VALVE Magazine Online!
My Own Triode Input MK III
A remarkably bad graphic representation of a very pretty amp. Squint to see the Tung-Sol 6550's, new tube arrangement, Vitamin Q caps, new bias pot location, and cool paint job.
Some of you may remember that I was so enamored of the Mklll's I modified for Chris that I wangled a pair for myself. As I recall I bought a Fisher 20A, swapped that and some other stuff for a MkIV, and then swapped the MklV and a pair of Ampex monoblocks for the Mklll's. Then I swapped a Philco Model 60 cathedral radio for a quad of Tung Sol 6550 tubes.
Then I made my list of parts for the mod. I ordered polypropylene caps and metal film and power resistors from Mouser, and axial electrolytics from Antique Electronic Supply. Mean time Eric scored some 1.0 mfd @ 400V Vitamin Q caps for me on their way to a dumpster. I stripped the chassis, repainted everything in Classic Radio colors and enlarged the filter cap mounting hole into a proper octal socket cut out. Output tube sockets went to the outer edge of the chassis, where the rectifier and filter had been, and the rectifier and a new 6SN7GTB driver were installed where the output tubes had been.
The circuitry was rewired like the original mod. Write us if you'd like a copy of the original article (grid resistors were reversed in the schematic. See Corrections, p.3). I added a balance pot, based on the Boak mod in TAA. This pot was installed where the bias pot had been, and the bias pot was installed through a plate mounted In the preamp power socket hole on the front panel. Filters were crammed under the chassis as follows: first stage - 50 mfd @ 900 VDC (two 100mfd @ 450v in series with look ohm equalizing resistors) second stage - 100 mfd @ 600 VDC (two 200 mfd @ 300v in series with 100k Ohm equalizing resistors) each following stage - 100 mfd @ 450V. Bias caps were increased to 100 mfd @ 160V low ESR.
The power cord was replaced with a grounded shielded model, and MOV's were placed across the mains and from each leg to ground. The selenium bias rectifier was replaced with a silicon fast recovery diode. Once again I used my cool 'star ground bus'. I take a long terminal strip tied to the chassis and solder # 12 wire along all the terminals. All ground leads go to this strip, with input ground at one end and power supply ground at the other. Liberal use of terminal strips under the chassis makes all these mods doable if not perfectly accessible.
The Vitamin Q's sit on the PC board (to show them off, of course) and couple the phase splitter to the 6SN7GTB. The 6SN7 is coupled to the 6550's with 630V polypropylenes, as I haven't collected quite enough Vitamin Q's to do the job. During the smoke test everything went great. When I connected input from a preamp and I got a little hum. I've been testing for audible hum with Eric's A7's or my Stretchorn. If hum is going to show up, your gonna hear it through a horn loaded woofer!
It turned out that the original schematic spec'ed a 10 Ohm resistor from the input minus to ground. When I smoke tested with a grounding plug in the input, this resistor was shunted, so I got no hum. Replacement of the resistor with wire solved the problem and I got another nice quiet amp. In the original mod, I found that coupling the screen to the plate (for triode input, get it?) in the pentode section of the 6AN8 with a 100 ohm resistor was not as smooth sounding as coupling it with a .22 mfd capacitor. On this amp the resistor worked great, resulting in a bit more detail without highs getting too edgy. Fabulous midrange presence, great detail.
It juices my lizard.