Cheap Battery Showdown

[fitek]

Despite some advice to just spend away on expensive batteries, I've decided to buy cheap NiMH, NiCad, and Lithium batteries and let them go at it head to head. I placed orders for 9 batteries.

All batteries are 70 cents/watt-hour or less (except the lithium, which are 71 cents) in quantities required for a 48v ~8ah pack.

Sub C NiMH: 3.5Ah, 3.8Ah, 4.5Ah
AA NiMH: 2.6Ah
C cell NiMH: 5Ah
D cell NiMH: 8Ah, 10Ah
D cell NiCad: 7Ah
lithium ion 18650: 2.2Ah

Some are Tenergy brand, the rest are mystery meat.

Testing Goals
I'm running my bike at 24V. 30A provides enough power for me. If these cells can put out 2C and survive, I'll be happy. The Sub-Cs are rated up to 10C. Therefore, I propose the cells be tested at the following discharge rates:
10C
5C
4C or 3C
2C
1C
0.5C

I need to charge my bike in about 6 hours. So 0.2C charge rate would be sufficient. Most of the cheap chargers seem to only go up to 2A so this is about right for an 8Ah pack. Might be fun to try faster charge rates.

Testing Apparatus

My accomplice and I have a 2 channel digital storage oscilloscope at our disposal. It can record for 2.78 hours. Hence the 0.5C lower discharge limit. For charging, we have some troubles, since the scope can't measure over a long enough interval. Instead an LM431 shunt regulator will probably have to do, signaling a stop watch of some sort that the charge is done.

We're still working on the circuit to make all this happen and input is appreciated. http://www.zeva.com.au/tech/LiFePO4.php has a schematic for a system using LabView, but it's a bit more than seems necessary here.

Comments and suggestions welcome

 

[Ypedal]

AWSOME !!!.. pictures please !!

Oh.. and i consider tenergy " mystery meat with a label " lmao..

We need more of these threads ! i love it .

 

[TylerDurden]

 

[tomv]

My experience with Tenergy NiMH is that their batteries work, but their ratings for both capacity and discharge rates are waaaayyy optimistic. Like capacity 2/3 of what's claimed and discharge rate is multiplied by 3x My experience is with 10 Ah D cells, but there are plenty of reports on places like resellerratings.com.

Bottom line is that price per real Ah is pretty much always the same - if you buy say Panasonic or Sanyo Nimh, they'll just be rated (correctly) at 6.5Ah instead of (overstaded) 10Ah by Tennergy. Batteryspace.com cells seem to be in between, just slightly optimistic.

So that gets you to $1/Wh point - and then you might as well go with Makita or Milwaukee lithium packs that will be better value for high rate/long cycle count use.

Anyway, curious to see your test results! Nothing's better than real measured data...

 

[safe]

I'll especially like to see your SubC data. My plan is to run 10C SubC cells at about 2C to 5C so that as they age and lose capacity (which is inevitable) you still can get you peak amps even though the cells might have 500+ cycles on them. This way you can sort of "cheat" and get effectively more than the rated cycles because you continue to use the batteries after the accepted 80% capacity compared to the start.

The biggest mistake people make is to setup things so that there is such a narrow margin between the spec and the needs of their bikes that the moment the capacity drops the pack can no longer deliver.

 

[Riles]

Wow, this is great. I have been waiting for somebody to do this! I was actually toying with the idea of buying some test equipment and doing it myself a few months ago, but then work got busy and, well, here we are.

I am an avid rc modeler, so I have always been interested in the price point of tenergy stuff, but I know that they overrate. The question has always been "just how over-rated are they?". My girlfriend and I just picked up a pair of Currie eZip bikes (our first electric bikes), and now I'm even more interested in batteries. These things are a blast and I can't wait to upgrade the batteries to get more mileage out of them, as we currently only get about 8 miles in hilly riding.

I'll be watching this thread intently, thanks for doing this!

 

[fitek]

The second batch of batteries arrived today. The Maker Faire is upon us and I will be preaching the religion of electricity to the masses this weekend, so no battery tests-- just yet. Before I upload the battery pr0n, here's a few questions for those smarter than I:


When is a NiMH cell discharged? Wikipedia says 1.0volts is as low as the voltage should drop (for some period of time I assume, rather than momentarily)

When is a NiCd cell discharged? (No idea, guessing 1.0volts as well)

When is a NiMH cell "charged?" Wikipedia says full charged is 1.4volts, apply 1.6v for charging, but watch for cell reversal near end of charge (I suppose I can just use alligator clips to charge the D,C, and SubC cells using my AA nimh charger)

When is a NiCd cell "charged?"

When is a "3.7V" Lithium Ion cell charged? When is it discharged

The test rig will be very large resistors with heat sinks, maybe even heating coils, and an LM431 shunt regulator to signal when the battery is discharged (I don't want to kill the batteries, just test them).

How does that sound?

 

[tomv]

When is a NiMH cell discharged? Wikipedia says 1.0volts is as low as the voltage should drop (for some period of time I assume, rather than momentarily)

You can discharge down to 0V and it won't kill the cell completely but it will hurt it. If you "reverse it" - flow current thgouth the cell backwards, that will seriously damage it. This is still much better than lithium - discharge it to 0V and it's dead for good.

When is a NiMH cell "charged?" Wikipedia says full charged is 1.4volts, apply 1.6v for charging, but watch for cell reversal near end of charge (I suppose I can just use alligator clips to charge the D,C, and SubC cells using my AA nimh charger)

Uhuh. That could generate some smoke. You can't fast charge NiMh using voltage cutoff. You need to monitor temperature (or better dT/dt). NiMH can be trickle charged at very low rate but that hurts them too. Just no explosion.

Also - at the end of charge cells don't reverse, but sometimes dip voltage slightly. This is unreliable and depends on charge rate, temperature, etc.

When is a "3.7V" Lithium Ion cell charged? When is it discharged

Depends on particular chemistry. LiCo is usually charged up to 4.2V and then kept at 4.2 until current is bellow 100mA (depends on cell size). To prolong cycle life some people charge to 4.0V and get 80% capacity but more cycles.

Discharge cutoff also depends on particular cell. Usually 3.0V for LiCo.

 

[safe]

NiMh Timed Charging

I'm considering getting four low priced high amp chargers for some NiMh SubC cells in a 25 cell (30 volt) This would be set up so that two sets of 30 volts would produce 60 volts so that I could use that new 60 volt controller from TNCScooters.

36V @ 4.2A 150W POWERSUPPLY1 POWER SUPPLY - $18.95 each

http://www.powersupplydepot.com/productview.asp?product=13657+PS

...and then just use a timer ($22.99):

http://www.westsidewholesale.com/index.cgi?HN_SessionID=@@@@1209773874.29891@@@@&sort_order=sales&pid=973416&CATEGORY=376&back=48

...and what I would do is charge for what I expect it needs (since I would know how far I rode) and then take some voltage measurements to see how close to full I was. If the voltage showed a 1.4+ voltage level then I'd be done. This at least limits the damage because the charger would be unable to overshoot and miss a delta signal which is apparently easy to miss with NiMh anyway. Temperatures would be capped because the timer goes off before the heat could build up too much. It would probably be a good idea to charge in steps... maybe do 30 minutes, measure, then another 10 minutes... then maybe 5 minutes. After a while you would get a sense for what it needs after each ride and with a fast charger (1C) you could ride several times per day. (it makes the smaller SubC packs more useful)

 

[fitek]

Well I sidestepped Nimh and NiCd charging issues by using my camera battery charger.

Now, battery pr0n, as promised.

I will omit the picture of the charging setup. It's too explicit for these forums, but briefly, it involves tape, alligator clips, and c clamps. You could have a party with list of ingredients like that.

 

[safe]

Hey where did you get this one?

At 14Ah that's a lot of capacity for a "D" cell and I've got a bike that was originally designed for "D"'s but I was looking to use SubC's as a replacement because the normal 10Ah is rated at 1C and that's not enough to give the amps I want.

Where to buy and what price?

And I'm going to be very interested if you can test those to find out if they really deliver 14Ah...

 

[D-Man]

Looks like you can get them here but they are discounted because they have capacity problems.
http://www.batteryjunction.com/marathon-d.html

 

[kbarrett]

Looks like you can get them here but they are discounted because they have capacity problems.
http://www.batteryjunction.com/marathon-d.html

Translation ... the manufacturer puts lies on his cells.

 

[fitek]

Yes, I saw that the "14Ah" cell is actually 8.5. We'll see.

The tests have begun! The Tenergy AA is now running at 5C and not holding up well at all. The voltage sagged to 9.38V immediately, and it started dropping fast after 225 seconds. We've guesstimated the actual capacity at 5C to be roughly 0.63Ah. The scope data was fed into Excel, but we're too lazy to figure it out right now.

Running another AA at 2C now. It sagged to 1.11V at the start. It's settled in at about 1.08V, and is running OK.

 

[fitek]

Well, we finished the 2nd AA (600s at 2C to get to 0.9V), and proceeded onto the 10Ah Tenergy D cell. It was actually more disappointing than the AAs. Unlike the AAs, it had a very linear voltage drop. The AAs drop slightly at the beginning, then hold steady for a while, then drop all of a sudden. The D cell dropped slowly in a linear fashion, then all of a sudden took a nose dive. It also lasted roughly 600 seconds at 2C.

If the premise that all batteries cost roughly the same amount per real watt-hour is true, then the only other major considerations will be lifespan and pack complexity. We're not measuring cycle life or shelf life in these tests, so I can only suggest that it will be much easier to make a D cell pack than a AA pack.

We've decided to change the tests to save time; rather than using the logging scope on EVERY test, I'm going to use my Fluke for 2C tests on all the batteries to determine which are the best 2 or 3; then we will use the scope to further test those cells.

I've also decided to add my Sanyo AAs to the test, and maybe some known "good" cells like an A123 as a comparison.

 

[safe]

Battery Warmups / Stretching

Many chemistries don't give their best results until they are "warmed up" with a few excercise cycles. Even my SLA's that had sat dormant all winter are now starting to spring back to life as I use them again. Batteries that sit for a while tend to settle into a stable state and you need to sort of shake them up with use to get them active and flexible again. I suspect it's the activity that actually wears them out, but it's also how you get them to really perform.

So don't take your first test as the true behavior of the cells... keep trying...

 

[fitek]

I tested the no name NiCd when I got home (the train broke down? what's up with that? I got back after midnight) and it fared worse than the NiMH d cells. I also tried the 4.5Ah Sub-C, but it went down to 9V in two minutes. I will recharge it and see-- maybe I have a dud.

 

[tomv]

It's great to see your results! I kept looking at those Tenergy batts because the price is so tempting. It looks like their products suck across the range . I'm still hoping you'll find a gem in the rubble. How about that LiIon 2.2Ah cell? That's a pretty conservative rating for 18650 so it just might be ok quality. I would not try to discharge it at 2C. They're designed for more like 0.5C.

As for battery selection criteria, I would say capacity and discharge rate capability are important but by far not the only criteria you should look at. Other that matter quite a bit are:

- Self discharge. NiMH and NiCD are horrible in this. If you don't use your batteries often you can assume they are 100% empty when you pick them up. Exception: SANYO ENELOOP, great, but price vice not an option.

- Cycle life: If you count total Wh you'll get out of batts before throwing them out picture changes big time. Like 200 cycles for SLA, 500 for good NiMH (and more like 10 cycle for bad chinese crap), 3000 for A123.

- Energy density and power density. If this is not an issue for you go SLA. Nothing better than that in terms of shelf life, self discharge, current capability, convenient form factor.

- Operating temperature range. Generally nothing works ok outside of 15-30C region. But some are better than others

- Mechanical issues. A123 and many other lithiums have aluminum cans that make it hard to weld/solder them. Nimh/NiCD need gazillion cells per pack, so those become unreliable (welds/solder joints fail). SLA - heavy.

 

[mingonn]

Hanging out for more results

I am trying to set up my first ebike and am tossing up between sub c RC car packs and rc lipo at the moment.

About the same price for 36v at 4000mah but I am guessing NiMh will be a bit more abuse tolerant. Some super cheap 12C lipo 5000mah packs on ebay, even at a conservative(probably realistic) 4C discharge I should still get 20 amps continuous

Derek

 

[fitek]

Sorry for the delay. I'm off for a few days tomorrow (again), but I will jump back on this next week. Trying to finish my tiki bar by June 7th...

 

[Riles]

Sorry to be a bother, but have you had a chance to test any other batteries?