Finding excessive load on large circuit board

PFloyd36069 · 2017-05-03 01:55

Hello All,

I recently acquired a Denon AVR-3313CI A/V receiver that is not working properly. I traced down the malfunction to there being excessive load on the 3.3v circuit pulling the voltage down to around 1.5v and causing the DC-DC converter to get overly hot. I unsoldered the two zero ohm resistors that connect the converter to the rest of the circuit and the voltage came back up to 3.3v. Is there any way to trace down the cause of the excessive load or is the unit trash? I was thinking maybe use a thermal camera to see what components get hotter than the rest.

Thanks,
Bryan

kwinn · 2017-05-03 02:06

First thing to look at are the electrolytic capacitors. Remove any that look heat damaged or are bulging and check the voltage. Replace with new ones if that solves the problem. A thermal camera is a good next step.

Peter Jakacki · 2017-05-03 02:18

Electrolytics usually go dry so that's not the cause of excessive current but tantalums do short although I can't see any. Multi-layer pcbs are hard to check but I would put an external 3.3V onto the board that can supply a bit more current but still limited and then you should be able to feel the heat. btw, I find that you don't even have to touch the board because even by holding the side of your face near the board you can sense the heat quite readily. If it's one of those chips though then it probably isn't repairable.

whicker · 2017-05-03 02:33

Found a bad (shorted) tantalum capacitor with a thermal camera.
It was the only thing glowing in comparison to everything else on the board.
For this particular unit, the power supply kept shutting down and retrying, making the shorted component hot enough to see, but not noticeably warm to the touch.

Tracy Allen · 2017-05-03 02:40

The thermal camera can help but may be subtle. If there is a dead short somewhere, that point won't dissipate power, but with a thermal camera you can see the hot trace or net leading up to it. On the other hand, sometimes the thermal camera will detect an obvious culprit that you could also detect with your finger.

Another way to trace is with a precision voltmeter (like 6 digit). Or with a less accurate voltmeter and a DIY battery powered amplifier. Trace the voltage drops from the source to known power pins on the ICs. In general, the greatest voltage drop will be near the fault. That works best if the board has kind of a star layout and is much easier if you have a schematic and layout diagram.

Do check with a multimeter in first the ohms and then the diode position to see if the effect is a pure resistance or if it is a has a threshold. The 1.5V reading suggests a diode drop. Even better, use a current limited variable supply and increase the voltage slowly to see at how the current increases with voltage, linear, or at a threshold. A dead short will be linear, trace resistance only.

PFloyd36069 · 2017-05-03 23:05

Wow, that's for all the ideas guys! As soon as I find time to work on it again, I'll try some things you mentioned and get back to you!

frank freedman · 2017-05-04 03:00

Check also the connectors where the power comes into the board, visually or IR. Look for discoloration of the connector shell and oxidation on the conductors. Have had connectors go resistive and they will drop part of your voltage, generate heat or even melt connector shells and pins.

Peter Jakacki · 2017-05-04 05:08

The external power supply is the fastest solution from my experience with these things. You don't even need 3.3V but a bench supply with current limit and current metering. Sometimes I use meter probes from the supply just so I can track down a more elusive short. The current meter will read the highest the closer you get.

PFloyd36069 · 2018-07-24 19:25

Hello All,

I just wanted to give an update on the A/V receiver. I used the thermal camera and found the HDMI switching IC to be causing the excessive load. I swapped the IC from another board I had (the other board had corrupt firmware) and the unit has been working properly for over 8 months. Thanks for all your help.

Bryan