Spud Farmin' Pt. Deux

[valveitude]

Ok, after coming to (or rather, being pushed to) the conclusion that I would need some kind of follower arrangement if I wanted a manageable input capacitance on the “Tater Patch”, I reviewed John Broskie’s TubeCad Journal for some ideas. I had printed the pdf on “Complimentary Distortion Compensation” earlier, and decided to configure my circuit around the concept to see how it would simulate. Well, the distortion was incredibly low, but, it did this by squashing the 2nd harmonic, leaving the higher harmonics. In other words the harmonic spectrum resembled a P-P. If I wanted that, I would be focusing on a P-P. I then ran a simulation of just the “follower” output stage…yowser!! The distortion wasn’t as low, and, the higher harmonics were still present, but it did have a nice regression with the bulk of the distortion being 2nd and 3rd harmonics. So I put the voltage amp stage back, but this time I didn’t “match” the load, instead I presented it with an easier load and came up with this circuit…



First off, I’m not convinced the 6S45Pi is the best choice for the voltage amp, I still have a pretty high input capacitance. I would welcome some suggestions here. Because the follower output doesn’t suffer from the Miller effect it is easy to drive, so I should have a lot of options here. That said, the circuit “seems” to have some pretty stellar performance.
The 1 Watt and4 watt distortion is very low for a SE amp, and the1 Watt and4 Watt FFT’s show that there are higher order harmonics, but they are so low I believe they would be buried in the noise floor.

With the values shown, the –3db point is 80hz, a little lower than my target of 100hz but it’s doable. With a source impedance of 5k,its down 1.5dB at 100khz .

This is the part were the “Frontal Lobe Brigade” is invited to tell me what I’m overlooking this time. I was forced…forced I tell you!!…to reconsider my approach once, I wouldn’t be surprised if it happens again.

Let the critique begin…..

Casey

[SY]

I'll let EC8010 throw bricks when he's done with his afternoon nap and pablum (he does very well for a codger his age, and sometimes can even remember his name with only a bit of prompting). In the meantime, let me ask what the assumed output transformer turns ratio is?

I'd seriously think about a 6SL7 or similar for the first stage.

[rdf]

Is this intended to power the top end of a subwoofered system (~-20 dB at 20 Hz)? It also appears there are two staggered LF roll-off points, ~20 Hz and ~100 Hz. Could be intentional as well, just sayin'.

[valveitude]

I'll let EC8010 throw bricks when he's done with his afternoon nap and pablum (he does very well for a codger his age, and sometimes can even remember his name with only a bit of prompting).

Please God...Don't Disturb The Nap!!

In the meantime, let me ask what the assumed output transformer turns ratio is?

57.74:1 turns ratio, 1667ohm:1/2 ohm impedance ratio. I need ~130v p-p on the grids of the output stage for 4 watts (2 volts p-p) at the secondary. Currently the 6C45Pi voltage amp delivers this with a input of ~2.8 p-p.

I'd seriously think about a 6SL7 or similar for the first stage.

OkeeDokey...I'll look for a model to play with tonight.

Is this intended to power the top end of a subwoofered system (~-20 dB at 20 Hz)?

Yep I am building this to power a dual 36" ribbon that I hope x-over around 100hz...a blue skies attempt.

It also appears there are two staggered LF roll-off points, ~20 Hz and ~100 Hz. Could be intentional as well, just sayin'

Yep again..I am shooting for a gentle slope/minimum phase shift initially for a smooth transition, and a steep second leg to keep the really low stuff from damaging the ribbon.

[SY]

57.74:1 turns ratio, 1667ohm:1/2 ohm impedance ratio.

That's going to be a fun component to source!

[valveitude]

That's going to be a fun component to source!

Yes it would be. I plan on winding the output and choke iron. I've already modeled the output with yvesm's OPT program.

[PRR]

> ask what the assumed output transformer turns ratio is?

It's on there, in SPICE notation. 7.2H:2.16mH. The inductance ratio is 3,333:1, and (assuming K is nearly 1) that's also the impedance ratio. Given 0.5 ohm load, looks like 1,667 on the primary, or 5K per tube. Reasonable.

Is that SPICE really allowing K to be 1? That's ideal, and unreal. 0.999 to 0.999,9 are more likely with very good winding technique. Power transformers can be below 0.99.

> need some kind of follower arrangement if I wanted a manageable input capacitance

Well, you moved from 3 medium-gain grids to one high-gain grid. Even without looking up that tube, I can guess that your input capacitance dropped to half?

Go back to the plate-loaded output stage. It didn't suck bad except you seem unhappy about the input capacitance. Then use that same 6c45pi (or whatever) as a cathode follower input. Input capacitance will be like 1/30th of the original.

[Yvesm]

Is that SPICE really allowing K to be 1? That's ideal, and unreal. 0.999 to 0.999,9 are more likely with very good winding technique. Power transformers can be below 0.99.

Or you may include the computed values of Primary and Secondary R as long as the LL and the parasitic cap.

Results would be of great interrest for me

Oh, and among other (know and unknow) bugs, you must multply the lenght of wire by the number of paralelled ones, sorry

Upgrades soon to come.

Yves.

[valveitude]

PRR…

Well, you moved from 3 medium-gain grids to one high-gain grid. Even without looking up that tube, I can guess that your input capacitance dropped to half?

Well, by a third actually. Not a bunch, but enough to get a reasonably easy load for the proceeding stage.

Go back to the plate-loaded output stage.

Maybe I will… maybe I won’t (blows raspberry ).

It didn't suck bad ...

High praise indeed.

...except you seem unhappy about the input capacitance. Then use that same 6c45pi (or whatever) as a cathode follower input. Input capacitance will be like 1/30th of the original.

That was my plan, then I got to thinking (always a questionable endeavor) , if I am to succumb to the “evil” follower in the signal chain, why not see how it simulates as the output, rather than the input. Well, over a order of magnitude lower distortion (from 1.1% to .1%) at full power, that’s what. Just for giggles, I lowered the load from .5 ohms to .25 ohms last night to get a handle on the dampning, and found the distortion dropped to around .05% at full power…looks like I’ll be playing with the trannie design to optimize the load. I , like I suspect a lot of people, don’t have a real handle on distortion vs. sound quality, but my “theory” is this… Higher order distortion bad, lower order distortion not so much. All things being equal , lower distortion overall is better, as long as the harmonic structure of the distortion is regressive, ie, each successive harmonic is lower than the previous. I know this is a very simplistic view, but hey, the view fits the viewer. I suspect I may be falling into the “numbers’ trap here, after all, the huge majority of SE amps I’ve seen are plate loaded…we’ll see.

Yvesm…

Or you may include the computed values of Primary and Secondary R as long as the LL and the parasitic cap.

Results would be of great interrest for me

Will do…as soon as I settle on a final target for my transformer (see above).

Oh, and among other (know and unknow) bugs, you must multply the lenght of wire by the number of paralelled ones, sorry

Let me get this straight. You spend countless hours developing a comprehensive, easy to use, and reasonably accurate transformer design program. You then give it away free to anyone who wants it knowing it’s not perfect…the nerve .

I am extremely grateful to have this software..warts and all, and I look forward to any future release..thank you.

Casey