Large scale battery test done by me as part of an internship

[tom66]

As part of my internship at BitBox Ltd. (a local firm specialising in electronics engineer and embedded software design), I was set the task to prototype, design, manufacture (schematic capture, PCB layout, soldering SMD etc.) and program a battery tester. I wrote software and protocols on both ARM microcontrollers and Windows PCs. The results were VERY surprising... it seems that cheap batteries perform very well compared to expensive Duracell and Energizer cells...

http://www.batteryshowdown.com/

I also designed the website for the results.

Comments welcome!

[madires]

Great job! Some tests in local magazines indicated too that no-name batteries from the nearest super market offer the best value for the buck.

[jucole]

Superb!! and as a web designer myself, your site is great!  I like the battery line-up pic best! ;-)

[nitro2k01]

Do you have full discharge voltage curves available for all batteries? Would you consider making all the raw data available in some machine readable format, such as CSV?

[tom66]

Do you have full discharge voltage curves available for all batteries? Would you consider making all the raw data available in some machine readable format, such as CSV?

There's a whole lot of data to be categorised and uploaded at some point, but for now I've only done graphs for the three main chemistries of batteries. I would like to re-do all of these tests at some point to get some duplicate data to compare against tolerances. But I'm leaving for uni tomorrow so haven't got enough time.

[Monkeh]

Very interesting.

It's nice to see the Sony batteries I've been buying for years do pretty well. Have to keep an eye out for those Kodak ones, though.

[tom66]

They sell them in Poundland. The real surprise was the Costco Kirkland Signature batteries, which came 2nd for an alkaline in terms of capacity yet they only cost 21p each. We buy them all the time in packs of 48 and find them to be good in cameras, comparable to the Duracells, well the test shows BETTER!

[Monkeh]

They sell them in Poundland.

£1.50 plus petrol to get there.. I think not.

The real surprise was the Costco Kirkland Signature batteries, which came 2nd for an alkaline in terms of capacity yet they only cost 21p each. We buy them all the time in packs of 48 and find them to be good in cameras, comparable to the Duracells, well the test shows BETTER!

I stick to lithiums in my camera. I can fill a card several times with a set of lithiums, and they don't mind sitting around. I also feed them to my torch.

[tom66]

I stick to lithiums in my camera. I can fill a card several times with a set of lithiums, and they don't mind sitting around. I also feed them to my torch.

For low drain applications, they aren't worth it - when in higher drain applications they start to shine though and may be worth it. The results are very close to call and very dependent on the particular camera.

[retiredcaps]

I use Sanyo eneloops or Duracell pre-charge/stay-charge for my Canon A80 camera.  Rumor has it that the made in Japan Duracell white top pre-charge/stay-charge are Sanyo Eneloops.  When they are on sale, the Duracells are about $2 CDN each.

The local Canadian Costco used to sell the Kirkland 48 pack AA for $10.99 (or 23 cent each), but jacked up the price to $12.99 (or 27 cents each).  Rumor has it that Duracell makes the batteries for Kirkland.

[Monkeh]

I stick to lithiums in my camera. I can fill a card several times with a set of lithiums, and they don't mind sitting around. I also feed them to my torch.

For low drain applications, they aren't worth it - when in higher drain applications they start to shine though and may be worth it. The results are very close to call and very dependent on the particular camera.

My experience is that my camera kills alkalines (it can't cope with the low voltages). My torch also works much better with the lithiums, it may be more expensive, but I get much more usable life out of it. Picking up a torch to find out it's dead is less than useful.

[retiredcaps]

http://www.batteryshowdown.com/

I also designed the website for the results.

Excellent work and website. It loads fast, easy to sort the columns, devoid of advertising and I love the photo of all the batteries side by side.

[tom66]

The local Canadian Costco used to sell the Kirkland 48 pack AA for $10.99 (or 23 cent each), but jacked up the price to $12.99 (or 27 cents each).  Rumor has it that Duracell makes the batteries for Kirkland.

They are definitely some type of Duracell variant, possible Procell but better performance, as the end cap has two little dots on it which are in the same place on every other Duracell and no other battery I have found. However, the Kirkland are made in USA, whereas most Duracells come from Belgium. Either Kirkland buy the batteries in and relabel them (sounds unlikely to me) or Duracell has additional factories in the US for the US market.

http://www.batteryshowdown.com/

I also designed the website for the results.

Excellent work and website. It loads fast, easy to sort the columns, devoid of advertising and I love the photo of all the batteries side by side.

Thanks - I just realised though, I forgot the Kirkland in the battery line up. I think every other battery is on there.

It does have advertising, I suppose - it's designed to promote BitBox Ltd. and demonstrate some of the capabilities the company has, all five analysers were built in the UK; one by myself and the other four by the production team, they were designed by me personally but that wouldn't have happened without the support from the three other engineers (Matt, Pete, and Quentin - the manager/boss - who is also the engineer (and does some of the engineering work for the company still), which is rare for today.)

[retiredcaps]

Any chance of a 9V shootout when/if you have time away from your upcoming studies?  9V, obviously, powers most of the older generation of multimeters.

[tom66]

Any chance of a 9V shootout when/if you have time away from your upcoming studies?  9V, obviously, powers most of the older generation of multimeters.

If I get the time I'd like to test C/D/AAA cells and 9Vs, as well as rechargeables. I also want to test some nickel oxyhydroxide cells at some point, I wasn't able to get any easily in time.

Also, I was going to post it on Candle Power forums but I don't have enough posting priveledges, if anyone has the ability to do so I would really appreciate it. Thanks.

[NiHaoMike]

Using a 168MHz ARM for that data logging is massive overkill. I'm surprised it didn't just use a PIC18.

[Mechatrommer]

maybe the PIC18 $/Hz is overkill?

[tom66]

Using a 168MHz ARM for that data logging is massive overkill. I'm surprised it didn't just use a PIC18.

Of course but the STM32 module only costs £10. I would have used an STM8 or STM32L if I didn't use a module but why not use something I can be sure works and reduce the development time.

[AndyC_772]

Using a 168MHz ARM for that data logging is massive overkill. I'm surprised it didn't just use a PIC18.

If the object of the exercise is to learn to use and program an ARM processor, it's not a bad choice at all!

Either way, I'm really glad to see that someone has been able to carry out this research - and what a shame battery manufacturers aren't obliged to state the actual capacity somewhere on the packaging. (I believe there's a rule coming in fairly soon which makes it compulsory to state this for rechargeables, but not for primary batteries).

Off to Farnell for a pack of 20 Procells... for some reason it's £4.29 for 10, and £4.38 for 20. Go figure!

[tom66]

The 20-pack Procells cost less than the 4-pack Ultra Power batteries, which are insanely overpriced... But I suppose they are just relying on a good reputation, few people have actually tested a large number of batteries like this in such a methodical manner.

[Fraser]

A word of warning....

I have ceased buying anything made by Duracell, including Procells. They are too prone to leaking and I have had enough of cleaning out equipment battery compartments after a leak ! I have even had them leaking whilst still in the box awaiting use ! Total crap IMHO.

I have jsut started using teh cheap Kodak Xtralife cells and had no problems to date. No leakage and decent performance. At GBP1 for six I will be stocking up on these, especially after the test results shown here.

Fraser

[bullet308]

There is an additional variable that, of course, would be less practical to test, and one that is very important to me personally: shelf life, where the lithiums would really shine. I have, among other things, an assortment of flashlights  that need to sit for, oh, a decade or so, and then perhaps be used under difficult conditions. If somebody could prove to me that Energizer Ultimate Lithiums really have a usable shelf-life of fifteen years as-claimed, THAT would be massively cool.

Inevitably, your survey is a bit UK-centric. I suppose it has to be somewhere-centric, at least for starters. Perhaps a mail-in battery donation campaign from around the world? For instance, it would be nice to get some US market value brands (Ray-O-Vac, Enercell by Radio Shack, Sunbeam from Dollar General, etc) into the game.

A very good effort at a very difficult thing, though. This has to be the most comprehensive test of its kind to-date. Great web site as well.

[saturation]

Great job on your project, both in making the circuits and the final report format.  Easily an A for a grade.  These days I only use LSD NiMH batteries and leave alkaline to very low drain devices like TV remotes and wall clocks.

[bilko]

RayOVac rechargeable alkaline cells leak as bad as Duracell, and that's using the manufacturer's own charger. Don't even go there, utter JUNK!

[Jon Chandler]

There are sure a lot of negative comments on Hack-a-Day about this.  Makes a person really question whether the recognition is a good thing.

Too many of the people there are of the mindset that they know the best way to do anything (everything?) and any other approach is wrong.  Of course, they know everything, so they never actually do any thing.

[IanB]

Also, I was going to post it on Candle Power forums but I don't have enough posting priveledges, if anyone has the ability to do so I would really appreciate it. Thanks.

Done. See here:

http://www.candlepowerforums.com/vb/showthread.php?344201-Large-scale-alkaline-and-other-primary-chemistry-battery-performance-test

[billclay]

It would be helpful to see battery dimensions since you have all of these available to you.  Some batteries are a tight fit in certain equipment.   Thanks.

[rollatorwieltje]

Isn't 200mA huge for an alkaline cell? I basically only use them for clocks. I didn't bother to measure the current draw of an average quartz clock, but given how long a battery lasts it should be well below 0.5mA. In this case it would be more interesting to see the difference at extremely low current draw. Maybe some data can be derived from the internal resistance of the cell.

[tom66]

0.5mA would take... assuming 2000mAh... about 4000 hours (half a year.) I didn't have any where near this amount of time at my internship, the tests were done in about 2 weeks!

[SeanB]

200mA for an alkaline can be low, as a torch bulb can draw 1A or so, and a camera can have pulses of 4A or more during charging of the flash capacitor. A clock typically draws under 1 uA, with a pulse of under 100mA for a few microseconds every second as it operates the clock motor.

[saturation]

You can easily project modified AH for different loads uses Peukert's law.

http://en.wikipedia.org/wiki/Peukert%27s_law

The trick is establishing the Peukert constant of your battery and is a function of battery wear.  If you presume a new battery and discharge it at various rates you can set it as one.  Also, NiMH, Li chemistries will discharge close to k=1 even if old,  these cells are more limited by storage than by internal resistance.

One thing you can do is establish 2 discharge curves, one at 1A and another at 0.1A.  The difference in measured AH for that battery will allow one to calculate k for that battery.

[LaurenceW]

An interesting quick calculation - Electricity from primary cells such as these is 500-2000 times more expensive than the same energy from the mains!

[orbiter]

Great work here on these battery tests tom66.

I'm another one that gave-up using Duracell batteries a good while ago, mainly because as others have mentioned, I'm noticed they were starting to leak a lot sooner than they used to. In fact when Duracell first became available I can't remember them suffering much with leakage problems at all.

Anyway I decided to have a read through a few online battery tests etc, and what surprised me was that GP Ultra batteries generally came out pretty well in such tests, so I'm pleased that in these tests.. especially as I've been using them for ages, that the GP Ultras are still showing pretty high up on the list  

[tom66]

http://www.batteryshowdown.com/index.html

I have just added 12 new batteries, and Ikea won. Yeah. Ikea. The flat pack company.

2270mAh @ 200mA discharge, repeated for verification (got 2260mAh so considered it to be fairly consistent), 10p each cell (box of 10 for £1.)

 

And as requested, discharge graphs, and variable cut-off voltages. Plus some limited source code.

[Conrad Hoffman]

Great tester, great test, great web presentation! I've done something similar manually on a very small number of batteries and reached a similar conclusion that the bargain store batteries perform very well. The previous comment about leakage is extremely important. Though the performance of zinc is dismal, they don't destroy the device you use them in if left discharged for a brief time. Akalines are a disaster in that regard and I try not to use them unless required. The physical construction, I think, makes them inherently more apt to leak. All my AA wall clocks and similar get zinc batteries from the discount store. The temperature performance of zinc is bad, so my outdoor weather station modules get akalines. Again, great job!

[tom66]

I have to disagree about zincs leaking.

I have a pile of batteries which have been discharged. All but one zinc battery has leaked; the alkalines are fine.

Alkalines take a lot longer before leaking.

[lewis]

...Though the performance of zinc is dismal, they don't destroy the device you use them in if left discharged for a brief time. Akalines are a disaster in that regard and I try not to use them unless required. The physical construction, I think, makes them inherently more apt to leak....

Strange, I can't remember ever having a leaky alkaline, but have had many problems with zinc chloride cells leaking if they're left discharged for a while. Although other users report leaky alkalines too. Time to change all the batteries in my multimeters maybe?

[IanB]

The leakage mechanisms between the two kinds of cells are different and that can affect the outcome.

Zinc cells leak when the electrolyte eats through the zinc shell and then finds a way through the steel outer casing. This is a slow and progressive process and will generally only happen if the cell is completely dead (since consuming the cell consumes the zinc outer electrode). The leak itself is a slow and creeping leak due to a fairly dry electrolyte paste and you have a chance to catch it before it gets too bad.

Alkaline cells leak when chemical side reactions occur inside the cell and produce hydrogen gas. This gas builds up pressure inside the casing like a pressure cooker until the seal bursts, and then the cell spills its guts with a fairly rapid splurge all over the insides of your equipment. The process is relatively fast and shows little sign of impending doom. The process may also happen to fresh cells, not only discharged cells, since the gas producing side reaction can be triggered by impurities and defects during manufacturing, or bad storage conditions like excessive heat.

So while zinc cells do eventually leak, I find them to be much more stable and predictable. On balance the risk of leak damage is lower.

[Fraser]

I can confirm that I have had many Procell (Industrial Duracell) alkaline cells can leak badly whilst in equipment, and even whilst still in the new carton! Cells have been in date and not discharged. The manufacturer admitted to me that any application that applies "excessive pressure" to the ends of the battery can cause it to leak. In my case my D cell Maglite regularly caused battery seal failure.

I no longer use Duracell or Procell and the battery leak problems have ceased. I now use Kodak and Energizer Alkaline batteries.

Fraser

[billclay]

I can confirm that I have had many Procell (Industrial Duracell) alkaline cells can leak badly whilst in equipment, and even whilst still in the new carton! Cells have been in date and not discharged. The manufacturer admitted to me that any application that applies "excessive pressure" to the ends of the battery can cause it to leak. In my case my D cell Maglite regularly caused battery seal failure.

I no longer use Duracell or Procell and the battery leak problems have ceased. I now use Kodak and Energizer Alkaline batteries.

Fraser

Very interesting about the pressure on the battery end caps!  Some of my meters have a VERY TIGHT AA battery fit with "tab" contacts on the meter.  Not the kind with "spring" contacts.  There were three batteries, two Duracell, one Energizer (OK, I shouldn't mix battery types) and the two Duracells leaked all over, the Energizer was fine.

Also, it's worth noting what kind of batteries the manufacturer supplies with it's meters, Fluke supplies Energizer.

I think I just might put one of each type in a clamp, apply a little pressure and see what happens!
Has Kodak or Energizer commented about end cap pressure?

[G7PSK]

I can remember the Zinc batteries not only leaking but swelling up and getting stuck in torches and other equipment with tight battery compartments, all that disappeared with the introduction of the alkaline cells but recently they have started to have a leakage problem with the Alkaline batteries particularly Duracell, not sure if the problem is with trying to increase capacity or cutting costs on the casings.    https://www.eevblog.com/forum/Smileys/default/confused0024.gif

[bilko]

I just found a 1998 GP Ultra AA cell, unused and it is still showing 1.56V, and NO LEAKS !!

[tom66]

Funny, because we just received a pack of Powerone (Varta) alkaline cells. Expiry 2014 (would normally be 2017~2020.) Capacity just 300mAh instead of 1200mAh at 1 amp discharge, but oddly 1600mAh compared to 2000mAh typical at 200mA discharge.  So different, very odd.

But one thing we have learnt is open circuit voltage is a poor indicator of battery capacity.

[ivan747]

Great study! The results are fascinating for me. I never though IKEA batteries had such a great performance. And Duracells aren't that great. I guess they waste all their margins in marketing rather than engineering