Cathode Bypass Effecting Tone Stack Bass Control

[ToneHenge]

Hi all,

I'm going down the endless tweaking wormhole on a design. I thought the amp was sound a little flat with all tone controls set at noon, and decided to boost treble a bit to make things sound more lively. A convenient approach is adding a bypass cap to the cathode of the triode gain stage that precedes the tone stack. That gave the desired effect but now the bass control seems to behave more like a gain knob than an EQ knob. The gain stage is V2a below:


I put a large .1uF orange drop in as the cathode bypass cap on V2a. Going off of the cathode bypass calculator on ampbooks.com, I wanted to place the gain boost more towards the treble frequencies at 1kHz+. My guess is that I've affected the way that the tone stack is being loaded by the preceding gain stage but I'm not entirely sure.

Any ideas on what's going on here?

Edit: Took out the cathode resistor bypass cap, and this bass control still sounds more like a volume knob to my ears. Maybe that's how Fender tone stacks are? I might try something like a baxandall next time around.

[loudthud]

The blackface tone stack is pretty bass heavy to begin with. Many people run with the bass control on 2. To get away from this you need some HF lift in a couple of places. Try a little bass cut by reducing one of the coupling caps. That .1 driving the mid network or the .022 driving the effects loop.

Are you really using 15K for the cathode of V1a? That's going to choke off the cathode follower V1b. The CF can't deal with too high a plate Voltage on the preceding stage. Not much signal can get through and what does will have ugly distortion.

[Chuck H]

I don't see how the cathode bypass cap should affect the tone stack at all. But then you mention that the bass control seems like a volume control without it too. Though maybe not as much. LT is right about that tone stack being bass heavy, but you should still be able to perceive the difference between volume and 'bass volume'. So I'm wondering if something might not be mis wired? Or maybe something in the voicing of the rest of the circuit is having some anomalous affect that makes the standard type tone stack discordant?

I simulated your "Mid Control" in SPICE and it's pretty weird. The actual center frequency can be shifted between about 50Hz and 325Hz. And I'd call that bass frequencies. So it's a bass control? The available cut is about 15dB at this center frequency though at some extreme setting it can get pretty flat with only a small bass cut. At moderate settings there seems to be cut in the bass or upper bass/low mid frequencies.

This could be anomalous with the tone stack later since you have the standard low mid cut that comes with the circuit AND that tone stack, known to be bass heavy, raises and lowers from a center of 25Hz with the bass control.

Basically, this seems like strange voicing and you might be hearing some odd behavior where the bass control seems to carry most of the loudness of the amp, if not a wide frequency range. I wouldn't expect this though because your "Mid Control" has almost no affect on frequencies above 1k in any position so the amp should be quite bright but for a couple of LP filters that look to only affect very high treble.

I think you should definitely check your wiring for the tone stack and reassess the function of the "Mid Control" circuit. Where did you get that circuit anyway?

[ToneHenge]

loudthud,

Yup I've got 15K on the cathode of V1a. Basically just wanted to drop the gain at the first stage and was too lazy to wire up an interstage voltage divider. I didn't realize that would cause cause issues with the CF that follows though. I was hoping to get a kind of compressed sound by cascading these triode stages with lower gain. Going for a more mid-gain sound at maximum rotation of the gain knob. What change would you recommend?

Chuck,
I'll quadruple check my wiring but I think it's right. Might be worth checking if any of the components are failing somehow, but I'd be surprised. Good catch with the mid control! Those were old values that I had been playing with to see how it sounds. I've since change all the cap values to 2.2nF with moves the mid scoop to a range of ~500-1000Hz. I actually pulled this circuit from Merlin's preamp book! Pretty happy with how it sounds now that I've moved that mid scoop up a bit, and switched the shift pot to a linear taper. I was looking for something that would act like a parametric eq but didn't want to deal with adding any op-amps into the mix.

[loudthud]

I would use 1K or 820 on the V1a cathode. As long as B+ is 300 or so, you won't run out of headroom on a first stage. You can make a Voltage divider by sticking a resistor above the Volume pot and a small cap in parallel will lift the highs. Try 1 Meg and 270pF. For my money, unbypassed cathode resistors take away the even harmonic distortion that tubes are know for. I don't even like bypass caps less than 1uF. If you want or need to boost highs, use a Voltage divider with a small cap.

[SoulFetish]

I don't see how the cathode bypass cap should affect the tone stack at all. But then you mention that the bass control seems like a volume control without it too.

Well, partial cathode bypass can raise the output impedance at lower frequencies by quite a bit. This can cause bass frequencies to become heavily loaded if the bass network has a low input impedance.

[Chuck H]

Another downside to the unbypassed first gain stage is noise. Not just filament noise either, but actual emission noise. I did some A/B testing just to hear it. The amp I tested on didn't exhibit any filament hum. Though I had expected a little and that's what I was testing for it wasn't there. But I did hear a sort of low/mid frequency swirly hash that would change increase or decrease in volume relative to the value of the bypass cap. Full bypass equaled no hash. I can only assume it was emission noise I was hearing since I don't really have a way to test for that. But, considering the usual potential for hum anyway and my experience with this other noise I always design with a fully bypassed first gain stage now. All the uber gain partial bypass first gain stage designs be damned.

Like LT said, you can create a useful voltage division with a single resistor behind the load of the volume pot.

[Chuck H]

Well, partial cathode bypass can raise the output impedance at lower frequencies by quite a bit. This can cause bass frequencies to become heavily loaded if the bass network has a low input impedance.

Would you say that partial bypassing actually raises the low frequency impedance? Or wouldn't it decrease the high frequency impedance. I don't think that's exactly the same thing.

[SoulFetish]

Would you say that partial bypassing actually raises the low frequency impedance? Or wouldn't it decrease the high frequency impedance. I don't think that's exactly the same thing.

Well, interesting question. But it depends on the perspective of where you’re starting from. If you were to say fully bypassed vs partially, then I think my description is accurate and acceptable. But unbypassed vs partially bypassed, then your description is more appropriate. IMO.
You dig?

[Chuck H]

You dig?

Down like a clown Charlie Brown.

In this case I was considering the schematic in the opening post. And even at that I think your point remains the same. If the impedance is higher in the LF it's going to be more susceptible to the tone stack impedance either way I suppose. But I didn't peg the typical TMB stack in guitar amps as a low impedance input. Even "relatively". One reason I've never understood the common lore that the cathode follower in the 5f6a/Marshall design was to prevent loading by the tone stack. There's only a savings of a couple dB. You could do a lot better with an additional gain stage Just a happy accident that the CF makes a desirable tonal contribution. But now I'm getting sideways.

[Helmholtz]

Well, partial cathode bypass can raise the output impedance at lower frequencies by quite a bit. This can cause bass frequencies to become heavily loaded if the bass network has a low input impedance.

Why partial cathode bypass?

A full bypass cap will lower output impedance above the corner frequency by less than a factor 2, compared to unbypassed. Partial bypassing (split cathode resistor) will have even less effect.

The 0.1µF bypass cap as mentioned by the OP will not noticeably change low frequency output impedance.

[SoulFetish]

I just ran some FMB frequency plots. Its not much more lossy when driving with a 63k source impedance vs 38k. So “heavily load” is a bit overstated. In this case I would say the boosted gain caused by bypassing is probably the dominant influence affecting the sound

[Chuck H]

I might easily have considered that since the typical TMB stack presents something like a 100k to 200k (depending on knob settings, etc. impedance at it's pass frequencies that a difference of 38k to 68k in the input signal could be significant. It's really only because I've observed the difference in cad programs many times that I know it's not.

[ToneHenge]

Okay pretty interesting discussion, still trying to grasp some of it.

FWIW, I'm using twin reverb iron (the the PS, PA, and PI are from a blackface twin reverb). So the voltage for the PS node supplying those two stages is ~400VDC. I think I'll give your approach a try with dropping the V1a cathode resistor and place a divider before V2a, but I am confused. I don't really hear anything that sounds horribly wrong with this arrangement - want to make sure I understand. And can you guys spell it out for me on how to leverage the gain pot and an additional resistor to create an interstage divider. Would this resistor be placed from the wiper to ground? I suppose that parallel resistance would pull down the resistance of the wiper-to-ground side of the pot.

Another thing I've been trying to figure out (might belong in a separate thread): What is the design approach to ensuring that the phase inverter is hit with a sufficiently large signal? I ask because this build just sounds little quieter than I'd expect with the gain control set low (12 o'clock for example) - not quite that deafening twin reverb loudness. I've checked the bias on the power amp tubes and measure ~47mA on each 6L6GC. I'm wondering if I'm not hitting the phase inverter hard enough. I tried to make that series effects loop as close to unity(if not a bit higher) as possible, so I don't think that's the issue. I suppose the most likely cause is what I'm doing in V1 and V2

Thanks again for the help!

Here's updated schematic for clarity's sake

[Helmholtz]

You are using a very low value (270R) NFB series resistor. Together with the 100R shunt resistor this means a lot of NFB and consequently low power amp gain. Try increasing to 820R (as used in TRs) or higher and take the feedback signal from the 4 Ohm output if available. This will increase power amp gain (and stability margin). As a result less PI drive is required for full power.