Aleph4 gate proection zener failure

[Allen Newby]

I have been given a 1/2 dead aleph4 to repair. One channel is OK, the other outputting +45V.

On closer examination it turns out that the gate protection zeners on the lower driver boards have failed and are short circuit. (ie. the normal failure mode for zeners). The devices are 9.1V 1N4739's and are rated at 1W.

Now, obviously, in 'normal' operation they have no effect. Even if they do kick in, it is hard to see how sufficient current could flow through the device to damage it.

If you look at the board. The zener is mounted in close proximity to the heat sink. I suspect that the zeners have failed due to ambient thermal stress rather than current flowing through the device.

If I am right, a 50cent device has blown a megabuck amp and speakers which sort of defeats the object!

Qu1. Has anyone come across a comparable situation?
Qu2. What to do? replace them or remove them?

[Nelson Pass]

Classic example - protection devices cause the failure.

[Allen Newby]

Thanks for the reply. I am still wondering what to do.

I think it is a question of thermal derating caused by high ambient temperatures and possibly ∂V/∂t triggering: There may be sufficient gate capacitance, 6x1300pF for fast avalanche breakdown.

Do we:
i) remove the zeners
ii) replace them and hope for the best
iii) move them to the mother board where it may be a bit cooler
iv) use a resistor in series with the gate protection zeners to limit the avalanche current

[Nelson Pass]

Replace them, bring the amp up on a Variac while checking
for DC across any of the Gate resistors.

[Allen Newby]

I replaced the zeners and was puzzling to work out why the lower output transistor wasn't conducting. Turns out that the track from the long tailed pair on the underside of the board had fried!?

I peeled off the remaining loose bits of track and replaced the track with a jumper wire. Turned the amp back on and all was fine. Lovely!

A few days later, I screwed the lid back on, ready for Mani to collect it. Turned it on for a last minute sound check and pfzzt! accompanied by the dreaded smell of fused components.

Apologies to Mani, back to square one, lid off again, more dead parts ordered and replaced. This time the root cause was obvious. All along, the the long tailed pair were mounted too high on the mother board. The heatsinks were shorting to the underside of the lid! Dammit, why didn't I work that out first time round.

Maybe they had been protected by a few microns of aluminium oxide and a bit of thermal movement, or perhaps a knock and the short would occur.

Lessons?

1. I am glad I took Nelsons advice and replaced the zeners. Since these fail to a short, they protected the output Mosftets. Replacing these would have been a pain. Matched metal can devices as rare as hens teeth.

2. In diagnosing a fault it is too easy to think that the circuit diagram in front of you actually represents what you actually have on the bench. Fused and disappeared tracks and case shorts that don't exist when part disassembled throw you off the scent.

3. Don't look for a complicated explanation when a simple one exists, even if you can't see it yet.