by EWBrown » Thu Apr 28, 2005 8:39 am
Did you use the gold-plated contact tube sockets? sometimes, the process of soldering, the solder will stick and flow very nicely to the gold, but then the plating can lift off the metal pin and cause this symptom.
While you're conducting the forensic investigation, look for any cracked resistors, they can do some pretty weird and annoying things.
Lastly, pull out the 12AX7 and 12AU7, just leave in the EL84s, and listen for any suspicious noises, this will help isolate the problem even further. Since the EL84s are capacitor coupled (thus DC isolated) the tubes should just silently idle at their present current (approx 35 mA).
If just the 12AX7 is pulled, and the 12AU7 lis eft in, the amp will oscillate loudly, due to the PFB from the UL/SG tap (been there, done that).
A good trick to isolate a possible bad component or solder connection, is to gently tap each component with an insulated (plastic or dry wooden) rod or dowel, anything that is bad, microphonic, or poorly connected will show up readily
In the "X-files" or "twilight zone" category of problems , I did have one that I may have posted here last year, thhe left channel had some strange intermittent noises and distortions, and the biasing on one of the tubes would become very unstable. I suspected a high (ultrasonic or low RF) frequency oscillation was screwing things up. Upon further investigation, I found a low positive voltage on the offending tube's G1 grid (they should be zero or just a few millivolts). I tried changing tubes, even changed the coupling caps, suspecting a leaky one, none of this cured the problem.
My last ditch effort was to disconnect the UL taps on that side, and try it in "triode" mode - problem solved. There was something strange with the OPT (It was a Chinese made "DARED" brand OPT) and even though the UL tap showed the correct voltage ratio of 43%, the winding may have been reversed internally. So, the permanent fix was just to change the amp over to triode mode on both channels -something I wanted to try anyway, and I'm very happy with it. The power level is closer to 5WPC rather than 17, but it sounds truly excellent! For triode mode, just disconenct the UL taps, and connect a 100 ohm to 1K 1Watt resistor between the plate and SG/UL connection at each of the four terminal strips. And remove that little 15-20 pF cap, as it is no longer needed, and will introduce other strange situations... The bias may or not have to be re-tweaked, but that is quick and simple to checkk.
.
HTH
/ed B in NH
Real Radios Glow in the Dark