The STV-3H diode is available from PartStore:
http://www.partstore.com . I didn't know you needed one. The wait can and will be lengthy, but you can get them.
Something which Markthefixer and I have briefly touched on is that, in the original circuit, the D4 diode, STV-3H, is the one used to bias the output stage. The D11 diode, the STV-4H, is simply used as a shunt to bypass the D4 diode and the bias adjust pot...essentially acting as a 'bias limit' diode. It also protects the output devices in the case of the bias pot opening up.
Simply put, the D11 STV-4H diode limits the bias adjustment range by imposing an approximate 2V limit on the voltage that can be generated between the collector of Q5 and the collector of Q7.
Well, let's see...there's another device that can force a nearly-steady voltage between two points: A zener diode. The Panasonic MAZ4020 is a 2V, 370mW zener diode in a DO-34 axial package. It's characteristics may not match the original diode to a 'T', but as far as acting as a 'safety-valve' in the case of a failed trimpot, it should work fine. (Digikey PN MAZ40200LFCT-ND)
Note that the anode and cathode are reversed as opposed to the original STV-diode in order for the zener to perform properly.
The DO-34 axial package can even be mounted fairly easily by covering the leads with tiny shrinkwrap, and a very small ring terminal (such as the #5 hole ring terminal A27150-ND part from Digikey). The wire insert section can be straightened out and hammered flat, and the diode attached using JB Weld, which is a good conductor of heat.
Certainly I've never tried this, but I see no reason why it shouldn't work. I'd want to monitor the bias over a range of heatsink temperatures, but if it all looked good, I'd call it a done deal.
Edit: I meant to mention that, on many Spec 2 driver boards, one of these diodes (D4 or D11) has the silk-screening backwards...I cannot specifically remember which one. Check the orientation of the diodes currently in place before anything is removed. Do not rely on the silkscreening to get it right!
Edit #2: I'd be tempted to add a small film cap (such as a Panasonic 0.56µf 63V stacked film) across the zener to counter the effects of the dynamic zener impedence with AC signals. I'm not sure if it is important here, but it couldn't hurt anything. The cap will essentially perform an AC short from the collector of Q5 to the collector of Q7, and can be added to the back of the driver board (so no hassle with trying to remote-mount it physically with the diode).