new stuff from jc morrison

last weekend was etf weekend (links to collections of pics) and I had the pleasure of traveling to the north-of-france location together with jc morrison. my friend Daan van Egmond kindly provided the amsterdam–etf–amsterdam ride.

jc clearly had set himself a goal for his first-ever etf visit: spread the circuits he has been working on for the last year as far and wide as possible. he has recently returned to designing hi-fi tube amps after a long hiatus and yes, what follows here are in principle cornerstones for his work for silbatone. he says anyone is welcome to implement these type of circuits into their own amps, so that prior art exists and prevent some shifty type patenting these.

enthusiastically sharing circuits with anyone at etf, jc gave me these four sheets for immediate publication on this blog. so here we go, the schematics are in a funky spice form, but I think we’ll manage.

‘why does it always have to be B+?’ jc says. so here we have one B- supply and 2 direct coupled stages, both with full-µ gain [edit: depending on transformer loading]. important to get your head around: the input triode (U1) is not a kathode follower, it is an inverted gain stage. the reference for this tube is the junction of the input transformer and D1 and as far as this tube is concerned, the rest of the amp is moving up or down. this means it sees current sink I2 in series with the power supply on top of its plate and it puts out its full µ of gain:

for all these circuits jc is extremely dowhatchalike with the biassing of both tubes—(tube) diodes, resistor (+ cap)—and for the current sink—mosfet, (cascoded) tubes, choke—use what you trust. during his talk at the etf jc sketched a variant of this circuit using a choke.

next: ‘why fight city hall?’ jc says. so let’s use a pentode for what is good at and maximise transconductance. the result is a gain of 750 ‘without even trying,’ think phone front-end:

jc’s sketch at the back of this printout shows the principle. a variable current source (pentode) pushes against a fixed one (I1) and the signal current is converted by R1 to (a high!) voltage gain:

jc likes that the screen is simply hooked up to ground (although the burden is moved to the 24V power supply), I like that the signal creating current loop is extremely short. I think it could avoid going through any power supply if the 220k resistor was hooked up to the plate of the pentode instead of ground.

next is cascoding and I must fully admit that if cascoding is ‘build your ow pentode,’ that I do not understand why this one does not need a resistor for I/V conversion:

finally—even more cunning—is folded cascoding. the fet is a p-channel device:

I am pretty sure that here that jc likes to tune the supply voltage of the fet to further linearise the whole shebang. speaking of, jc says he build all of these circuits in many variants and they checked out fine.

stay tuned for jc’s silbatone DA-100 power amp design and an application of the same architecture for the GM70.