Massachusetts Institute of Technology
Media Lab
Matt Hirsch
Laptop Repair
Introduction
When I smelled the distinctive smell of electronics burning, I knew intellectually that something bad was happening. But nothing bad did happen, immediately. So I continued on, doing what I was doing, in the vain hope that the problem would just go away.
Earlier in the day, I had been very confused to find my laptop battery dead. I had left it plugged in, charging all morning, and now that I wanted to cut the umbilical cord and roam free I found myself powerless. But when I'd plugged the cord back in, all the right lights came on and the charge indicator started to creep up. I'd just run a software update, so maybe this was a bug of some kind?
But now, 12 hours later, I am sitting here in my darkened living room, trying to finish up a project, and the unmistakable stench of electronic death is wafting out of my laptop. I pull out the battery, remembering the incident earlier in the day, and this webcomic. And then I start running all my unison profiles.
A short time later, my out-of-warranty tx1000 zaps off, and will not power itself from the wall. I find that it will continue to run from the battery. So something in the DC-DC converter has failed.
The Cause
Never one to turn down a challenge, the next day I get down to business and open the thing up to see what happened. HP kindly makes their service manual available, which is quite helpful considering the myriad tiny screws scattered throughout the laptop.
Having fixed my share of electronics in the past, I'm expecting to see a blown electrolytic capacitor, along with whatever collateral damage occurred in the failure. So I was pretty surprised to see this:
Some IC has burned a hole in its case. I test all the capacitors in the area and they all seem fine. And sure enough, this chip is in what looks like a buck converter circuit next to the power input:
What is it?
So of course, if I can replace this $0.50 chip that'll be better than buying a new $400 motherboard. But first I have to figure out what it is. After unsuccessfully googling 4423, I wrongly assume that 4423 is some kind of lot or date code and that the true part number is burned off.
Thus begins a long, painful, and ultimately fruitless campaign to get HP to tell me what the part number of that chip is. Over the course of the next week I am misunderstood, willfully misconstrued, hung up on, lied to, and ultimately blown off, by all manner of tech support and customer service goons. I curse the names Hewlett and Packard. The profit motive is not on my side this time.
Finally, I decide to take my girlfriend's advice and just ask Reddit.
One hour later, I get an answer. D'oh.
The part is a p-channel MOSFET, the Alpha and Omega AO 4423. I'm now pretty sure that this is a buck supply.
Getting the part
Now my problem has become getting one of these. As far as I can tell on FindChips.com, the 4427 is only available in large quantity.
I request a sample of the 4423 from Alpha and Omega, but when I don't hear back in a day, I start looking for a substitute. It turns out that the AO 4447 is a reasonably close match, which can be bought in small quantities. I tack on a few to the next digikey order.
The Replacement
When the 4447 finally arrives, I'm eager to give it a shot. I measured the resistance across the burned out chip to be just a few Ω, so I guess whatever damage was going to be done to the board has already happened. There's little to risk in replacing the chip with my not-quite-perfect replacement.
The first step is to desolder the old 4427. The case is so brittle it just disintegrates when I put a little pressure on it.
The silicon wafer comes out of its packaging. I've never seen a chip do this before.
Another view of the remainder of the old chip.
I remove the old chip bit by bit. It's actually pretty easy, but I'm being careful not to scratch the motherboard and break some trace (I've done that before).
Parts of the 4427.
Finally it's fully removed.
New and shiny.
Putting some solder on the pads (tinning them) will make it easy to stick a surface mount part down to a board.
I place the chip and solder it down.
All set.
I nervously attach the battery and plug in the power cord.
Happily, it passes the smoke test (nothing explodes) and the battery charge light comes on!
Epilogue
Upon reassembling the laptop and trying it out, I am a little disappointed to find that my modified buck supply doesn't deliver as much current with the 4447. I can't both charge the battery from below about 80% and run the laptop from wall power at once. But being able to do them one at a time is better than nothing at all.
I have since heard back from Alpha and Omega, and they very graciously sent me a sample AO4427. When I have some time one of these weekends I'll swap out my 4447 for a new 4427. With any luck I'll have a fully functional laptop once again.
Thanks
Thanks very much to Reddit, and jarko_ for finding the part. And thanks to David Carr, who offered me good advice and analysis, as always.
And finally thanks HP, for teaching me patience... and not to buy HP products ;P