Some IMR (and others) discharge testing...

[flashflood]

If we can agree on some common test currents for certain size cells, I'll gladly do additional tests. I just don't want to do a bazillion more tests as this little endeavor took the better part of a week+. It's the recharging of the cells that I hate...

-DF

For the smaller cells (18650 and below) I would suggest 1.4A and 2.8A because they are such common drive currents. Moreover, adding those two curves should be sufficient to allow interpolation. For the larger cells, I'd suggest two currents between the min and max -- ideally logarithmically spaced, i.e. the ratio between currents should be the cube root of max/min. That will give you four nicely spread curves, two of which you have already generated.

 

[csshih]

DFiorentino, can you come up with any reason why the AW IMR14500 performed so much better at higher discharge rates than AW's IMR16340's and IMR18350's? This doesn't make any sense to me.

14500s were developed by battery manufacturers much more than 16340s and 18350s (which are arguably niche hobbyist cells) - I know the 16340s are completely niche cells not used anywhere else.
14500s are used in camera battery packs and the related.

Craig

 

[Colonel Sanders]

Very interesting thread!

Does anyone know how well the IMR26650s perform around 7-8a?

 

[xul]

One thing I noticed with battery testing is that some people try to space the currents equally on a log scale, i.e., 1, 2, 5, 10 or 1, 3, 10, but I think it's more informative to closely space the currents when the curve is rapidly changing.

This would require two tests; one to see where capacity vs. current is rapidly changing and the next to pick suitable currents.

The short gaps can be filled by graphing capacity vs. current and eyeballing the intermediate zones or fitting a curve. One time, with curve fitting, I got a correlation coefficient, r, of ~0.9 doing this.

 

[DFiorentino]

Update: I'm redoing everything using the DCHG+ expanding discharge function of the charger. It uses the balance port to monitor the cell voltage separately from the current carrying wires. I made a balance connector that has its wire soldered directly to the brass contacts along with the load wires. I just did a 23A discharge and the voltage difference was within 0.1%. So another few days of discharging and recharging and I'll be right back where I was. After I redo the original work then I'll take on additional discharge requests.

-DF

 

[Black Rose]

Nice work DF :thumbsup:

 

[jasonck08]

Update: I'm redoing everything using the DCHG+ expanding discharge function of the charger. It uses the balance port to monitor the cell voltage separately from the current carrying wires. I made a balance connector that has its wire soldered directly to the brass contacts along with the load wires. I just did a 23A discharge and the voltage difference was within 0.1%. So another few days of discharging and recharging and I'll be right back where I was. After I redo the original work then I'll take on additional discharge requests.

-DF

Cool, and you should also discharge to 2.5v. Trust me, there is a lot of mAh left between 3 and 2.5v at higher currents (due to voltage sag). Sure not all the voltage will be usable in all applications from 3v down to 2.5v, but the cells capacity is rated to be discharged to 2.5v.

You should also test the current accuracy of your hobby charger (verify it with a DMM at lower currents). I know the iChargers are usually pretty good within 1-2%, but sometimes other hobby chargers the current reading is literally 10% off.

 

[DFiorentino]

Things are looking better as far as repeatability and accuracy. I'm confident in the voltage signal now and I just did a current check; it's within 0.000A-0.005A @0.05A discharge, but also 0.000A-0.005A @0.50A discharge. It looks like its resolution is limiting its accuracy (by percentage) at very low currents, but for my purposes of 0.50A and up, any variation is probably not practically appreciable. As far as cutoff is concerned, apparently I'm locked into the 3.0V cutoff. Using the LiPo selection (3.7V/1S) the lowest discharge is 3.0V. If I could select LiIo (3.6V/1S), I can go down to 2.5V, but it won't let me select that option as the battery peak voltage detection expects 4.1V and I get an "over" warning. Oh well. One change I am making is I WILL NOT be using my Ultrafire chargers for this testing anymore. While my completed charge voltages were varying by only small amounts, I found the available capacity varying significantly from charge to charge. I plan on using the iCharger or my Triton (@1C) or the 4Sevens charger as they have proven to be fairly consistent.

-DF

 

[jasonck08]

Things are looking better as far as repeatability and accuracy. I'm confident in the voltage signal now and I just did a current check; it's within 0.000A-0.005A @0.05A discharge, but also 0.000A-0.005A @0.50A discharge. It looks like its resolution is limiting its accuracy (by percentage) at very low currents, but for my purposes of 0.50A and up, any variation is probably not practically appreciable. As far as cutoff is concerned, apparently I'm locked into the 3.0V cutoff. Using the LiPo selection (3.7V/1S) the lowest discharge is 3.0V. If I could select LiIo (3.6V/1S), I can go down to 2.5V, but it won't let me select that option as the battery peak voltage detection expects 4.1V and I get an "over" warning. Oh well. One change I am making is I WILL NOT be using my Ultrafire chargers for this testing anymore. While my completed charge voltages were varying by only small amounts, I found the available capacity varying significantly from charge to charge. I plan on using the iCharger or my Triton (@1C) or the 4Sevens charger as they have proven to be fairly consistent.

-DF

You should be able to use the Ni-mh discharge setting and program in 2.5v cutoff. It's usually not as picky. At least thats the way it works with the cheap Turnigy style chargers.

 

[Battery Guy]

And this was the data that led me down this curious road:

FM5x26500 host, Kiu socket, 64458 bulb, DD, AW 26500 cells... That's 1500 lumen lost in 44 seconds.

-DF

DF

How did you make these lumen measurements?

Cheers,
BG

 

[DFiorentino]

How did you make these lumen measurements?

In my DIY Integrating Spheres...

-DF

 

[Battery Guy]

In my DIY Integrating Spheres...

-DF

That is fantastic! I want one!

Any chance you might make these for sale on CPF?

 

[DFiorentino]

As much as I would appreciate an income right now, both of my spheres were incredibly tedious to make. So, sorry, no plans to start a production run at this time. But I appreciate the compliment.

-DF

 

[alexborr]

Question: I AM FROM ITALY; where can i buy IMR batteries without spending too much?
I am avoiding many USA stores because of 15-20$ shipping fees :sick2:
thanks very much guys

PS:
hi guys this is my first post on CPF but i love this website since years! lovecpf <3
i buy cheap flashlights from D....X..... (i don't know if i can type store names in the forum)
i am going to:
1) mod an akoray flashlight with a XLM T6, so i need a IMR battery because my 18650 normal cells can supply only 1,5 A -.-
2) mod any 10440 and 14500 flashlight i have, with these awesome cells, because i dont care about capacity, and i like so much this insane Current output.

 

[alexborr]

this test has been very usefull to me, ty!!