@hk_fan And once again you are wrong. My 336 caps failed which I returned to Levinson for repair and was told that ALL 3xx amplifiers suffered the same issue. It’s not the caps themselves, it’s the way they are used in the circuit which shortens the life of them.I can definitely agree with this. I believe ML knows this too. However the Phillips caps were a prominent weak point. Their service bulletins state that Phillips caps should be replaced with United Chemicon caps but I have read many cases where they replaced failed caps with Phillips caps again. Also these caps do have a finite useful life. If you have 10 to 15+ years on your amp then you should not be upset if they do fail. The amount of actual usage hours will affect this. When ML publishes safety bulletins (in an act of supreme irony they don’t publish safety concerns to the public) they are identifying areas of the amp that make engineers raise a brow. Three areas of concern were the Phillips caps, the contact points of the main caps and the in-rush resistors. The cap contact issue was rectified with mylar spacers The potentially violent failure of in-rush resistors was addressed by using open wound chromium wire resistors. Neither of these are parts they will sell to the public or 3rd party repair shops even though they are safety related. They will however automatically perform these fixes on any amp that comes across their service bench, at your expense. The main caps are punished every time the amp is started from a dead-off position. It is exponentially worse if one of the 3 in-rush resistors is failed open. One open resistor means the remaining survivors and the caps have to bear the in-rush; the next resistors will fail in short order. If your ML starts occasionally blowing fuses on power up, you potentially have a bad in-rush resistor or the main caps are going bad.
I will venture to say that those that keep the ML in idle (not off) when not in use will have longer cap life. In-rush current is not nearly as bad when turned on from idle. The problem with idle-off is you are burning approx 200 watts for absolutely no reason except to keep the output transistors at a nice toasty temperature. I for one am not going to watch my electric meter spin around faster with an amp chewing up 200 watts as I sleep or when I am at work. In fact I don’t know if I like the idea of keeping this hulking power supply of an amp alive at all when I am not physically home.
I cannot say the technology to bring the power supply up slowly with pulse width modulation was available at the time the 300 series was introduced, due to the current demands. I really wish it was. It would certainly make sense on modern equipment. These caps could really live a lot longer if they were slowly brought up to nominal voltage over the span of 30-60 seconds. I have even read of some guy trying to slowly bring these amps up to voltage with the use of large Variacs. The brute force approach of using single stage in-rush resistors with the Phillips caps that falter under this stress seems like an unfortunate mix.
Even with all of the above, which some might consider negative, I am here to say that the ML is still a superior product.