Replacing Driver Transistors (Help)

I'm thinking that replacing the driver transistors may be the next step in addressing bias creep in my RD65-112. I've got the transistors from Mouser but am getting a little anxious on how to replace them...
I've got a fair amount of experience in building/repairing point-to-point amps (and replaced all electrolytics and bias resistors on the RD65 PC board). However, I've never dealt with soldering delicate things like these driver transistors, which I assume are very heat sensitive...
Can any of you with such experience please advise on the install process (e.g. iron temperature, heat sinks, anything else relevant!)

My 2 cents:
Get the iron hot so you don't have to spend too long soldering. Those transistors aren't *that* delicate. If you know how to solder a 1/4W resistor without frying it, you should be good.
Without knowing the back story: When you say 'bias creep' - does that mean that the current draw of the output tubes/transistors is unstable over time but the actual bias voltage is stable?
Cheers,
Lars Verholt

Thanks Lars! What I'm getting is an increase in readings from the emitter resistors -- I've observed an cumulative increase 6 ma over a 1.5 hr period. I've tried different sets of tubes (they are 6L6s in this version), and still get the upward trend...

Ok, so the power tube/driver transistor pair on both sides of the push-pull show this trend? I would measure the actual bias voltage. If this voltage creeps upwards over time, it would explain why both sides are pulling more current. Another culprit could be the usual leaking 20uF electrolytic coupler.
Cheers,
Lars Verholt

Thanks again, Lars for troubleshooting hints on this one. FYI, I've replaced all electrolytics on the board with Atoms (as well as the main filter caps in the doghouse), so doubt that a leaking cap is the issue. What I've been measuring is voltage across the 3.9 ohm emitter resistors, as per the Service Bulletin #1 procedure -- I get creep across both resistors. Given the weird way these things are wired, how do I measure actual bias voltage?

The bias voltage is created by two diodes in series (D7 and D8). The 16V (positive!) main supply voltage is fed through a 3.3k@1W resistor (R32), through D7 then D8 to ground. If both D7 and D8 are working properly there will be a steady 0.6V across each unit meaning that there should be 1.2V at the junction R32/D7 and 0.6V at the junction D7/D8. A trimmer pot is connected in parallel with D7 meaning that the bias regulation voltage fed through the junction R34/R36 will be somewhere between 0.6V and 1.2V.
If either diode is no longer stable, the bias voltage can drift.
Another possible culprit (and I just thought of this now) is of course the 22V grid bias supply to the 6L6 tubes. The 22V supply is governed by Zener diode D14.
To measure the 22V supply voltage, measure the voltage at pin 5 of the 6L6 tubes. If your creep in voltage across the 3.9 ohm resistors correspond with a rise in voltage at either pin 5 of the tube or at the base of the corresponding driver transistor, you have solved the puzzle.
Clear as mud?
-Lars Verholt

Just a little added info-I usually(if I get paranoid) use a pair of forceps on the transistor legs to act as a heatsink AWAY from the transistor. Does it work? Don't know. Haven't torched any yet but I HAVE been soldering transistors, fixing boards, and in general, tearing up stuff for 58 yrs(I'm old) so I go pretty fast. You are looking for good wetting and minimal time in soldering on trace boards. Just do NOT try any of that lead free crap until you are GOOD at it and don't try to MIX the two-period. JMHO. Mike.