throbscottle
Well-Known Member
I have had a Devolo AV Desk 200 sitting in my "to repair" pile for far too long, and decided I'd have another crack at it whilst the bench is relatively clear. Lucky it's one of a pair, so I could compare with the other one.
So, what I've established is, it gets an ethernet connection alright, but no ethernet over mains connection - the link light doesn't even come on, apart from flashing briefly at power-on. So far I've re-flowed the board with a heat gun, and replaced the high voltage smoothing cap since the smp seemed jittery. Smp looks better now but still no link.
The device uses an OPA2674 to drive the communication over the mains, which gets it's signal from an AD9865. The opamps are sitting with their inputs and outputs close to mid-supply.
Comparing signals, the good unit gets a nice burst signal on the outputs of the OPA2674, and what on my scope looks like pulses feeding into it, though probably burst signals, I just can't see them. Without going to the effort of actually tracing the signal path, it looks as though the opamps are fed from the IOUTx outputs from the AD9865 - similar signals seen on good unit. I will have to study the data sheet to get a better idea.
On the faulty unit, one pair of IOUT signals is missing. Not sure yet how this ties in with the opamps having no inputs at all - a few passives in between. I don't think the opamps are faulty. Have to re-test.
Only other bit of evidence is that a series resistor in the smp had been getting so hot it completely blackened the board under it right the way through, but doesn't get hot now. Good unit same resistor has also blackened board, but not nearly as much. So methinks something was drawing too much current but isn't now it's failed.
So, does anyone think it's likely to be the AD9865 at fault? Or should I be looking at the discrete devices around it? Anyone with experience of these things able to shed some light on it?
So, what I've established is, it gets an ethernet connection alright, but no ethernet over mains connection - the link light doesn't even come on, apart from flashing briefly at power-on. So far I've re-flowed the board with a heat gun, and replaced the high voltage smoothing cap since the smp seemed jittery. Smp looks better now but still no link.
The device uses an OPA2674 to drive the communication over the mains, which gets it's signal from an AD9865. The opamps are sitting with their inputs and outputs close to mid-supply.
Comparing signals, the good unit gets a nice burst signal on the outputs of the OPA2674, and what on my scope looks like pulses feeding into it, though probably burst signals, I just can't see them. Without going to the effort of actually tracing the signal path, it looks as though the opamps are fed from the IOUTx outputs from the AD9865 - similar signals seen on good unit. I will have to study the data sheet to get a better idea.
On the faulty unit, one pair of IOUT signals is missing. Not sure yet how this ties in with the opamps having no inputs at all - a few passives in between. I don't think the opamps are faulty. Have to re-test.
Only other bit of evidence is that a series resistor in the smp had been getting so hot it completely blackened the board under it right the way through, but doesn't get hot now. Good unit same resistor has also blackened board, but not nearly as much. So methinks something was drawing too much current but isn't now it's failed.
So, does anyone think it's likely to be the AD9865 at fault? Or should I be looking at the discrete devices around it? Anyone with experience of these things able to shed some light on it?
