steve wrote:It's more of an educational exercise for me than anything else. I wanted to design an amp without having a schematic or kit or anything to force me to go through all the design restrictions and math, so that I would learn how to do it rather than just making something that someone else did the hard part for.
That's funny; about a year ago, I snatched up a crate of weird NOS tubes from Canal Street in NYC for about $20, and I've been looking up datasheets and designing and breadboarding weird amps ever since. Forcing myself to design an amp around a random tube is pretty rewarding and has led to some discoveries I don't think I would have otherwise blundered across.
I do have a couple of nice things that makes this easier though, such as a power supply bolted to a plank of wood, and a set of octal sockets with screw terminals on the topside, so I can get something up and working without even firing up my soldering iron.
steve wrote:The VA section is a cathode-biased single 5703 subminiature triode -- excellent bandwidth and about 20V p-p available. The power section couldn't be simpler: a single 6B5 tube. This is an unusual and long-outdated dual-dissimilar triode tube that has its own driver triode internally cathode-coupled to a triode power section. This saves me the trouble of designing as driver, one of the most contentious areas of amplifier design, and one that would surely cause me much hemming and hawing.
I'm seeing a lot of Internet schematics lately that call for the 5703. (And I think that tiny Z.Vex head uses them, as well ...) I'll have to buy a few and mess around with them.
One interesting tube I found: The 13EM7, a dissimilar dual-triode orginally intended for TV use; the bandwidth is also excellent. And, since they're TV tubes, the audiophile nuts haven't yet started hoarding them. I got a bunch for next to nothing.
steve wrote:The output transformer is a custom Electra-Print. This company was fast and easy to deal with, and I recommend them.
Wow. I just checked out their website ... $145 for a SE OPT that's flat out to 55kHz? I'll have to check out their stuff. Sounds pretty reasonably priced.
steve wrote:The power supply is a C-L-C Pi filter. Haven't decided whether to use diodes or a couple of 3A3 rectifier tubes I have laying around. I'll probably use diodes on the breadboard until I'm sure everything is working, then try it with the rectifier tubes. I'm a little concerned about the complexity I'll be adding with another filament transformer dust for the PSU, and these tubes were originally intended for high-voltage clipper use, and I don't know how well they'll work as PSU rectifiers.
I'm starting to like diodes rather than tube rectifiers, for a number of reasons:
1. As you mentioned, there's one less filament to power.
2. Reliability.
3. Less noise.
4. The bass is audibly better.
Granted, people talk about switching noise and "hash" with SS diodes, but the choke can pretty reliably kill that. Putting 0.1uF/1000V caps across the diodes helps, too.
How did you arrive at that choice of tubes, by the way? Were they just things you had laying around the workbench? Or did you choose them on their specs?