Bump Charge fully explained - we may have to deal with it

[Medion]

Post 1: Explanation and Summary

Post 2: Expands on the bullet points from post 1

Post 3: Further investigation needed

My last post on this was an analysis. I had some theories but clearly stated that they were theory, and not fact, based on numerous tests. Well, some parts of my theory were correct, and others turned out to be false. After some more tests and some help from someone with the patience to sift through two kernals, we've figured it out. I didn't want to publish these results without checking the NOV OTA for changes, as these findings were finished around the time the NOV OTA started rolling out. So, here's what's going on (See Post 1 for conclusions, and Post 2 for elaboration on each point).

1. The Droid Incredible does not support power via AC.

2. Due to point one, even when plugged in, the Incredible runs off the battery.

3. The kernel used in the Incredible sets two battery charge thresholds, 96% and 80%.

4. The 96% threshold is set to prevent constant recharging (which over time drains the life of the battery). If the phone reports 96% or more charge when plugged in, you'll get a green light.

5. The 80% threshold is the point at which the Incredible will begin charging again.

6. The above is controlled via software. If you turn the phone off, it will act as a typical dumb-charger and do a complete charge.

 

[Medion]

This post is meant to further elaborate on the bullet points above. The first post is meant as a summary, whereas this post is the meat and potatoes for those who want a complete explanation.

1. The Droid Incredible is not capable of running off AC. For one, AC is disabled during a bootup, so the battery would have to be inserted and have some degree of charge just to get through this. Secondly, removing the battery when powered on will cause the Incredible to turn off. NOTE: I have now had the opportunity to test this on 3 Incredibles (Two AMOLED, one SLCD). If you wish to dispute this, I'm open to two methods. One method is to ship me the phone that you believe runs off AC (I will pay shipping both ways). Alternatively, you can record a video. What I want to see is the phone plugged in, back cover off and battery exposed, phone turned on. Then, with the phone plugged in, remove the battery. If the phone stays powered on and functional, you've just proven me wrong. At that point, we'll need to talk, as I'll want to test this to see what hardware/software differences are causing this, and if it can be replicated to remove the need for a bump charge (bump charge is caused by the current kernel, but a custom kernel could be modified to work with this kind of phone for better battery life).

Furthermore, it is currently inconclusive if the lack of AC functionality is a hardware or software limitation. It could be hardware limited (where the battery completes the circuit), or it could be software (security measure, like the original Sony PSP which would not work without a battery inserted, unless running a hacked ROM which removed that security). If software limited, we can write a custom kernel or app to remove this issue, and thus eliminated the need for bump charge. If hardware limited, we're screwed.

2. Due to point 1, the Incredible, when plugged in, continues to use the battery. If charge > usage, then the battery slowly charges. If charge < usage, then the battery slowly drains even when plugged in. After the first 2.2 OTA, charging was drastically slowed to reduce battery heat (caused by a battery being charged and drained simultaneously, brilliant HTC!). This caused certain very high drain apps (games, Google Nav) to drain the battery faster than it could be charged. The NOV OTA fixes this by speeding up charging, though I'm not yet certain if it's back to 2.1 levels.

3. The two thresholds will be better explained in the following bullets.

4. Lithium-Ion battery can, in theory, only go through so many complete charge cycles. However, charging from 20-80% causes less wear than charging from 80-100%. As such, constantly charging that last 20% can quickly reduce the effective life of the battery. Doing full drains and charges are also bad. So, since the battery is being drained while being used, HTC didn't want the charger to constantly flicker on and off. Once the battery is fully charged, the charger turns off, the light turns green, and the battery begins to drain and not charge, even while plugged in. Despite being under 100%, the light stays green. If you unplug it, and find that you have at least 96% left, then plugging it in will STILL give you a green light and no charge (but turning off the phone and using it as a dumb charger will give you an orange light and cause it to begin charging to full again). This is the firs bump-charge issue.

5. The second threshold is 80%. This is where HTC decides that it's time to charge again. If your phone stays plugged in, in theory, it could drain as far down as 81% and still report a green light! However, once it hits 80%, it will begin to charge again, and the light will turn orange. This means that if you went to bed with the phone charging, you could wake up with a charge anywhere between 81-99% and still have a green light. If you try to bump charge with the phone on, and the light stays green, this means you have a 96% or better charge. You'll need to power down to bump. If you get an orange light, this meant you were in the 81-95% range, and powering down won't be necessary to bump.

6. If you turn the phone off to use as a dumb terminal, the software is bypassed, and the battery will charge to full. However, LiIon does lose a residual charge over time (typically 1% per day). So, if you unplug the charger and plug it back in, yes, the light may change orange. Also, battery charge against a type of resistance, meaning the first 50% charges faster than the last 50% (hence why some devices report a charge time of 1 hour for 80%, 3 hours for full). So, that last 1% could take 5-10 minutes to turn green. Really, it's just one percent, so don't assume that a 10 minute bump means you lost 10% charge. You didn't (if the phone was off).

 

[Medion]

So, where do we go from here? We need to determine if the lack of power via AC is a hardware or software implementation. If software, we have to find where this is at (it won't be in HTC's open-sourced Kernel if it's a security check). If hardware, then we're limited to a hardware mod that most wouldn't want to do.

In the previous thread, two users claimed for certain that the Incredible they each had would work off AC. If this is true, and since all Incredibles run off the same software, that would mean we're looking at different hardware models. If this is the case, a custom kernel can be tweaked to not use those thresholds on THOSE phones, to eliminate the need to bump charge (and thus, extend the life of the battery used). People with a non-AC powered phone, like me, would be left in the dark.

But, if it's a software limitation (like the aforementioned Sony PSP example, post 2), we should then be able to isolate it. But I'm starting to think this is unlikely. I see no need for HTC to use a security measure on the Droid Incredible that is based on the battery being inserted. On the PSP, it made sense, as the battery itself was used to trigger the service menu and bootloader (hence the Pandora's Battery hack). We have no such battery-driven menus on the Incredible.

But, if it is a security measure, that can eventually be broken, and then custom kernels could be edited to get rid of the 96/80% thresholds, and thus, extend the life of our batteries.

 

[jbuck1984]

Good work. I like control measures on the batteries that seem to be used to make sure that the batteries don't degrade faster. Personally I'm ok with the bump, and with the Kernal I have with virtuous 3.1 I no longer have a battery life problem.

 

[howarmat]

great thread and evaluation. thanks!

 

[jes0212]

It all makes sense now... Thank you.

 

[RoboCuz]

I have started two threads (one was a poll) about phones running without a battery. In fact, I had a replacement of my first phone sent to be from Verizon, as the tech on the other end confirmed that his THREE phones that he had to test with all ran without a battery. We walked through the scenario together, and confirmed the different behavior. There definitely seem to be some phones that run off AC. I wish mine was one, as I suspect they also may exhibit better battery life in general (my original thought was they have less draw from the system, and therefore in general just use less juice).

PLEASE continue this, I hope someone can provide a phone or video to you that shows this behavior. I wish I could do more, but of course, mine runs only with battery.

For reference, here are the two relevant threads:
Real issue with power consumption, easily testable
Battery Problems - Interesting Test


Cheers
Rich

 

[sdrawkcab25]

I have started two threads (one was a poll) about phones running without a battery. In fact, I had a replacement of my first phone sent to be from Verizon, as the tech on the other end confirmed that his THREE phones that he had to test with all ran without a battery. We walked through the scenario together, and confirmed the different behavior. There definitely seem to be some phones that run off AC. I wish mine was one, as I suspect they also may exhibit better battery life in general (my original thought was they have less draw from the system, and therefore in general just use less juice).

PLEASE continue this, I hope someone can provide a phone or video to you that shows this behavior. I wish I could do more, but of course, mine runs only with battery.

For reference, here are the two relevant threads:
Real issue with power consumption, easily testable
Battery Problems - Interesting Test


Cheers
Rich

The first 2 incredibles I had both ran without a battery, but the 3rd refurb they just sent me does not run without a battery and the battery doesn't seem to be that different, but I'm also running a different ROM than I was on the other ones that is better on battery, so I'm sure if I went back to stock I'd have worse battery life.

PS: Nice write up Medion, nothing to disagree with you on this one, except that there is definitely Incredibles out there that run without a battery.

 

[Medion]

I don't doubt one bit that there are phones that can do this. I just cannot test it myself without seeing it.

If a phone has the ability to power on, run, and power down all without a battery, then there is hope for two things.

1. Isolate the difference and see if this is correctable on phones that can't do this.

2. Modify one section of the kernel to remove bump charging. However, should this kernel be made available, those who have non-AC powered phones should avoid using it, as doing so would lead to rapid battery degradation.

There are other factors, I'm being simplistic. But for now, I'm hoping to get a phone that I can do some tests on. Sure, a video would confirm the existence, but I would literally need to have my paws on it to further examine this.

 

[Medion]

sdrawkcab just brought up a valid point.

For those who have or had a phone that would run off AC, what ROM(s) did you run? For the AMOLED Incredible I own, I've run stock 2.1, stock 2.2, stock 2.2 (NOV OTA), and now Virtuous (first custom ROM). The other two phones I toyed with (AMOLED, SLCD) were running stock 2.2 (late AUG OTA). If this is a custom ROM thing, then we're looking at something that could change via software. In my case, Virtuous is very much stock, and I would assume that all ROMs using the Incredible's framework would exhibit this issue.

When I get home, I think I'll have to try Frankenrom and Cyanogen just to see if they affect this issue at all.

 

[sdrawkcab25]

sdrawkcab just brought up a valid point.

For those who have or had a phone that would run off AC, what ROM(s) did you run? For the AMOLED Incredible I own, I've run stock 2.1, stock 2.2, stock 2.2 (NOV OTA), and now Virtuous (first custom ROM). The other two phones I toyed with (AMOLED, SLCD) were running stock 2.2 (late AUG OTA). If this is a custom ROM thing, then we're looking at something that could change via software. In my case, Virtuous is very much stock, and I would assume that all ROMs using the Incredible's framework would exhibit this issue.

When I get home, I think I'll have to try Frankenrom and Cyanogen just to see if they affect this issue at all.

I'm running the MIUI rom now(invisiblek's #16 kernel) on the phone that doesn't operate without a battery, didn't try pulling the battery when stock, but the Incredibles that ran without a battery I was running Virtuous and Skyraider and completely stock(Froyo and Eclair)

 

[HoochieCoo]

Mine is a new Inc (purchased about 1 month ago) with SLCD. Running Sky Raider 3.5 and Ziggys latest kernel.

Mine does not run without the battery.

But...

While I see a sharp drop after unplugging from charging all night (usually from 100% to 94-95 within minutes), I don't see that drop if I just charge to 100% and immediately unplug.

I rarely leave it plugged in over night. If I set my phone down at night with it charged to 100% (not plugged in), it's common for me to see 96+ when I wake up in the morning. This was true on the stock ROM too. In fact, the other day, I was playing with KingxKlick #6, and I woke up in the morning (6 hours later) and it was sitting at 98%

 

[The_Law2]

I have an early AMOLED Inc running Cyanogen and the phone does not work on AC with no battery.

 

[Soccerdude588]

When you say the phone can, when removed from charger be anywhere from 81-96%, does that mean the phone actually says that, or it labels it as 100%, but only has the life of an 80 something percent battery?
I charge my phone every night, overnight. Also, while the phone is plugged in and sitting at 100% (right now), I can't for the life of my get it to drop below 100%. I've turned on BT, GPS, Navigation, 3g, Wi-fi tether, and played some NFS Shift. It just sits at 100%.

 

[jamor]

Well thanks for refreshing the noobies, Medion, but mostly everyone already knows the bump charge was done for a reason as a precautionary measure and that HTC had no plans on 'fixing' it.

 

[Medion]

When you say the phone can, when removed from charger be anywhere from 81-96%, does that mean the phone actually says that, or it labels it as 100%, but only has the life of an 80 something percent battery?
I charge my phone every night, overnight. Also, while the phone is plugged in and sitting at 100% (right now), I can't for the life of my get it to drop below 100%. I've turned on BT, GPS, Navigation, 3g, Wi-fi tether, and played some NFS Shift. It just sits at 100%.

The phone will report a higher charge (maybe 99/100%), but over the course of about 15-20 minutes it will tick down to the actual charge. Someone else here demonstrated that putting the battery in a compatible phone will report the actual charge immediately.

Well thanks for refreshing the noobies, Medion, but mostly everyone already knows the bump charge was done for a reason as a precautionary measure and that HTC had no plans on 'fixing' it.

Was that really necessary? If you don't find the information useful, then don't participate. We were having some heavy traffic in the older thread, and I had finished my findings. And while many of us figured that there was intentional battery management going on, the actual numbers weren't known.

 

[jamor]

Was that really necessary? If you don't find the information useful, then don't participate. We were having some heavy traffic in the older thread, and I had finished my findings.

Unfortunately, it was necessary as it's old news.. but thanks again for refreshing the new guys.

 

[Medion]

Unfortunately, it was necessary as it's old news.. but thanks again for refreshing the new guys.

Hey, thanks for enlightening me. Please link the thread that discussed all of this (including the specific numbers) so that they can be merged.

 

[bjanow]

Medion, your work is thorough but no need to tell someone not to participate. It's an open forum and they can say what they want as long as it stays within the rules. Also, only mods or guides can merge threads.

As for bump charging I have a simple theory. It's a design and safety feature. When a phone is on, the hardware correctly assumes you are using it. If you remember back a few years when there was no safety cut off on Dell laptops (or was it the iPhone.. whatever you get the idea) the battery exploded on someone's lap and the phone caught on fire in some poor guys front pocket. (that hurts my balls just thinking about it) So the lawyers got together and said, "dudes.. cut this charging thing off at X% so we have a leg to stand on when we go to court." And there you have it.

So IMHO, you can test all you like, pull batteries, put them back in, and even change kernels. You will always need to turn the phone off to get a full charge. And that is across the board, on my Eris and my wife's LG flip phone from 3 years ago.

 

[imdickie]

Unfortunately, it was necessary as it's old news.. but thanks again for refreshing the new guys.

Not everyone who is on this board has been on it forever. To make a statement like this is elitist and completely out of line. Medion took time to do a deep analysis of something that affects many of us, myself included. It may have been discussed at length previously but Medion's coverage of the topic was clear, concise, and most importantly all in one spot and current. Beyond that, he offers potential solutions. All you offered was criticism of some one trying hard to help the community.

 

[jamor]

Medion, your work is thorough but no need to tell someone not to participate. It's an open forum and they can say what they want as long as it stays within the rules. Also, only mods or guides can merge threads.

As for bump charging I have a simple theory. It's a design and safety feature. When a phone is on, the hardware correctly assumes you are using it. If you remember back a few years when there was no safety cut off on Dell laptops (or was it the iPhone.. whatever you get the idea) the battery exploded on someone's lap and the phone caught on fire in some poor guys front pocket. (that hurts my balls just thinking about it) So the lawyers got together and said, "dudes.. cut this charging thing off at X% so we have a leg to stand on when we go to court." And there you have it.

So IMHO, you can test all you like, pull batteries, put them back in, and even change kernels. You will always need to turn the phone off to get a full charge. And that is across the board, on my Eris and my wife's LG flip phone from 3 years ago.

True. I have to agree.. somewhere along the lines of not wanting to reach a certain voltage which is bad for the battery and like you said regarding the safety cut.

 

[Medion]

Medion, your work is thorough but no need to tell someone not to participate. It's an open forum and they can say what they want as long as it stays within the rules. Also, only mods or guides can merge threads.

He didn't participate, he trolled. That was my issue. He added nothing to the thread. My request for a link was facetious, because this hasn't been discussed and clarified to this point before, but had he provided the link (if it were to exist), we could request a mod to merge it. Without the link, he really is just trolling.

As for bump charging I have a simple theory. It's a design and safety feature. When a phone is on, the hardware correctly assumes you are using it. If you remember back a few years when there was no safety cut off on Dell laptops (or was it the iPhone.. whatever you get the idea) the battery exploded on someone's lap and the phone caught on fire in some poor guys front pocket. (that hurts my balls just thinking about it) So the lawyers got together and said, "dudes.. cut this charging thing off at X% so we have a leg to stand on when we go to court." And there you have it.

So IMHO, you can test all you like, pull batteries, put them back in, and even change kernels. You will always need to turn the phone off to get a full charge. And that is across the board, on my Eris and my wife's LG flip phone from 3 years ago.

Actually, what the Incredible (when turned on) is doing is something different. All Lithium-ion (or polymer, or other types of Lithium batteries) have a hardware cutoff built into them. When you charge a battery to 100%, it's not really at 100%. Here's an example with fudged numbers.

A battery is advertised as 1300mha. It may actually have 1500mha capacity. When charged to 1400, it will report a 100% charge and cease charging. When it discharges to 100, it will report 0% and cease operating until charged. This is both for safety (as you described), and to preserve the life of these batteries. Batteries in cell phones are intended to be usable for 2-3 years. Batteries in hybrid cars use much more aggressive management as they are meant to last 10-20 years.

Where the Incredible is different is that it's actually still using the battery when plugged in. Most consumer electronics, when plugged in, do a handoff from the battery to the AC. AC is split between operating the device and charging the battery. When the battery is fully charged, it ceases charging, and the device continues to run off AC.

In the Incredible, this handoff never occurs. AC continues to charge the battery as the device continues to drain the battery. This "in and out" occurring simultaneously is not good for the long term health of the battery, which is why HTC used this more aggressive method of battery life mangement. It's why this device has such an obvious need for a bump charge when most consumer devices do not.

 

[imdickie]

Medion, your work is thorough but no need to tell someone not to participate. It's an open forum and they can say what they want as long as it stays within the rules. Also, only mods or guides can merge threads.

As for bump charging I have a simple theory. It's a design and safety feature. When a phone is on, the hardware correctly assumes you are using it. If you remember back a few years when there was no safety cut off on Dell laptops (or was it the iPhone.. whatever you get the idea) the battery exploded on someone's lap and the phone caught on fire in some poor guys front pocket. (that hurts my balls just thinking about it) So the lawyers got together and said, "dudes.. cut this charging thing off at X% so we have a leg to stand on when we go to court." And there you have it.

So IMHO, you can test all you like, pull batteries, put them back in, and even change kernels. You will always need to turn the phone off to get a full charge. And that is across the board, on my Eris and my wife's LG flip phone from 3 years ago.

You make an extremely valid point.

I work for a large computer OEM and was in the middle of that battery debacle that happened several years ago. By forcing a battery to limit max charge to 96% and keep it from starting a charge until you get below 50% you effectively triple the life of your battery. In notebook computers that have 3 year batteries all they do is add a circuit to the battery that hard codes the thresholds.

I have gotten around the battery life issue by having multiple batteries and using a battery charger not the phone to charge the batteries. I just wake up in the AM an swap out my battery and take a freshly charged one with me in case I need it. My biggest concern and the reason I read this thread originally is how to get my phone to operate using the power adapter instead of the battery.

I am on the road much of the time and use my phone for everything. I use a bt headset (I'm in CA where you have to), I use a bt A2DP to stream my music library via Subsonic into my car stereo, and I use Google Nav to get me to my appointments. Because the dinc only uses battery power and I am a heavy user my phone battery discharges faster than the power adapter can recharge it. It also makes the battery extremely hot.

I'm hoping that Medion's analysis leads us to a way to bypass the "Use battery power only" on some phones, including mine. It will make for a much more functional device for me at least.

 

[rman18]

My Dinc with SkyRaider 3.5 will not run without the battery.

 

[dibs ODDJOB]

Great posts, good to see some analytical discussion of battery life as opposed to the usual "My battery goes for 12 days, you are doing something wrong!"

Do you know how any of this relates to Hibernation Mode, found on some newer ROMs brought over from the Desire HD? I'm running Skyraider 3.5, when I enter Hibernate mode, the LED light is not on, but I've read that the battery is still charging. I haven't tested it yet to confirm.

My hope would be that when it Hibernation mode, the battery charges as if the phone were off. But since it is actually still on, an app could potentially schedule a time to put the phone in Hibernate for a period of time, let the battery reach 100%, then exit Hibernate mode, just as I am ready to wake up and unplug my phone. It would essentially be an automated "bump charging app."

It's similar to turning a Blackberry on and off. BBs are never really off, just in a low power hibernate mode.

Thoughts?