"The Evil Sandman" Preamp

[wrenchone]

This thread documents a preamp project I've been contemplating since a little before Burning Amp '09. I'm christening. I'm christening it "The Evil Sandman" because its pristine vacuum state is marred by some bits of nasty silicon. The RIAA portion has remained pretty much static since its conception (triode based, passive equalization), but the lineamp portion has seen quite a few changes as I wrestled with this and that circuit concept. I settled on a non-inverting stage with gain based on a commonly available power pentode.

The picture below shows the first breadboard of the lineamp. I actually powered up up that crazy little cats-cradle of parts just a half hour ago, and the bias voltages look right. When I have time, I'll take a look and see whether it actually makes noises like a lineamp... Schematics will follow when I'm reasonably satisfied. I'll then hammer away at the RIAA preamp.

This project will use all commonly available (at least in the US) 7-pin tubes. More later.

[wrenchone]

Well, the lineamp prototype certainly makes noises like one. Attached is the 10 kHz sine wave response.

[wrenchone]

Being the suspicious sort, I always give the input a kicking with a 10kHz square wave within the linear range of the circuit to see what comes out - in this case, some undershoot and overshoot.

[wrenchone]

A compensation cap of 20pF cleans things up rather nicely, and, and I happen to have a bunch of 20pF NPO discs hanging around at the home front. Cleanup here was done using a sleazy-a** Z5U disc, good enough for illustrative purposes.

[wrenchone]

Homing in on the edges makes the compensated line amp look like a slow poke, but the big picture shows otherwise... I'll doctor the schematic and post it when I get home. It's evil and rather sandy.

[wrenchone]

Here's the circuit I'll be using for the line amp. In the rough prototype, Q1 was replaced with a ring-of-two 1mA current sink, but the JFET eliminates a low voltage positive supply that won't be on the real deal. I ran the prototype with a DC filament and a Xantrex HV bench supply. The real supply will involve some sneaky subterfuge with a 50VA dual input 12V + 12V toroidal transformer. More on that when I verify that I can get away with said subterfuge.

[planet10]

A lovely piece of sculpture

dave

[Bonsai]

I would have called it the 'Dirty Sand Line Amp'

Just kidding. Looks very nice.

[Geek]

Hey Dave, we need some little gadget like that at AudioFest!

Cheers!

[wrenchone]

More evil sand. This configuration makes better use of the 10 mA allotted to the tube portion (more transconductance for a given plate current). Despite all the sand scattered about, the tube is the boss. I may omit the input cap and put the volume pot pot (with a safety resistor from wiper to ground) right at the input. Some care may be in order, as these slippery RF tubes are reputed to run a few microamps of grid current even when reversed biased. Anyway, it won't hurt to breadboard the thing to see how it performs.

This amp accomplished several things : low gain, no inversion, and an output voltage centered reasonably close to 1/2 plate voltage for wide dynamic range.