Noisy dead router after possible power problem

[xieliwei]

I'm not sure if its a power problem, since no other electrical appliance in my house was affected (even those that were more sensitive to power issues), but its too much of a coincidence that two different network routers running from the same power strip failed mysteriously at the same time last night with similar causes of death.

The first router had a damaged wall wart. The second router had a damaged power supply (within the router, I don't know specifically which part of it died, more on that later). It should be noted that while both routers feed on DC power from seperate wall warts, the first one (which is damaged) is switched-mode while the second one (which still works) is linear.

The reason I mention damaged is interesting. Whatever happened did not totally break the wall wart/power supply, instead, they seem to perform well until a certain load.

This was confirmed for the first wall wart as a seperate (but same model) router that was working fails under high load (more devices plugged in, more data flowing) as well when I swapped the wall wart over. Thus the resolution for the first router was simple, replace the damaged wall wart with an identical spare I had available and its working well again.

The second router is much trickier since it is the on board power supply that was damaged (meaning, I can't just swap it). The symptoms are similar, everything is fine until a certain load (in this case, when the wireless module starts), then the power becomes unstable and the router locks up and supposedly the onboard watchdog resets it. One additional symptom however is that there appears to be a hissing sound coming from a inductor when the power beomes unstable (I can see the power light flickering according to the hissing).

Preliminary tests show that:
1. Input and output electrolytic filter capacitors (C20 and C95, maybe C8) are not shorted nor open and resistance slowly goes up to infinity when charging, but I don't have a capacitor mode on my DMM nor can I measure their ESRs. C20 has a mild bulge though. (I also can't determine if the caps are fully functioning well or just fails under load)
2. The polyfuse F1 seems to be working properly and does not trip when the router is turned on or when the power becomes unstable.
3. The voltage regulator (VR1) provides a ~12+-1V output, within its specification's limits.
4. The unknown diode (D1) (by DN? Model 514?) seems suspicious as I can measure a voltage drop in either direction with the diode mode on my DMM. The forward voltage drop is about 0.115V, and 0.794V when measured the other way round.
5. The inductor coil at L11 is the source of the hissing sound. I don't have any tools available for inductance testing but I know the terminals at both corners measure close to 3.0V with respect to ground.
6. There are two more voltage regulators (VR2 and VR3) on the other end of the board, both seem to be working properly.
7. Q2 is (I think) a PMOSFET. I don't have the necessary time to test it, but I guess it can be a cause too.
8. I don't think there is any unusual heat generated. Most of the heat comes from the Atheros controller and the ethernet switch.

With the information above, my question to all is:
1. Why did all this happen? Did I really experience a power problem?
2. What is wrong with the power supply of the second router, how do I fix it?

The wall wart of the second router is DC 12V 1A, pin positive.
Below are photos of the circuit board of the second router.

http://www.rvnet.ath.cx/xieliwei/router/DSC04817.JPG
http://www.rvnet.ath.cx/xieliwei/router/DSC04818.JPG

All help appreciated!

 

[zevon8]

The cap with a bulge is very suspect. Replace it and the issue is likely solved. It probably has high ESR now, and under load will not perform properly, resulting in voltage fluctuation "noise" on the DC supply output. This is causing the extra mechanical noise from the coil.

The linear wall wart likely produced a high output during a voltage surge, or its own capacitors ( if any ) died, resulting in poorly regulated output at the DC input of the router.

The dead switcher wallwart may have simply killed itself with an internal fuse, or similar.

New capacitors ( C20 and in the wall wart ) will probably fix the router. If it works up until a heavy load, my bet is that there are power supply issues.

 

[xieliwei]

Thank you for your response zevon8.

The bulge is very mild, but I guess any gas build up is suspicious. I'll try changing it along with the other six I see on the board today.

The linear wall wart produces about 16.4V unloaded and drops to and fluctuates about 14.1V when plugged into the router and turned on. I know about wall warts producing a higher voltage than rated, but is 14.1V running a sign of damage? Should I get it replaced as well?

I should note that I've tried powering the router with a generic 12V 1A switched-mode wall wart (that should be working fine) but the problem persists.

In case the capacitor(s) is/are not the problem, does anyone have other suggestions? My next suspicion is D1, why is it that I can measure a small (less than 1V) potential drop in either direction?

Thank you to all!

 

[zevon8]

The wall wart provides a basic suppply voltage, the internal supply in the router does final regulation, along with any secondary voltages needed. If the wall wart cannot provide a solid supply because of a weak capacitor, the internal supply will not be able to do its job. If the internal supply has been damaged, even a clean wall wart will not help, since the internal capacitor in the router cannot hold a DC level.

D1 likely has other components accros it that will give you a misleading measurement. Unless you remove it, you cannot trust an in-circuit reading without knowing waht other circuitry is connected.

 

[xieliwei]

Hmm, it is the capacitor alright! The input filter capacitor's bad.

I'm having a problem replacing it though. I guess the manufacturer used solder with a higher melting point than my iron can handle. It refuses to melt even though the iron has no problem melting the solder I use. I'm currently attaching the replacement capacitor in parallel with my own solder.

Anyone with any idea on how I can fix this?

 

[Nigel Goodwin]

Presumably it's lead free solder, which has a higher melting point, however a normal iron should handle it OK. If it's temperature controled then turn it up higher, if it's not temperature controlled, then it should be hot enough anyway.

For a start, resolder the offending joints using your normal solder, this will lower the melting point of the entire joint.

Any capacitor which is bulging is faulty - REPLACE it.

 

[xieliwei]

Hello, thank you for the answer!

Hmm, I managed to remove the old capacitor and (barely, hole became very small, I could not totally remove the solder) replaced it with the new one using Nigel's suggestion.

The router is now working very well! Thanks to all!

(I might not have noticed the offending capacitor if not for another same capacitor on the board. The bulge was so small that it seemed normal.)