Well, I know the Fender engineers placed this diode in the circuit for a good reason. But here's another way to look at it. People who like and use the (older versions) Princeton Reverb many times get around to adding a bias pot to the bias circuit. I've done this many times and I own a few of these amps myself (older versions, BF and SF). Once the zener diode is removed (from the RI version), it looks just like the 'lightly modded' bias circuit (with added bias pot) that we've added and seen so many times. The pot itself gives plenty of variation for either hot or cold biasing. These Princeton Reverb bias circuits have been chugging along for 50 years + and seem to work great without any real inherent problems.
When I benched this PR-RI in question, the first thing I noticed was no bias voltage on Pin 5 of both power tubes. Tests revealed that the 1N5370B zener diode had shorted, causing the bias problem. Once removed from the circuit, the bias performed just fine (the zener is in parallel with the 100uf@100volt bias cap). As I didn't have this zener diode in stock and the guy needed his amp back right away for gigging, I checked out the bias schematic carefully and saw that it was exactly the same to an older Princeton Reverb with an added bias pot. The bias pot had a broad enough range window to accommodate a wide variety of 6V6 (and related) tube types, bias numbers and preferences.
I'm sure I've repaired well over a hundred Princeton Reverbs over the years and this is the first time I can recall this model (or any other Princeton/Princeton Reverb) having a bias problem/failure. So why add the zener? The other thing is that this is a relatively new amplifier, not a 50 year old amp needing lots of TLC and suffering from many drifted out of value parts. The guy who owns and plays it is an electric player in an otherwise all acoustic band. I doubt he gets the volume over 2 for most gigs (He's not really driving it hard at all).
Comment