Distortion profile versus Foreplay models?

[greg788]

How does the Moreplay compare in overall THD, type of distortion (predominate 2nd and/or 3rd harmonic, etc.) and profile (flat, rising at high frequencies, et.) versus the previous Foreplay models? I'm asking in general terms knowing that each tube contributes its own signature, esp. in such a simple zero NFB triode connected circuit

PS, my question has nothing to do with the "low distortion movement" I'm seeing through audioscience.com. It's more from measurements I saw on the CJ PV10, which is also a simple triode linestage. And comparisons of the Moreplay versus a CJ PF1, which is solid state

[Paul Birkeland]

I didn't have anything quite like this ready, so I had to spend some time putting something together and thinking about what made sense for testing.  Ultimately I figured that grabbing an FFT shot at about 2V of output made sense, but bear in mind that I didn't make any effort for noise reduction as this is only really about THD.

Anyway here's what the THD looks like in the Moreplay at 1kHz and 2V.  There just isn't much beyond that 2nd harmonic and I don't even really see the 3rd pop up above the noise floor until over 2.5V of output from the preamp.  The 60Hz/120Hz noise that is seen is stuff that we generally deal with in planned future upgrades, and we are in board design right now for one of those. 

I also increased the loading on the Moreplay from 10K up to 100K, and there is a commensurate reduction in THD accordingly.  C4S loads will also bump this down in the future. 

What's somewhat unique about the Moreplay is that the THD level is not dependent on the level control (unless you go to mute), so the distortion percentage and profile at 300mV of output is identical to what you see at 4.5V of output, and this is quite a bit different from something like the Foreplay that has a THD increase with level, or a lot of vintage preamps that are going to have a curvy relationship with output level and THD.

[Paul Birkeland]

Here are some measurements from a non-C4S Foreplay II that has LED biasing but is otherwise relatively stock.  What I found rather odd is that the THD performance is a bit better into 10K rather than 100K, which is exactly the opposite of what I would have predicted!

[greg788]

Paul, is THD or THD+N more valid here? And this is with the stock Russian tubes I assume? What contribution do you estimate to THD are just from the tubes?

[Paul Birkeland]

Getting good THD+N would take more time, as I'd need to more carefully dress what I'm using for the inputs and outputs to the FFT, as well as paying a bit more attention to what is earthed where. 

The tubes in the preamp are some old stock 6V6s.

[greg788]

If I wanted to reduce the distortion by white wiring a feedback loop around the 6V6, how would I do that? I assume an RC network would be needed as well? I can go down to unity gain no problem. I think this is 12dB of gain so cutting 6dB or even 12dB, no problem

[Paul Joppa]

Gain is 9dB, so a 9-dB inline attenuator at the input will reduce the signal level in the circuit by 9dB. That in turn reduces the second harmonic by 9dB, third harmonic by 18dB, 4th by 27dB, etc. This is the easiest way to reduce distortion. Feedback is not necessary.

Introducing feedback would be a re-design of the circuit; either an anode follower or a cathode follower would be candidates. The anode follower is complicated by the balance control topology; the cathode follower would need a bipolar power supply.

[elcraigo]

I have these measurements from a Foreplay I. Amazing I still have these.

December 12, 2000 - Foreplay Measurements by Craig Lewis (elcraigo)
All measurements are unweighted into 100K ohm load.
tubes: GE 5963 (came with kit), Kit stock except magnet wire.

Equipment used:  "Audio Precision System I"  with DSP (digital signal processing) option installed.

Input      Output      Gain   THD+N% @1000Hz
1Vrms      1V           0dB    0.2%
1Vrms      10V (max volume)    20dB    0.9%
2.25Vrms   22.6V (max volume)     20dB    5.1%
Signal to Noise (s/n) @ 1kHz = -79dBr (preamp gain = 0dB, 1Vrms input)
Frequency Response (-0.2dB, +0.07dB, 10 Hz to 50 kHz; -0.05dB @10Hz,  -0.16@50kHz)
   Input: 1Vrms
Output: 1Vrms
Reference: 0dBr @1kHz
Phase Input to Output (0Hz to 50kHz)  -196deg to -177deg (yea it inverts or -180 deg)

[Paul Birkeland]

If I wanted to reduce the distortion by white wiring a feedback loop around the 6V6, how would I do that?

Make it into a unity gain cathode follower.

[greg788]

Paul, two tube rolling questions since I've been doing a lot of it lately

1. if you're gentle and slow with tube extraction from the bases, are there risks to loosening the pins in the bases over time, or do the ones you use stay good for a long time? I notice you can't tweak the pins with these bases

2. based on the gain stage and voltage the tubes are run at, is it conservative and promoting long tube life? Or does a triode gain stage stress tubes significantly even at the lowish voltage they're run at? In other words, should I expect the 3 years rule of thumb from new, or shorter or longer on average for tube life?

[Paul Birkeland]

You can tighten the pins on these bases with a very, very small flat head screwdriver.  If you do enough tube rolling, eventually you will wear out the sockets. 

The 6V6 in the Moreplay runs at about 1.6W of dissipation and is rated for 9W in triode mode.  I suspect from a set of new old stock 6V6s that you should see 5-20 years of tube life with regular use.  I have no idea how long the new Russian and Chinese tubes last though. 

[greg788]

What about 6K6GT?

Also, delving into the philosophic / subjective realm, perhaps, how do the different standard topologies (triode, cathode follower, etc.) sound differently from each other?

[Paul Birkeland]

I haven't tested the 6K6GT.

Triode refers to three signal handling elements in the tube (cathode/filament, grid, and plate).  A cathode follower trades away the available voltage gain in favor of decreased output impedance.

[greg788]

Paul, I'm familiar with the different topologies. My question is, in your experience, how do they sound differently? All else being as equal as possible (same tube, etc.), can one infer a cathode follower sounds leaner, more "SS" than triode by virtue of lower output impedance and distortion?

[Paul Birkeland]

There isn't "cathode follower vs. triode".  A circuit with a single triode can be designed as a cathode follower, grounded grid amp, inverted triode, grounded cathode amp, and several triodes can be wired into a far wider array of topologies (like the older Foreplay preamps).

If you want something that sounds a lot like a solid state amp, cascade two 12AX7 grounded cathode gain stages with 30dB of feedback around the whole enchilada and you'll basically be there, albeit without much drive current capability.