[Aug 7th] Investigation of battery capacity claims. EVO AMZER serious ripoff everyone

[d0ugie]

We all want better battery life, no question, more than processor speed (some of us even underclock), resolution, 3D drivers; pretty much everything takes second place in our hearts to battery life. In response to that demand there are a lot of companies out there selling batteries for our phones either at the same mAh rating of the OEM (the one that came with the phone), a higher mAh rating but the battery's the same dimensions as the OEM and does not require a deeper cover and batteries that are both larger in capacity and physical size -- or so they allege. Usually these batteries are cheaper than the OEM's, mAh for mAh at least. By the way, mAh ratings, or milliamp hours, think of it in terms of the size of a gas tank you installed in your car. The higher the figure, the longer you can use your phone.

Got a Nexus One? Say hello to our sister thread! No wait, just go straight to forum.batteryboss.org right now, forget this thread.

There are battery threads all over XDA but only with vague data so I invested in equipment to rate the batteries myself, under the guidance of electrogeniuses as you can read in this thread, in order to tell you what to expect in addition to giving you figures such as how many amp hours you get on the dollar as maybe a cheap battery with an exaggerated or oddly defined rating is worth getting over an expensive and better one if you're on a budget. Not counting all the man hours that Telek, amit77 and I have put into coming up with the most reliable way to produce data, so far I've spent over $200 on different equipment to get to where we are now. By the way, OEM = made by the people who made the battery that came with your phone. I'm not asking you to donate yet or reimburse me for my troubles, rather I'd like you, the next time you buy a battery I haven't already tested, to mail it to me first so I can run a couple tests after which I will FedEx it to you at my expense. If that interests you, PM me. If you are a battery manufacturer or reseller and you stand by your claims and trust me to run legitimate tests on it that are consistent with every other test I publish in terms of procedure, PM me or post publicly.

The procedure:
I am using the Computerized Battery Analyzer III. The software which is somewhat sophisticated plots out milliamp hours (mAh) burnt over the descent of voltage from 4.14V to 3.5V which we have determined to be the level of voltage in the battery at which point the phone decides to stop charging itself and decides it's time to turn itself off because it's too low on juice. The CBA software plots out data in graphs, PDFs, CSVs, the whole deal. Looks a little like this:

In order to produce the best real-world information, numbers that are relevant to you and not just to academics, I charge the batteries with the phone as you do until the phone decides the battery is charged which for the Rhodium/Touch Pro2/Tilt 2 is 4.14 volts regardless of capacity rating. I hook the batteries up to the CBA which is plugged into my computer. With the software that came with the CBA I have the CBA test the batteries at 250mA, a current in the neighborhood of what the average user would average were he to do his thing (including having push-mail fired up with the screen on bright, downloading and browsing rss feeds, the occasional call, bubblebreaker etc) without interruption. To get a better idea of what kind of current you're using when doing various things use the attached acbPowerMeter software. I'm not Geraldo Rivera out to get the third party guys that exaggerate their numbers a little bit nor am I here to rewrite Wikipedia's take on capacity calculation industry standards. If you're a manufacturer or a battery company sympathizer and want to break my balls about voltage cutoffs, read this simple explanation which I feel sums up our position well. The point, in short, of all of this is to supply you with information that will help you choose which battery to buy.

Doug Simmons

 

[d0ugie]

Test results and other information.


This table is a hotlinked image to data on the mother site of this, batteryboss.org on which the actual links work. Hit refresh if you've been here before in case your browser cached the image of the table.

Updates:
August 7th: First did a dry run then I the AMZER 1800mAh for the EVO. So far it's in first place for being the biggest ripoff on the gallon. First place.
August 5th: Received Carl's AMZER 1800mAh for EVO, doing a dry run discharge now, hopefully get some data for you tomorrow morning.
July 28th: Finished Carl's Seidio 3500mAh for EVO 4G. Seems the Seidios all rate at 81% of their claim.
July 6th: Completed round one of EVO stock (John Doward). Got the coveted Amzer 1800 and a Seidio 3500, both EVO, en route thanks to Carl Willi.
June 12th: Completed first run of a Mugen 3200mAh for the Hero. Both disappointing and unsurprising. Most cost ineffective battery I've tested.
June 11th: Jasper and Dan's batteries on the way back to them. Thanks again. Hey, Mugen 3200 for Hero and EVO 4G stock on their way! Another update, just received that Mugen 3200, charging now.
June 10th: Completed testing for a no name Hero battery and the stock Incredible battery with the EVO 4G stock on its way. Nice. Returning those batteries to my man Dan and my other man Jasper.
June 9th: Seems a seller I linked may be committing fraud. Please read.
June 7th: First test complete of an oversized no name Hero battery, second one in progress.
April 26th: Got some press. And a little more.
April 15th: Hoo-F'ing-ray, we have a winner, Wade's HTC 2150mAh clocks in exactly at 2150! high five, HTC!
April 14th: Taking the HTC 2150mAh for a spin right now. Finally! Thank you Deathmonkey!April 12th: Rotohammer's Seidio 1600mAh for the N1 has arrived, charging. Exact dimensions as OEM, wish I had a scale.
April 9th: In a continued effort to outdo himself Rotohammer just ordered a 2400mAh-rated Cameron Sino, on its way to me. Lucky I got his attention. Extremely helpful. Thanks.
April 8th: N1 Seidio 1600mAh should show up today, thanks to Rotohammer.
April 3rd: Finished N1 Seidio 3200mAh, five runs. Learned that it's rated slightly more honestly than Mugen (not saying much) but is the most expensive battery per tested amp hour. Still, highest capacity. I got a new and fast and really badass server and now have a our own forum which you can fire up at forum.batteryboss.org. Finished the new Andida for the TP2, pretty weak, but for some of you the price may be right.
March 30th: Completed dry run of a Seidio 3200mAh for the N1. Not fantastic but Seidio has taken the lead against Mugen in honesty.
March 29th: Mugen "engineer" responds (see table). Rotohammer's Seidio arrived, charging now baby, yeah! Should be very interesting. Thanks Roto.
March 27th: Shawn's OEM a fake (but a well performing fake). Activity building in the Nexus One thread. Mugen sent me an exchange for Jeremy's and it sucks even more.
March 20th: Just ran the first test of the Nexus One's OEM, not bad.
March 18th: Just ordered a Google Nexus One. I got an extra battery so the first thing I'm using this for is to prepare a battery for testing. Need to figure out if it has different voltage cutoffs, need to figure out how to present the data and what to do with my site, .. hmmm.
March 16th: Mugen wants me to send me another battery to test, I agreed and mailed them back Jeremy's battery. Also mailed Sean/Telek his OEM 1500mAh. Thank you both fellas. Also DeathmonkeyGTX offered to sponsor a test of the HTC 2150mAh -- thank you!
March 13th: Finished no name #2 3600mAh (2466mAh ). In search of voltage cutoffs for Touch Pro/Fuze, please help.
March 12th: Mugen has expressed interest in sending me another battery to test, I expressed willingness. And to you I express curiosity into which device to expand the testing.
March 8th:Finished round two of no name #2 and fake OEM #2. Waiting on another ebay OEM to verify authenticity and a fresh Andida courtesy of my main man Shawn Martell.
March 7th:Added intriguing head to head chart matches.
March 6th:Completed a few more including fresh standard legit OEM, also discovered two counterfeits.
March 2nd: Completed no name #1, cheapest per mAh so far. Dropped Jason's battery off in the mail as promised.
Feb 28th: Completed tests of the Seidio, mailing it to jasonweaver.
Feb 27th: Just received Seidio 1750mAh from jasonweaver in addition to 1500mAh no name ebay cheapo. Nice.
Feb 27th: Mugen 1800mAh testing completed, table updated. Thank you very much jcr916 who bought the battery and had it shipped to me, now I'm going to mail it to him.
Feb 22nd: Thank you jasonweaver and jcr916 who are hooking me up with a barely-used Seidio and a brand new Mugen 1800mAh respectively. Those test results should be interesting as from what I've gathered those two brands have the best reputation and are priced accordingly so let's see if they deserve it.
Telek and I just laid down some dough for five more batteries this weekend. So I'll have a lot of testing to do shortly. Stay tuned for the results!

Batteries I would like to test next so PM me if you want to help. New and used batteries welcome.
AT&T/HTC 2150mAh Pricey worth it. Really want to test this one. *En route*
HTC 2150mAh - Same model as AT&T but cheaper, doesn't come with door. *En route (same battery)*
Cameron Sino - Found five favorable/neutral reports.
A new Seidio 1750mAh - Tested a used one already but need data on a fresh battery.

Google Froogle search for more.

Tips and other reflections:
AT&T people, beware that if you order a battery that requires a deeper back cover but is for the Touch Pro2 and not the AT&T Tilt 2 that the cover may not have a hole for the PTT button nor may it latch on. Maybe you can burn one through with a hot screwdriver or you could just crazy glue. Beware of the usual dangers of ebay obviously, only use sellers with high ratings and consider buying straight from the company's website or Amazon or a name you've heard of. I have found over two counterfeits from sellers with high ratings. Read your phone's warranty before using a third party battery. Get the return policy before you buy. If you do get burnt with a counterfeit OEM, immediately give negative feedback using language like "counterfeit" and email the seller requesting both immediate restitution and that they remove their listings of that particular battery or at the least any reference in the listing to the battery being either an OEM or giving an OEM-like serial number, anything misleading, and in return offer to neutralize negative feedback. Reporting fraud to ebay is up to you but I would take those steps first. It's a longshot but the information we really want from the dealer is their supplier but so far I haven't been able to get any of them to cough that information up. If you're not sure if the thing's a counterfeit and want to find out, send it to me to test.

Testing hardware:
I am using the West Mountain Radio CBA III (Computerized Battery Analyzer) which you can buy along with some toys from these guys for $149. I bought something else from them, didn't like it and they offered to shave the cost of the thing I didn't want off the price of the CBA III without even asking me to return it. Good people. The CBA III is the most accurate and reliable device we could find for these testing purposes and we spent many hours arriving at the final testing procedure. No corners cut. There is no indication whatsoever that the results it's produced are inaccurate, certainly not relative to each other given its consistency. All testing procedures were identical including the current of 250mA, starting voltage and bottom cutoff (4.14V and 3.5V respectively, the top and bottom cutoffs of the Touch Pro2, which I use to charge the batteries with original HTC wall charger). The 250mA current may be a little high and won't produce as flattering a result versus a 100mA current, but it's both a normal current we burn when we're doing stuff on the phone, it keeps each of the three tests inside six hours usually and most importantly we use that current on every single test of every single battery so this is a standardized test. Finally the OEMs get 95% of their claimed rating on this current so we believe that that current is the sweet spot to supply you with information to use to buy your next battery.

Doug Simmons

 

[Telek]

Excellent thread - although from my experience with LiIon devices:

- the device itself typically has poor ability to determine battery usage
- you really need an actual external ammeter to monitor usage

With payment terminals that I used to work on the external ammeter (which I assume was accurate) was up to 15% off what the terminal itself reported - and wasn't consistent depending on the amount of current draw. The ability for the batteries themselves to accurately report their % remaning was also highly inaccurate and can only be used as a guideline. YMMV.

 

[d0ugie]

Thanks .. though you may realize this, my main interest is not in current at a given point but in total consumption from a full charge to when it dies. You're saying in order to get a definitive result, solid enough to recommend one thing over another, I need such a device to get beyond bad indications and confounding variables? Or with a difference between a few hundred mAhs over a few trials sound like enough to determine a clear winner, but ideally have an ammeter? Any cheap ammeters that not only clock the current but plot it down over time or do whatever it takes to yield the total? Not too familiar with these things, just a quick glance at wikipedia.

 

[Telek]

Well honestly I have no idea how either the battery or the TP2 fares for current reading.

I think you'd have to do a benchmark - find some way of running the same task on the TP2 that will use the same amount of battery power and run it a few times from a full battery charge. See if the mAh readings are consistent. My guess, however, is that the readings are going to be quite different which will be a combination of the battery AND reporting method. Even using an external battery reconditioner we'd get +/- 10% on successive runs with the same battery. The only way you could tell that the reconditioner was working was the batteries would go from 30 mins runtime to 90 mins after reconditioning.

Back when I did my testing I used a custom designed unit - had a arm7 microcontroller and threw a resistor in series and constantly monitored the voltage drop to determine current usage and plotted it that way. Most multimeters that you can hook up to your computer don't have high enough sampling rates or are too expensive.

 

[Telek]

how the frak do you delete messages? duplicate post...

 

[d0ugie]

how the frak do you delete messages? duplicate post...

dupe all you want, need this thing bumped.. working hard on it, getting some good info together.

 

[anit77]

As Telek said an external test is the only reliable way to get accurate results. The best way to do this to apply the same resistive drain to all of the batteries and charging them all with the same charger. A .5c drain should be safe enough to drain in a reasonable amount of time with out over heating the battery.

The main factor we need to know as far as the device is concerned is what the cutoff voltage is set at. LI-ion batteries can be drained down to less the 3v safely and most good ones have protection circuits that cut them off above 2.7v. None of that will matter if HTC has the device set to cutoff at 3.2v or higher.

Once the cutoff V is known then you time the drain from full charge to C/O.

I've done a lot of battery pack building and maintaining in radio controlled hobby's. With out having access to the right equipment this is going to be a daunting task and quite possibly more expensive than buying a few batteries. Hopefully the right person will see this thread and will step up to do the testing.

 

[d0ugie]

As Telek said an external test is the only reliable way to get accurate results. The best way to do this to apply the same resistive drain to all of the batteries and charging them all with the same charger. A .5c drain should be safe enough to drain in a reasonable amount of time with out over heating the battery.

The main factor we need to know as far as the device is concerned is what the cutoff voltage is set at. LI-ion batteries can be drained down to less the 3v safely and most good ones have protection circuits that cut them off above 2.7v. None of that will matter if HTC has the device set to cutoff at 3.2v or higher.

Once the cutoff V is known then you time the drain from full charge to C/O.

I've done a lot of battery pack building and maintaining in radio controlled hobby's. With out having access to the right equipment this is going to be a daunting task and quite possibly more expensive than buying a few batteries. Hopefully the right person will see this thread and will step up to do the testing.

Okay well in the interests of making this thread epically informative I am almost willing to buy myself an ammeter. Could you please recommend one that is both cheap and capable of getting the job done for something smaller than a car battery and a device with such settings to fine-tune? Something that I can rig up to the battery's contacts which are so thinly spaced together without a soldering iron? Here's a list...

 

[Telek]

Well for starters do you want to test the phone + battery or just the battery?

Testing the battery alone will be a LOT easier - in fact a general test can be done with just a heavy duty resistor and a voltmeter. Watch how long it takes the battery to go from full charge to C/O and you've got your mAh rating. I'd recommend this as it's easy and cheap.

If you want to check actual phone usage time and compare it to whatever reading you get via software you'll need to build a little rig that allows you to place the battery externally with leads running to the battery contacts in the phone. That part is not as hard as it sounds. However what that you're going to need some sort of way of tracking the current flow with high precision and high frequency. If you're any good with microcontrollers that can be done easily, otherwise I'm not sure. I'll check and see if I can find anything that would qualify.

 

[d0ugie]

Well for starters do you want to test the phone + battery or just the battery? Testing the battery alone will be a LOT easier - in fact a general test can be done with just a heavy duty resistor and a voltmeter. Watch how long it takes the battery to go from full charge to C/O and you've got your mAh rating. I'd recommend this as it's easy and cheap.

If you want to check actual phone usage time and compare it to whatever reading you get via software you'll need to build a little rig that allows you to place the battery externally with leads running to the battery contacts in the phone. That part is not as hard as it sounds. However what that you're going to need some sort of way of tracking the current flow with high precision and high frequency. If you're any good with microcontrollers that can be done easily, otherwise I'm not sure. I'll check and see if I can find anything that would qualify.

C'mon man, obviously I want the easier way that satisfactorily satisfies what I want to know without crazy wiring and microcontrollers. Just want to know how many milliamp hours these two suckers and any other batteries I might get along the way compare to each other, a drag race. I'll leave that other stuff to the guys from Popular Mechanics.

 

[Telek]

We also need to consider the resolution needed for such a device.

For example a 5.0V ADC (providing 0-5V range) at 16-bit resolution for a current shunt designed to provide a 50mV drop at 1A is going to give you 1.53mA resolution (5V 16-bit = 0.0763mV resolution; 0-50mV range = 655 binary range = 1.53mA range). Conveniently a 3.3V range for the ADC gives almost exactly 1mA resolution which is perfect.

However I can't find any cheap ready-to-go data loggers with that resolution and range.

Now if you wanted to make the setup a little more complicated and toss a simple op-amp into there you could probably get away with something like this:

http://www.dataq.com/products/startkit/di194rs.htm

$25 plus a few bucks to build a simple op-amp circuit (check out http://www.chem.uoa.gr/Applets/AppletOpAmps/Appl_OpAmps2.html with values 1kO and 2000kO V1=0.005V and V2=0) use the computer PSU to provide +/- 12V to the amp will give you ~ +/- 10V operational range (perfect for the data logger) and you'll get 2mA resolution with 240Hz sampling - far from perfect but sufficient to actually do some decent measurements of actual current draw from the phone. Log both the shunt resistor voltage drop through the op-amp and the raw battery voltage and you've got a nice little power monitor that can give you detailed measurements of how much power every little thing in your phone takes.

 

[Telek]

Ok so grab yourself one of these:

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=HPC1210JCT-ND

Or any generic 10 ohm resistor capable of dissipating at least 5W (wire-wound will work). Ask at your local electronics shop and they can probably help you. Should cost less than $20 including the multimeter.

It'll draw about 700mA and dissipate about 5W (so it'll get hot, should put a small fan on it) and should dissipate your battery in about 2 hours. Combine with a voltmeter where you write the measurements down at least every couple of minutes and that will allow you to plot the voltage drop curve and calculate the mAh rating of the battery.

 

[d0ugie]

Well if you can come up with what you think the right standardized procedures and various levels of this and that, again standardized, to use to test multiple batteries from multiple manufacturers of varying mAh claims (but presumably of the same voltage?) just to get a total mAh versus that of the OEM, and then tell me which device to buy to make that happen, I could contact these various companies, tell them to send me their superbattery for me to test accurately to see if they are legit and, in exchange for sending me the battery, I'll list the results on either of my sites which are fairly trafficked, enough for a company to want to give away a battery if it meant getting their claims verified in front of an audience thirsty for milliamp hours.

You see where I'm going with this?

 

[d0ugie]

Ok so grab yourself one of these:

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=HPC1210JCT-ND

Or any generic 10 ohm resistor capable of dissipating at least 5W (wire-wound will work). Ask at your local electronics shop and they can probably help you. Should cost less than $20 including the multimeter.

It'll draw about 700mA and dissipate about 5W (so it'll get hot, should put a small fan on it) and should dissipate your battery in about 2 hours. Combine with a voltmeter where you write the measurements down at least every couple of minutes and that will allow you to plot the voltage drop curve and calculate the mAh rating of the battery.

I've got a Radio Shack down the block. But 700mA is maybe unnecessarily/excessively strong -- firstly because running batteries hot and heavy wears them out over time so if I keep testing new batteries against one oem .. ahh nevermind i guess that wouldn't batter but normal usage is between 30-400mA. How many ohms might yield roughly 250, 300mA? Think I'd get lucky at RatShack?

 

[Telek]

Running batteries at 1/2 their capacity is generally pretty safe for full discharge - so a 1500mAh battery discharging at 700mA should be fine - not really considered "hot & heavy".

You could go 20 ohm (2.5W discharge 350mA) and monitor over 4-5 hours then, just means more work for you since you're going to need to record the voltage over time.

I somehow doubt that RadioShack would carry any sort of resistors like this, but a general purpose electronics hobby shop should.

 

[woaloo]

Thanks you very much with this test. Very instructive.

A question:
What is the best PocketPC software for mesuring mAh. I've try AbcPower, but I don't think it's very accurate when charging. normally, 500mA via PC USB.

 

[d0ugie]

powerguard

I believe it's powerguard which is frickin' awesome for all things battery-related (.. born on xda) but I can't get the consumption thing to tally which doesn't really help this project. That's either because the software doesn't like our phone or because the shareware is crippled not to do that among some other things. Other than this powerguard and abcpowermeter, I am unaware of any other WinMo app that will count up the consumption. So I sent the guy some euros (five, he asked for three) to get a registration file hoping that will do the trick but man he went out of his way to make it a pain in the balls to register this thing. Hope he checks his email soon because I want this thing unleashed.

By the way for those of you living in 10% increment land on your battery meters, grab Chainfire's 1% battery driver for Euro and apparently AT&T Touch Pro2s (gsm?) or Quentin-'s 1% driver for Tmo, Sprint and Verizon (I guess any cdma tp2). After you install either of those you'd better check your battery drain to make sure you didn't install the wrong one (you can tell if you notice your phone is using a lot more juice). If you did, not to worry, they are uninstallable.

Doug

 

[d0ugie]

Hopefully the right person will see this thread and will step up to do the testing.

I want to be that guy. I'm willing to put the time, money and patience into it. Only snag is is that I'm not a physicist/electrical engineer like you two.

What I want to set out to do here is have a means of performing standardized testing with the same equipment with the same settings with me doing the same things (including how frequently I watch it and what math I do at the end) each time in order to get true and consistently true mAh until c/o ratings, true enough to satisfy third party battery companies/dealers (and people who actually read and use what I post) that I've got a fair and trustworthy procedure to see if their batteries live up to their claims, posting tables of batteries, mAh, cost of the battery and cost per mAh, something like that, both here and on my website for people to read and copy to their own sites and forum threads.

I reread your comments a couple more times. What I guess I need to do is find out what the cutoff voltage is for this phone (do I do that by eyeballing the mV readings on this powerguard software until the phone clicks off or is additional equipment necessary?). Then I need to get myself this data logger -- nope, just a 10 ohm resistor and maybe this voltmeter or perhaps something like this would make this simpler? What's my shopping cart?

Once I have some setup like that I fully discharge a battery then charge it up with the phone off then connect it both to the resistor and the voltmeter, plot out the readings every three minutes until it hits the voltage cutoff point, repeat X amount of times, then crank out some math to figure out a total mAh reading from start to cutoff point (we'll discuss that later...) and finally I'm good? Something like that? How am I doing?

Just hit me with the shopping cart and I'll get that out of the way and read this while it all ships then check back here.

The goal is to generate and disseminate valuable information on something very mysterious and dubious. People who make these things know we have no idea how to do what we're fixin' to do in this thread so why not claim wild ratings for the same dimensions and get more people to buy our stuff? Let's tell 'em why and maybe score some free batteries along the way for our troubles.

 

[anit77]

This looks like it'd be perfect. but it's $120. You be able to test most any battery too. There's tons of this type stuff in R/C and electronics hobby sites. It just doing the searching needed to find what your looking for.

I'd say make a new thread for taking donations to buy one. PM a mod or Flar and ask where to post it or if they'll put links to the thread in some of the sub forums and you'd get people with other phone that'd be interested too. Could be a way to make some extra cash as well. If you don't want to deal with a progect of this size I'm sure you'd be able to find someone who is.