The Fat Lady



The Fat Lady is a 15 watt amp which was originally modeled after the Marshall 20 watt amps. At least to the point of having a bit more gain than a Fender (nowhere near as much as today's Marshalls) and a pair of 6BQ5's for the output. Its pretty much a bark and growl amp with a controllable sag.

Basically the circuit is one gain stage, a fender style tone stack (with some funky values since you can't find 250k audio taper pots these days), two more gain stages, and a cathodyne phase inverter. The final gain stage and the phase inverter are a 12AU7, the first two gain stages a 12AX7. As a concession to practicality, I ended up putting a master in at the grid resistors to the 6BQ5's, and it actually sounds pretty good with the master down, but the raison d'etre of this amp is to run wide open. There is no negative feedback loop - I love what this does for the sound, but one side effect is that the amp is particular about the speakers it likes to drive.

The other thing of note is that you can select one of two bias settings, and two bypass values (in addition to running the 6BQ5's unbypassed) from the back panel. I did this as an experiment - the 6 position switch turned out to be overkill. Switchable bias turned out to be pretty cool. Having the option of running unbypassed can also be interesting - it made things slightly more compressed, and a little fuzzier. The partial bypass values didn't do anything real exciting.

I was really chasing a musically controllable compression/sag with this project, hence the 5Y3 rectifier. B+ filtering/decoupling is rather interesting on this point. At one point I was trying to get the effective resistance as high as possible - and it actually worked against me. the first side effect was the IR drop to the plates was so high, that I had to estimate plate voltage/current with a quadratic equation. If I took a plate voltage measurement, figured out the target idle current, and set the bias for that level, the plate voltage was off considerably from the value I used for my calculations. The second problem was using really high resistance in the power supply chain was that the time constants became so long that I couldn't get any sag under load. I ended up using a minimally decoupled power supply - and decoupling each stage individually.

Schematic

Download the Postscript version of the schematic . (2/15/96)

For convience and neatness, I combined paralleled resistors. The transformers listed are the ones I used. If you decide to build this, I would recommend using a heftier power transformer, as the one spec'd in the schematic is rather underrated (but nothing's melted down yet).



Pictures



[Outside Shot] This is looking at the chassis from the front (the knobs should have tipped you off.) B+ decoupling lives in the box on top of the chassis, which also serves to shield the amplifier circuitry from the power transformer and rectifier somewhat. It also serves to prevent me from mounting this chassis a combo with a 3/4" baffle board, as it's a bit close to the front edge of the chassis.



Internal Shots

I tried to stay with the conventional color codes for wiring where I could. The blue bundle is a Fenderish attempt at shielding the wires to the tone controls - it will be replaced with a proper foil shield at some time in the future.


[Chassis Pic] This is a wide shot of the chassis from the underside, with the back panel on the top of the picture. Here you can see a bit of my chassis construction. Basically, it's two pieces of aluminum u-channel (the front and back panels) with a sheet of 0.060" aluminum bent around it in a 'C' shape, and riveted wherever I could stick a rivet. There's another flat sheet of aluminum that screws on to enclose the box, but that would be a rather boring picture. The box was surprisingly stiff, but for good measure, I used rather large oversize washers on the transformer bolts to prevent any deformation/pulling out. (The output transformer isn't too heavy, so I didn't bother.)


[Chassis Pic#2] Here you can see the filament transformer - I ran the preamp off 12.6v DC heaters. The aluminum BFR is an 8 Ohm - 50 watt dummy load - rather than short the output speaker jack like Fender did, I ran it through a dummy load. This is a good safety measure, an unbelievable convince on the bench, and a handy dandy power soak all rolled into one.


[Power Supply] Here's the power supply section. I yanked the regulator out of the preamp heater supply because the transformer, which was rated at 12.6Vac @ 450mA, drops too low - I'm getting about 11.5 Vdc after filtering. I should be drawing only about 300mA. I think I can squeeze a 15 VA transformer inside the chassis. If not, I really don't have any place to mount one externally.



Special Thanks

Special thanks go to RG Keen and Tom Balon for all their debugging help, and to Tom Martin for graciously letting me use his workshop.

Back to Nate's Projects Page


Nathan Stewart, npstewart@pobox.com
$Revision: 1.3 $
$Date: 1997/02/08 08:23:55 $