Thoughts on Vari-Mu design with S.E. output stage

So after running many tests including observation of harmonic content on a recent vari-mu build (a-la AM864) I started wondering how difficult / harmonically rich / pointless a vari-mu tube comp would be with a single-ended output stage. It would be designed to be rich in 2nd harmonic (when pushed?), compared to all the Push-Pull output stages where all the even harmonics cancel out. Even harmonics are musically related to the fundamental, whereas odd ones are dissonant, yes? Is this a big deal? Maybe not... I know there are designs out there with SE output such as the U73 and R&S comps, but replicating these designs requires rather special tubes, and the circuits are somewhat complex with very specific trafos. After-all, they were trying to get rid of this harmonic content which I am after. Yes yes, I know these tube comps do sound awesome and I love them, but maybe there is some way we can design something that favours particular rich harmonic sounds rather than us replicating old designs where they were trying to minimise THD...

Attached is a rough sketch from Eagle showing one idea. Please ignore any part names such as the trafos. I have put cathode followers after the vari-mu stage as with my bench tests it proved that I would need a very high-Z interstage trafo to link the balanced vari-mu stage to the SE output stage. Quality High-Z means pricey trafos in my mind, compared to nice cheap EDCORS which only go as high as 15K ohm. One more tube plus EDCOR iron can be cheaper than Sowter iron . Hence the CF's. I DC-coupled them because I wanted to get the sketch done quick. This won't work of course in reality as the anode voltages move A LOT on that first stage with GR. But would that interstage trafo arrangement work? DC current would cancel. I added R6 as these triodes (I'm thinking 12AT7) will surely need more DC-R than the edcor can give, considering that a 15K:15K EDCOR has about 300 ohms DC total on pri & sec.

I re-jigged the AM864 vari-mu on the bench with the output triodes (6SN7) paralleled, choke-fed, and output trafo cap-coupled from the linked anodes. Sure enough, testing this stage on it's own I can get strong second harmonic (plus others of course). I also tried with resistor instead of choke, and this still gave me enough headroom. In fact, it should make the stage overload easier. Using a variable pot H-Pad on the line out seems to increase 2nd & 3rd harmonic as it loads the output more when you turn it down. make-a-sense? I did try some interstage trafos between the 6SK7's and the SE O/P stage but Freq. response was poor due to too much loading on the 6SK7's. I was however getting a dominant 2nd harmonic...

I am hoping that some of you wise folk can shed some light on or debunk some/all of the above ponderings, rather than me go into a rabbit-hole of pointless harmonic chasing until my dog can no longer stand the sound of the sig-gen

There are lesser known RCA Photophone comps that are standard PP gain reduction with SE outputs as well.  The purpose in those designs relate to lower ratios and decreased power consumption for remote use with batteries. 

Thanks for your input Doug. Hmm interesting! Since reading your reply I have been searching for schematics but all I can find are power amps from the photophone era, and the 86a (may be from that era?), which as we know has a PP output stage. If you have any model numbers for me to go on that would be a big help.

Another thought... Why do the older limiting amps have such a high power output stage? Usually a pair of 6V6 producing up to about 10 watts! Doug, I think I read a comment from you in one thread (which I fail to find right now) saying that one reason for this hefty O/P stage was to give enough current drive to the sidechain to get fast enough attack. Can this be correct? I always assumed that the prime reason for the hefty O/P drive was in order to drive long, capacitive telephone lines all the way to the TX site, possibly many miles away from the Studio. I am lacking in any knowledge of Studio-to-TX linking from this era though.

I need to keep this in mind when trying an SE output stage, as I will need enough current/swing to charge that sidechain cap fast. Would a Cathode follower help much in the sidechain, seeing as I would only have one AC signal to take from the SE amp.?

I am just realising that maybe this thread should be in "The Drawing Board" rather than here. If any moderators agree, then please do move it over to there . Maybe we will get more discussion going, which I would be grateful of.

Another part I am looking at is the gain control stage and how I should implement it. Right now I am rather bored with typical pentode/triode vari-mu's, and am leaning towards either a pentagrid or even bridge topology. It seems the pentagrid died out pretty quick though, and I am wondering why. Expensive tubes maybe? Anyone with any thoughts on either, I would be glad to hear! Currently digesting DaveP's thread on the Bridge compressor. Brilliant data as always!

vinyvamos said:

Right now I am rather bored with typical pentode/triode vari-mu's, and am leaning towards either a pentagrid or even bridge topology. It seems the pentagrid died out pretty quick though, and I am wondering why.

Click to expand...

From what I understand, which is very little, a lot of the decisions stem from application/costs scenarios like many other things. One video I watched explained how the pentode took over where the tetrode fell short at low plate voltages when being used during a time that portability was the new market.....

Here's a video getting into the pentagrid a bit..... Same channel I heard the above info..... Pretty good stuff....

https://www.youtube.com/watch?v=SXDxnwGspck