltage, yet that value is changable. So, lets say that you connected a single Li-Ion cell and it reads 4.15V, do you edit it to 3.6V or 3.7V nominal voltage. It doesn't mention setting the voltage in the manual. I didn't look in the Help section.

 

[evan9162]

[ QUOTE ]

It seems my Excel knowledge is lacking in the line-graph making area. I can import the .csv (comma delimited text) file the software generates just fine, I just can't get it to create something useable. Maybe I need to twiddle with it some more.


[/ QUOTE ]

If you are given a time column, and a voltage coumn, then you could try making an X-Y graph.

 

[SilverFox]

Hello Bill,

I am not sure what that setting is for. I believe the program looks to see if the watts are within internal guidelines, but it does not effect the graph or the discharge rate. It may have to do with the programs guess on the number of cells...

I'll have to try putting in 12V on a single NiMh cell and see what it does...

Tom

 

[SilverFox]

Hello Bill,

On NiMh you can go up to 2.0 volts. If you go above that, it changes the number of cells. I set mine at 1.4 volts for NiMh and 4.2 volts for LiIon. Not sure if it makes any difference or not.

I was playing with some Energizer AA batteries and came up with this:

I think I got all the information in it.

The X axis is related to time by taking the Amp Hours and dividing them by the Amp Rate.

For example: The 2.3 amp rate test rated the battery at 1.9 amp hours. Taking 1.9 / 2.3 = 0.826 hours or about 50 minutes.

Likewise, the 0.5 amp rate rated the battery at 2.25 amp hours. Taking 2.25 / 0.5 = 4.5 hours or 270 minutes.

Comments?

Tom

 

[Tweek]

Silverfox: Yes, I did notice that the exported data was rather limited. It's not too hard to manually generate an approximate current column, though.

If they do offer an upgrade, I'll just drive back out there... /ubbthreads/images/graemlins/smile.gif

If I understood correctly while I was there talking with them, the Pro model is going to happen. If so, I can't wait, if this unit's any indication... /ubbthreads/images/graemlins/smile.gif

Wptski: I had one that was even better, my install CD was totally blank, never burnt! Didn't stop me though, I downloaded the software off the website, and installed it. Luckily it was as I had hoped, the drivers were included with the software, and I just had to point windows to the drivers directory where I installed the program, good to go. The folks at West Mountain were nice enough to promptly respond to my email, and provided a link to download an image of the entire CD. I really should burn it and see what else is on there.

Chris

 

[SilverFox]

Hello Chris,

Bummer about the CD, glad you found the download.

This is a lot of fun. I decided to check out the difference between a rechargeable 123 and a SureFire 123 at a 1 amp load and came up with this:

I did not know where to cut the SureFire cell off, so picked 2 volts. After this test, I continued to 1 volt and it dropped pretty fast. I think I got most of the data.

I should also mention that this is a SureFire battery I had laying around the house. It may have had a small amount of use prior to the test. When I get everything figured out, I'll put a brand new one in and check it out.

Tom

 

[evan9162]

You can get watt-hours of each by integrating the area under the curve. You can do this in excel by doing the following:

(t2-t1)*I*( (V2+V1)/2 ) for measurements 2 to N, then sum them up

basically, you're multiplying a timeslice (t2-t1) by the average power over that timeslice (I * average voltage), then summing up the timeslices.

 

[SilverFox]

Hello Evan,

Email sent.

Tom

 

[Tweek]

Ok SilverFox, send me a PM, make me a step-by-step blurb with how to make that graph, would you please? /ubbthreads/images/graemlins/smile.gif

I've been abusing a 14500 lithium cell lately, turned out to be a decent cell, have a look:

Chris

 

[evan9162]

[ QUOTE ]
SilverFox said:
Hello Evan,

Email sent.

Tom

[/ QUOTE ]

Thanks for the data.

Results:

SF 123A - 1.80 W-h
R 123 - 1.90 W-h

So, while under a constant current load, the SureFire cell lasts longer, over the course of the load, the R123 cell actually delivers more total energy (due to the higher voltage).

So, if it came between the R123 cell and the SF123 cell in something like a badboy (say, driving a 5W luxeon) it appears that the R123 cell would be slightly better for runtime.

[edit]Made a stupid mistake, updated values. Thanks for double-checking, SilverFox.

 

[yaesumofo]

I thought a few people would be interested in this thing.
It has been working great for me as well.
PowerPole conectors ROCK!
Yaesumofo

 

[Tweek]

Powerpoles are good, but I wouldn't want to make a lot of them without the crimper... /ubbthreads/images/graemlins/grin.gif

Chris

 

[SilverFox]

Hello Yaesumofo,

Thanks for letting us know about this tester.

An interesting thing happened last night. I tested a Pila 168S. This battery has built in under-voltage protection, so I set the ending voltage at 1. When the test was completed, it registered 0.8v. Does this seem right?

I charged the Pila back up and noticed that it was at 3.73v when the green light came on. I put it back on the charger for a couple of hours and it remained at 3.73v. I decided to re-test the battery. As soon as I put a load on the battery, the voltage went up to 4.1.

It seems that someone else has noticed Pila's at 3.7+ volts. What causes this? I assume that it has something to do with the protection circuit.

Tom

 

[Tweek]

I wouldnt' be surprised about the end voltage, seeing as the circuit isolated the cell from the load, and won't reset until it's charged. The 3.7V peak does seem a bit odd though. Could be the circuit's just acting odd, I don't know. I don't have any Pila's handy to play with right now.

Chris