1. Increased cathode bypass resistor on the 12b4a output tubes to 250 uF . No audible change, so no changes there
2. Add film bypass capacitors to the power supply and cathode bypass electrolytic capacitors. Smoother sound, less grain. Definate yes.
3. Add a 0.5 ohm dropper resistor in one leg of the filament secondary to drop the filament voltage from 7V down to 6.3V. Ideally, you would want to use two 0.25 ohm in each leg of the supply, but come on, 0.5 ohm is miniscule compared to the power transformer's DCR. Worked like a charm.
4. Build a IXYS IXCP10M45 CCS and replace the film resistors on the 12AT7. I find that with the CCS
- Gain is increased
- Transparency is increased
- Drive and also treble resolution are increased
- The warm "euphonic" sound effect is reduced. To be expected, but I kinda miss it.
- The miniscule amount of hum (only heard with ear 1 cm from driver) I had before is pretty much gone (increased PSRR, I guess).
Here's the new schematic:
Update: September 20, 2008 .
I occasionally noticed a little amount of "zzzzz" sound . It turns out there is a little blocking distortion action going on. "Kenpeter" suggested a couple of "fixes".I've tried the first circuit (http://www.diyaudio.com/forums/attachment.php?s=&postid=1607995&stamp=1221201218) The debuging saga is documented at http://www.diyaudio.com/forums/showthread.php?s=&threadid=129577
Here's a link on blocking distortion