Cathode bypass

[Dr. Toobz]

If one were to remove the 1000uF bypass capacitor under the cathode of each tube, my understanding is that gain would be reduced, but this would also be accompanied by a reduction in distortion. How much gain is lost when using an un-bypassed cathode resistor, and is the reduction in distortion enough to be measurable (or worthwhile)? Some of my 3S4 tubes seem strange in the bass region, sort of "rolled off" or wooly (this is without any Specos to load down the circuit - just as a preamp). I've wondered if taking the bypass caps out of the circuit might help stabilize the low frequencies a bit more, at the expense of some gain. I only need about 3dB of gain anyway, and I think the stock circuit gives 7dB, IIRC.

[Paul Joppa]

Removing the bypass capacitor will more than double the equivalent plate impedance. I have some doubts about whether it will reduce the distortion in the stock circuit, because the load impedance is pretty low to begin with. But it's an easy experiment - try it and tell us what you hear!

The bypass cap does not need to be large - I think anything over 20uF is enough. You might try a better cap, even a filme cap, and see if that makes the kinf of change you are looking for.

You did say "...SOME of my 3S4 tubes..." so it may be an artifact of microphonics. Just a thought.

[Dr. Toobz]

Actually, I have - some very interesting things to report as a result.

First, the Quickie with either chokes or CCS gives me more gain than necessary. In the case of the chokes, the distortion signature is gain dependent - so turning the pot up not only raised gain, but also raised 2H distortion, leading to a warmer, fatter sound with a collapsed soundstage. I like the sound of my S.E.X. amp so much that I have traditionally preferred going right from the source into the amp, as any (tubed) preamp I stick in between seems to raise the noisefloor, and diminish tiny micro-details in the treble. I want a little bit of gain with no additional "flavor." What I ended up doing is keeping the chokes, as they seem to be able to swing a bigger voltage than the CCS, leading to less noticeable clipping on really loud passages. I then removed the bypass caps, one at a time, as to compare channels. Interestingly, the bypass cap makes the amp have a bit more gain (of course) but also changes the sound signature. I found the bypass to make the amp sound warmer and fatter, especially in the mid-bass to midrange regions, but a bit less clear in the HF region. The amp sounds flatter and more "boring" with just the resistor, presumably due to the additional negative feedback at the cathode. I actually prefer this.

Another experiment was to add some local feedback around the tubes, to decrease noise, and decrease the idiosyncratic nature of particular tubes. 3S4's seem to really vary between example, and I don't like that - I'd rather just plug in a tube and have it sound the same every time. Putting a little bit of plate-to-grid feedback into the mix really cleaned up the amp and is letting through sounds like distant reverb and very tiny details in some recordings - opposite of what some might expect. I suspect this is due to lowering the distortion a bit, as I've found that the typical 2H "tube signature" tends to thicken things up and makes dominant instruments in a recording "stand out" at the expense of ones deeper in the mix. The loop was accomplished via a 1uF film cap in series with a 100k resistor. This seems to give about 3dB of feedback, just enough to flatten things out without making the sound sterile. If anything, I am now hearing the "DHT magic" people talk about. It's almost as though something "lively" or "real" is added to strings and other acoustic instruments, and I've never heard my IDHT tubes do this. I now feel that the preamp in the mix is better than just going straight into the S.E.X. amp - it not only seems to keep the entire bandwidth of my recordings the same, but also adds a little bit of something I can't quite quantify - "magic," if you will? I needed to clean the sound up first to hear it, however. Some of the tubes I have, using the stock circuit, sound WAY too warm and rolled-off for my tastes.

As always, YMMV - keeping in mind that I grew up with solid-state, not tubes! As a result, I suspect that I like a cleaner, flatter, brighter sound than your typical tube guy, but to each their own.

[cpaul]

Interesting, and thanks for your followup, Dr Toobz.  How long have you had this mod operating?  I'd be interested to hear a report back in a few months about how you like it.  I've experienced liking mods at first, then changing my mind later.  Anyway, thanks for your experiments.

[Dr. Toobz]

I couldn't resist more experimenting, and it now looks like I'm settling on a middle-ground approach with my Quickie. I just bought one of those HRT Music Streamer II usb-audio things.....and it sounded a little too sharp and treble-heavy for my tastes though the Quickie (actually, the HRT is a bit bright on anything I've tried, but I digress). I figured that dialing down the NFB at the cathode and/or plate would fatten up the sound a bit and perhaps tame the brightness.

I tried a few different voicing options: no local feedback and no cathode bypass caps (less sharp/crisp treble, but not as big of a soundstage), cathode bypass caps of lesser capacitance and better quality than stock, e.g., 100uF Panasonic low-ESR electrolytics, the ones that come with the Crack (warmer sound, more bass, more "tubey," yet tighter bass than the 1000uF stock. Highest treble lacks some sparkle/detail). Finally, I tried the aforementioned 100uF cathode bypass with slight plate-to-grid feedback. The latter seems to be the best compromise: I end up with a very clean, spacious sound with no sharpness and just a hint of warmth and presence in the mids. There is enough "sparkle" in the highest notes to keep me happy, though the treble is definitely most transparent with just a cathode resistor. The bypass cap controls the voicing of the amp much more than I would have expected, and it is interesting that I seem to detect a reduction in the highest treble anytime there's a cap in the path. With the bypass, I would surmise that I'm getting slightly higher levels of 2H distortion than I would when just having a cathode resistor, but this seems to balance the squeaky-cleaness of the local feedback so the amp doesn't end up sounding too solid-state.

The best part of DIY has been the ability to learn about what makes an amp sound like it does, as well as retain control over these characteristics as to get the most pleasing sound from my particular speakers, headphones, and sources.