A couple weeks ago, i had one of these AVR's brought in; fault description = totally dead.
Took the cover off (coveniently missing screws), and after a quick visual inspection, noticed a glaring issue - see attached photo no.1. Yep, great big honkin' crater in the MCU.
"Obviously" no chance of finding one of those on its own, and even if i did, i couldn't know if/what else was blown on the same board. Somewhat fortunately, i found a same model board from a seller in the US on fleabay:
http://www.ebay.co.uk/itm/2626112563...%3AMEBIDX%3AIT
As can be seen, sold as untested. Then again, nothing with craters in it, so at the very least, it's in better physical shape than the original one. I explained that to my client, and that i'd charge him for the cost of the board regardless of whether it works or not, and he agreed.
As a side-note, i like how the faceplate of the thing had started to melt (see attached photo no.2). I'm not sure i wanna know how cramped & hot this thing ran, when it still did...
It finally arrived yesterday. I install it, hook up all the connectors, and that's where the click on / click off thing began. After a few of those cycles, a puff of smoke from somewhere inside the depths of the thing, came out. That turned out to be a crapped out 7915 regulator (measured about 68 ohms input-to-output and vice-versa).
Didn't have any 7915's in my parts bin, so i "MacGyver'd" one with a 7912 with a LED and a diode in series with its ground pin (photo no.3) - this results in a nice -14.5v on the output - that's within 5% of the 14.9v that comes out of thr 7815 next to it, so i'm calling that "close enough"
That didn't help much though, but at least the dropper resistor "upstream" of it stopped smoking 
While i had the main board removed from the case, i noticed a "skidmark" on the inside of the case (photo no.4), as well as what had caused it. The 47k resistor between the AUX_L input and ground was toast, and the associated ground trace had been literally blown clean off the board (photo no.5).
During the testing and the "automatic" on/off cycles, i noticed the HDMI chip on the "new" board was getting unreasonably hot, in a big hurry. Turned out the 3.3v AVDD rail for the HDMI chip is shorted to ground. That's not the case on the original board (with the "well ventilated" MCU), so there's still hope for that one
Either way, all the series ferrites for the 3.3v HDMI supplies are removed for the moment, to eliminate that variable.
Then i decided to remove the MCU board altogether, and force the mains relay on (wire jumper between the MPOWERD pin and ground, on the wire harness coming from the stand-by PSU) - see pages 40-41 in the attached service manual. After a few seconds, a wisp of smoke came out from somewhere on the main board once again, but from a different place.
I removed the "mid board" (containing the speaker relays & binding posts and some analog I/O), and forced the thing on again. No more smoke, but i noticed some 1/2w resistors getting quite hot. These turned out to be the dropper resistors upstream of a pair of 6.8v zeners, which form the local regulation for Q4001, the audio mux / volume control chip - a Renesas R2A15218FP. Measurements showed the outputs of those to be -3.3v / +2.5v respectively. Safe to assume that's toast as well
Luckily those rails go through two wire jumpers on the top of the board (i haven't removed it yet again from the case). After snipping those two apart there's no more heat, and the 7912's LED doesn't wink out in-between the on/off cycles, with the MCU board reconnected.
So, as it stands:
- the 3.3v HDMI supplies are disconnected (pending chip swap)
- the +/-6.8v supplies to the R2A15218FP chip are disconnected (chip's toasted)
- the MCU "still" cycles the power on/off when connected
- the thermal fuse inside the power transformer is perfectly fine (getting about 6 ohm across the primary both before and after the fuse)
I don't suppose anyone has any idea of some "checklist" of causes which might make the MCU turn the main relay on & off, right?
Took the cover off (coveniently missing screws), and after a quick visual inspection, noticed a glaring issue - see attached photo no.1. Yep, great big honkin' crater in the MCU.
"Obviously" no chance of finding one of those on its own, and even if i did, i couldn't know if/what else was blown on the same board. Somewhat fortunately, i found a same model board from a seller in the US on fleabay:
http://www.ebay.co.uk/itm/2626112563...%3AMEBIDX%3AIT
As can be seen, sold as untested. Then again, nothing with craters in it, so at the very least, it's in better physical shape than the original one. I explained that to my client, and that i'd charge him for the cost of the board regardless of whether it works or not, and he agreed.
As a side-note, i like how the faceplate of the thing had started to melt (see attached photo no.2). I'm not sure i wanna know how cramped & hot this thing ran, when it still did...
It finally arrived yesterday. I install it, hook up all the connectors, and that's where the click on / click off thing began. After a few of those cycles, a puff of smoke from somewhere inside the depths of the thing, came out. That turned out to be a crapped out 7915 regulator (measured about 68 ohms input-to-output and vice-versa).
Didn't have any 7915's in my parts bin, so i "MacGyver'd" one with a 7912 with a LED and a diode in series with its ground pin (photo no.3) - this results in a nice -14.5v on the output - that's within 5% of the 14.9v that comes out of thr 7815 next to it, so i'm calling that "close enough"
That didn't help much though, but at least the dropper resistor "upstream" of it stopped smoking While i had the main board removed from the case, i noticed a "skidmark" on the inside of the case (photo no.4), as well as what had caused it. The 47k resistor between the AUX_L input and ground was toast, and the associated ground trace had been literally blown clean off the board (photo no.5).
During the testing and the "automatic" on/off cycles, i noticed the HDMI chip on the "new" board was getting unreasonably hot, in a big hurry. Turned out the 3.3v AVDD rail for the HDMI chip is shorted to ground. That's not the case on the original board (with the "well ventilated" MCU), so there's still hope for that one
Either way, all the series ferrites for the 3.3v HDMI supplies are removed for the moment, to eliminate that variable.Then i decided to remove the MCU board altogether, and force the mains relay on (wire jumper between the MPOWERD pin and ground, on the wire harness coming from the stand-by PSU) - see pages 40-41 in the attached service manual. After a few seconds, a wisp of smoke came out from somewhere on the main board once again, but from a different place.
I removed the "mid board" (containing the speaker relays & binding posts and some analog I/O), and forced the thing on again. No more smoke, but i noticed some 1/2w resistors getting quite hot. These turned out to be the dropper resistors upstream of a pair of 6.8v zeners, which form the local regulation for Q4001, the audio mux / volume control chip - a Renesas R2A15218FP. Measurements showed the outputs of those to be -3.3v / +2.5v respectively. Safe to assume that's toast as well
Luckily those rails go through two wire jumpers on the top of the board (i haven't removed it yet again from the case). After snipping those two apart there's no more heat, and the 7912's LED doesn't wink out in-between the on/off cycles, with the MCU board reconnected.
So, as it stands:
- the 3.3v HDMI supplies are disconnected (pending chip swap)
- the +/-6.8v supplies to the R2A15218FP chip are disconnected (chip's toasted)
- the MCU "still" cycles the power on/off when connected
- the thermal fuse inside the power transformer is perfectly fine (getting about 6 ohm across the primary both before and after the fuse)
I don't suppose anyone has any idea of some "checklist" of causes which might make the MCU turn the main relay on & off, right?
Attached Files
Comment