Repairing an unresponsive Kindle 3 (Kindle Keyboard)
Here’s the story of how I resurrected a Kindle 3 that wouldn’t turn on even after replacing its battery. TLDR: I reflowed cracked solder balls underneath the BGA CPU.
I have a Kindle 3 (or Kindle Keyboard) that I am quite fond of. While newer Kindle models are technically better on paper, I nevertheless enjoy using this “antique” model. Despite its 15 years of age, when it works, this Kindle remains a capable e-reader in 2025, albeit without the luxuries of a frontlight or access to Amazon’s online store. Unfortunately, my Kindle 3 had stopped working.
One common problem with older Kindle 3’s is that the Lithium Polymer batteries inside them age and degrade until they can no longer hold a charge. The internet abounds with testimonials where people claim that plugging in their Kindle for several days allowed them to magically resuscitate aged battery packs; this is not the solution when an old LiPo battery has degraded beyond the point of no return. The proper fix is to replace the original battery with a new part. This procedure is exactly what I did about three years ago. For much of that time, my fix worked quite well, until one day, the Kindle simply refused to turn on.
No matter how long I left the Kindle on the charger, and regardless of how long I pressed the reset switch, my Kindle remained unresponsive. Attaching a micro-USB charger, the amber LED would glow for a moment or two, but then it would turn off. The green LED (indicating the Kindle is booting) never activated.
Undeterred, I broke out my USB multimeter. Measuring the Kindle’s current draw revealed that it was pulling only 0.06A. This extremely low current told me that the Kindle was not charging its battery. (I expected to see a current draw somewhere in the ballpark of 0.5A to 0.8A.) Disassembling the Kindle, I extracted the three-year-old aftermarket battery pack. With my multimeter, I measured a grand total of 0V across the terminals of the battery, rather than the expected 3.7V. This proved that this battery was kaput!
Not really wanting to invest money into this ancient Kindle, I began the questionable job of disassembling the aftermarket battery pack to extract its BMS circuit board. My thinking was to swap this BMS onto a different LiPo pouch with similar characteristics to the original that I had lying around. Eventually, I came to my senses and stopped before doing something too dangerous. (It’s not wise to take LiPo batteries apart —-the risks posed by catching one of these on fire is not really worth the cost savings to me…) Anyway, unwrapping the dead aftermarket battery revealed something interesting: underneath the generic black label was an authentic-looking Amazon label from 2010! Whether or not this aftermarket part was a rebadged genuine Amazon part remains a mystery to me. This could be a genuine aftermarket part that was wrapped in an extra label to avoid getting stopped at customs for looking like a counterfeit. The outer label claimed the battery was made in 2020, but the inner label screamed 2010. If the 2010 manufacture date is correct, that would certainly help explain why the aftermarket part died after a relatively short period of time. The deception doesn’t stop with manufacture dates: the milliamp-hour ratings of the two labels also contradict one another. Who knows what the true capacity of the battery really was? Probably less than either of the labels claimed. The moral of this story: caveat emptor. Beware when shopping for aftermarket Kindle batteries —-someone somewhere probably isn’t being fully honest.
I safely discarded the guts of the autopsied battery pack and ordered a new one, crossing my fingers that this time, the replacement part would be genuinely new. Once my new-new battery arrived, I installed in into the Kindle ——and nothing happened. :(
Checking the internal voltages of the Kindle revealed that 3.9 volts were making it onto the logic board, which is 0.2V above the nominal voltage for the battery pack. No amount of time letting the Kindle sit around on a charger was going to get this to turn on if it wasn’t already doing so. Attaching my USB multimeter once again, I saw that very little current was being drawn —the same 0.06A from before replacing the battery. Interestingly, when I held down the reset switch for long enough, the current draw dropped further to 0.00A. My Kindle was very dead.
Searching the internet far and wide for answers, I waded through vast swamps of AI-generated slop and SEO nonsense until eventually, I reached a helpful forum post written by a real human. This post mentioned that sometimes the solder balls under the CPU can crack. A tiny grid of solder balls (BGA, or ball-grid-array) attach the CPU to the logic board, and if even one of these balls was cracked or otherwise damaged, it could explain the issues I was having. This Kindle was manufactured around 2010, at a time when manufacturing issues with lead-free solder balls were notoriously common (and not just with Kindles). At least two people claimed this was a known issue with this model of Kindle—- and their advice actually seemed consistent with the issues I was having —-power was getting to the CPU, but it just wasn’t doing anything.
The best solution would be to expertly remove the CPU, replace the old solder balls with new ones and solder it back down into place. However, at the moment, I lack the specialized tools required to perform SMD reballing and rework, so I couldn’t fix this Kindle the “right way”. I was determined to get this Kindle working —besides, I had just bought a brand-new battery for it. There was something else I could try however: I could blast the CPU with hot air until the existing solder balls underneath it melted, and then cross my fingers that this would restore the broken connections.
With little to lose, I disassembled my Kindle and extracted the logic board from the chassis. I pried up the RF shielding covering the CPU, and I began blasting it with hot air. More specifically, I began by preheating the logic board at a low temperature for a few minutes to prevent the PCB from warping, and then I targeted the CPU with even hotter air for about thirty seconds until the solder balls underneath it had a chance to melt. (The CPU is the largest chip underneath the largest RF shield.) I then gave the logic board a chance to cool down, and afterwards, I reassembled the Kindle.
To my excitement, the Kindle’s green status LED illuminated almost as soon as the battery was attached! Flipping the Kindle over to view the e-paper display, I watched as the screen came back to life, and the Kindle began to boot. My bricked Kindle was fixed!
I attached my USB multimeter, and saw that the Kindle was drawing a healthy 0.5-0.8A and beginning to charge up its new battery pack. I buttoned the Kindle back up, and began to enjoy it once more.
Once I mounted the Kindle to my computer, it began to draw 0.8A and the message about not charging went away.
