Acer Extensa 4620-4605 laptop bad cap

[severach]

U10: Intersil ISL6255HRZ Integrated Battery Charger with Automatic Power Source Selector for Notebook Computers
U2, U1: Fairchild FDS4435BZ P-Channel MOSFET (adapter power on circuit)
U42: Fairchild FDS6298 N-Channel MOSFET (CPU Vcore-Hi)
U39: Fairchild FDS6676AS N-Channel MOSFET (CPU Vcore-Low)
C437: SMD Ceramic Capacitor, at least 25v

The circuit of U2 and U1 isn't quite like what is shown in the ISL6255 datasheet sample circuit. It's more like the circuit in the TI bq24751 example except that the drain and source are reversed.

Adapter -> S-U2-D -> D-U1-S -> SYSTEM
SYSTEM -> D-U42-S -> Vcore -> D-U39-S -> ground
U42-D -> C437 -> U39-S (SYSTEM to ground for U42+U39)

SYSTEM is 19v running from adapter or 11v running from the battery.

I got this laptop in for service. It forces any power supply into short circuit protection. A short should be easy enough to find so I set to work. Being my first laptop motherboard more involved than a power jack I didn't know what to look for. I figured if I pulled enough parts I should find the problem.

Power adapters just short and shut down so I attach the bench power supply to the power jack and turn the voltage up. At about 1.1v 1 amp the first MOSFET U2 in line heats up. It had tested as not shorted but it's heating up so out it comes. Then I place the bench power supply at the circuit point orphaned by U2 and U1 heats up. On and on this goes. Each MOSFET in line U2, U1, U39, and U42 heat up and out they come. Each tests good in the DCA55 except for U42 which was destroyed during removal. The short is still there and this time C437 heats up. Out comes C437 and the short is gone.

Now it all makes sense. U42 and U39 probably did not heat up since the current for C437 does not go through them. C437 is directly between them and I was expecting the MOSFETs to get hot, not the cap. U2 and U1 got hot because they were functioning properly and trying to supply a short.

I got a replacement FDS6298 for U42 and pulled a ceramic cap off from a similar location on a bad board and the laptop is working. One more electronic gadget saved from the dust bin.

It's a bad cap but finding it quite different from finding bad electrolytics.

[momaka]

Re: Acer Extensa 4620-4605 laptop bad cap

Nice work!

Yes, apparently those little multi-layer ceramic caps like to short out rather than go open circuit or high ESR like electrolytics do. Worst of all, they don't look bad like you mentioned.
I myself was able to save 2 dead monitors for the same reason - shorted ceramic caps in the panel board. Took me quite a bit of time to find the problem.

[TheLaptopDoctor]

Re: Acer Extensa 4620-4605 laptop bad cap

Good Find Severach.

A shortcut I use to find parts that are over heating is to get a simple Crayola crayon and stick it on the part. The crayon melts at a low temp but hot enough to find the short. Another way is an ohm meter in Resistor mode or Diode, will show very low resistance 0-10 ohms when caps should be ~500 ohms one way and infinity the other if it is a bypass cap on a MOSFET circuit.

[severach]

Re: Gateway MA3 laptop bad cap

The cap on the Acer wouldn't ever overheat because a standard laptop power supply wouldn't run long enough to get it to heat up. A laptop supply has only one output voltage so can only be safely hooked to a small number of locations on the board. The bench supply delivers whatever I tell it to so I can safely run any rail on the board including ones that might not be shorted. If I'm wondering whether the 0.025v drop on a cap is a short or a Northbridge, I simply power it at a safe voltage and watch the current.

If one cap on a rail shorts then all caps on the rail test as shorted. The diode beep continuity scale finds the caps on the shorted rails and once you know what all caps show shorted, some current from a bench supply will single out the bad one out of many without blowing it up. It's good to know how much current is being delivered. One burned internal trace and the board is toast.

So far all I've needed to find these shorted caps is one hand to feel the caps and chips and the other hand to keep the red power probe on a randomly selected cap on the short list. Many of these shorted caps have visible defects. I've seen one with a hole in the side and others with white marks that I didn't cause with the scratchy meter probe. I keep a close eye on any cap that seems discolored.

I use the diode scale and mine beeps when the voltage drop is very low. The voltage drop of a short is almost 0.000v. The voltage drop of a CPU is as low as 0.008v. The voltage drop of a Northbridge and other high power chips range from 0.018v to 0.065v.

Here's a Gateway MA3 31MA3MB0021 "no power up" board I bought in hopes of doing it again. The shorted rail was easy to find with that many shorted caps. All the boxed caps show 0.001v voltage drop on the diode scale. The bad cap was hiding under a black sticker. 1.2 amps wasn't enough to rat it out. When I spotted the discolored cap and checked to be certain that it wasn't any of the other caps I started upping the current. 2.5A heated it right up.

I haven't completed this laptop but I'm confident that this one cap will bring it back to life. There's so many caps in parallel I don't even need to replace it. Just pull and go. Another fine example of bean counting.