Tube DIY Asylum: Gary Pimm CCS transconductance amp by John Swenson

I just finished up another one of my weird amps. This is actually the simplest one, and its SE not a differential.

I was looking at the schematic of Gary's CCS in regards to the CF threads that have been around here in recent days, Using one as the CCS on a CF and how to do it so that it cancels out all distortions caused by the load (thats a different amp then this, I haven't built it yet). While I was working on the above I realized that the CCS with a pentode on top would be a perfect transconductance amp all on its own. All you would have to do is drive the CCS with the audio signal instead of a constant voltage. I talked to Gary about it and he cameup with a slight variant of my design (which was even simpler). The audio signal is cap coupled onto the gate of the bottom IRF820 and C2 is ommitted, thats it. This weekend I built it.

Its built on the Bottlehead eXcite, I used all the iron as is and rewired the 9 pin socket for the el84 and added the CCS (no big tube in the 4 pin socket). I ran the PS as LC since the PS voltage was way too high with a CLC. Being just LC there is still a significant amount of ripple on the B+, but the circuit seems to be immune to it, the PSRR seems quite high. This gives about 2.4 watts max, about right for an SE el84 circuit. The output circuit is parafeed with a MQ BCP-15 choke and EXO-045 OPT.

One very interesting aspect of this circuit is that you can adjust the Gm by changing the current set resistor (R1 in Gary's schematic). You can easily make the GM high enough that you don't need a driver stage at all, the whole thing has one tube per channel (and the 4 MOSFETS in the CCS).

How does it sound, WOW! Its definately has a family resemblance to the diff el84 high current amp, but even cleaner, more open, resolves even more detail, and yet is smoother, the midrange harmonic structure is just amazing, and it even has deeper bass and better dynamics. The PP amps are supposed to have better dynamics than SE, but not this one.

Another interesting thing about this design is that the linearity of the tube is almost irrelavent. This ought to scale very easily to much higher output pentodes, many of which would sound horrible in normal circuits, but should sound great with this. Next on my list is a EL519 based version, but I'll need new plate chokes and OPTs for that.

If anyone wants to try this, remember its a high output impedance amp, designed specifically to drive speakers such as Lowthers. For most speakers it would not be a good match.

BTW the 22K screen resistor (R3) Gary mentions worked perfectly, the screen voltage was just perfect so the bias was right where it needed to be for 40mA.

John S.

Gary Pimm Self Bias CCS (Open in New Window)