Adding a bypass capacitor... effect on the NFB resistor?

[Enzo]

I get it, but it pretty much 86's the NFB.


Even if you lack clip wires, a cap has wire leads. Two dots of solder, and the cap is connected to whatever. It doesn't have to be sturdy, it doesn;t have to last for the ages, it has to sit there for a minute or two while you listen to the result.

[Leo_Gnardo]

Even if you lack clip wires, a cap has wire leads. Two dots of solder, and the cap is connected to whatever. It doesn't have to be sturdy, it doesn;t have to last for the ages, it has to sit there for a minute or two while you listen to the result.

Agreed! and tacking on a cap, or whatever other component, in that way relieves the possibility of setting up an unexpected reaction - ultrasonic oscillation I'm thinking - between the signal riding on the clip leads and whatever they may come near to in the amp.

[Enzo]

Sometimes clip wires introduce hum. I have enough discipline to ignore the hum and just listen to the change from the component. But that does confuse a lot of guys. Tack soldering does not add the wire lengths, so we see less of that. WHich I think is what you said.

[galaxiex]

I've been known to tack solder caps in various places, just to see what happens...
observing voltage ratings and polarity of course.

Got a cap substitution box too.
Makes instant comparisons easy, or to zero in on a value.
No electros in the sub box, all 600V films.

[Chuck H]

Ok... The point IS that I looked at toobdude's drawings and interpreted that he may wish to null the NFB as a switchable option AS WELL AS have the bypass cap on that stage affected by the NFB loop as a switchable option. The circuit I proposed solves for both.

[Dave H]

The 100R resistor as a cathode "tail" and FB shunt allows the switching of a bypass cap across that cathode resistor to remain an independent circuit from the NFB loop and makes manipulation of the NFB circuit independent of the cathode bypass cap switching circuit. That was my goal.

I don't think it will be independent as it's still inside the global feedback loop.

I see the three options working like this.

1. 1k5//0.15u cathode and no FB gives a 6dB treble boost.
2. As above + 100R, 2k2 FB gives a reduced treble boost (because of the FB) of perhaps 2 to 3dB.
3. 1k5//0.15u cathode + 33k FB 'Presence' circuit gives >6dB (9dB?) treble boost but this could be reduce by adding a resistor in series with the 0.15u cap.

[Chuck H]

I don't think it will be independent as it's still inside the global feedback loop.

Agreed. But it's as independent as the two circuits could be (NFB null and brilliance cap switching) since both involve the same preamp cathode circuit. Considering TD would like to simply switch the brilliance AND switch the NFB I still think my proposed circuit is the best option so far. Everyone else is suggesting changes to the plan and I don't see that as necessary.

[Dave H]

For a worthwhile amount of brilliance you'd have to switch in the 0.15u cap and open the NFB loop otherwise the NFB will flatten out the response. To have your cake and eat it (independent brilliance and NFB) the brilliance would have to go on the first stage cathode.

[tubedood]

Thanks to everyone in helping with this. I now understand the impact of negative feedback in gain stages better and I will definitely experiment clipping/tack soldering into the circuit with these values and ideas to get an ear for what changes and why . The original plan was to add the .15 into the first gain stage but I had read that it might have a better impact in the second stage so if all else fails I can keep it as originally planned.

Thanks everyone for your patience, I know some (all) of these questions I ask to you all are complete beginner questions, but to me it reminds me that I should have stayed in school! That math stuff is all over these forums.

Dale

[Chuck H]

For a worthwhile amount of brilliance you'd have to switch in the 0.15u cap and open the NFB loop otherwise the NFB will flatten out the response. To have your cake and eat it (independent brilliance and NFB) the brilliance would have to go on the first stage cathode.

NFB will be flattening everything coming out of that triode, not just the effect of the brilliance cap. The relative amount of brilliance with the cap switched into the circuit should be the same as any partial bypass preamp cathode. It's not as if the NFB is going to hunt down the effect while ignoring every other frequency. So, yes there will be a reduction in gain, but that's just what NFB does. It shouldn't specifically reduce the affect of the brilliance circuit. So the relative HF boost of the brilliance switch should remain the same.

[Dave H]

It's not as if the NFB is going to hunt down the effect while ignoring every other frequency.

Ha Ha! Good one That is what NFB does (sort of) The excess gain in the feedback loop is used to make the output a more accurate representation of the input. If there's a factor of 2 excess gain in the loop then the boost given by your brilliance circuit inside the loop will be halved (approximately).

[Chuck H]

I'm not studied in electronics, but... I have a hard time reconciling this. What is a "factor of 2" with respect to what the triodes output is> The gain stage says "Here is the signal" and the feedback loop says "And here's some of it back at ya." Once the feedback loop is on the 100R instead of the top of the cathode it shouldn't know the difference between no bypass, full bypass or partial bypass on the cathode resistor. I understand that more gain from the "brilliance" being switched in means more "brilliant" frequencies being fed back. I don't see how the resulting relative ratio of frequencies in the face of feedback is altered though. I'm picturing it like this, say you start with (arbitrary numbers) 6@100Hz and 6@3000Hz and connecting feedback cuts that in half to 3@100Hz and 3@3000Hz. Now... Switch in the brilliance circuit and you have a start of 6@100Hz and 12@3000Hz. If feedback cuts that in half then you have 3@100Hz and 6@3000Hz. Now...

It's clear that because of the boost in HF there is more cut in HF due to the feedback circuit (cut 3@100Hz and 6@3000Hz) but the relative frequency balance is still preserved in that there is only half as much 100Hz as there is 3000Hz. Just as before any NFB was applied.

I can't know I'm right because I don't have the tech chops. But the way I see it, if my perception is correct then we are both right.

[Dave H]

I understand that more gain from the "brilliance" being switched in means more "brilliant" frequencies being fed back.

That's it right there! It's negative feedback. More of the brilliant frequencies are fed back so more of the brilliant frequencies are subtracted from the output (relative to the lower frequencies). I've done an LTSpice of the circuit above. The first one is without feedback and you can see the brilliance boost is 6dB from 100Hz to 5kHz (as you'd expect)). The second one is with the 100R, 2k2 feedback network and the brilliance boost is reduced to 3dB by the NFB.

[Chuck H]

I love that you're always game with a spice simulation I can't say that I have my head around it entirely, but the simulation is right there. So, back to the issue at hand then...

I think a 3dB boost may be BETTER than a 6dB boost. Too much gratification on the switch could make it a useless novelty rather than a desirable tone enhancement. What I don't like, now, is that the HF boost ratio for the brilliance switch is so different depending on the whether the NFB is switched in or out.

I was just PMing with TD about a possible hybrid circuit where the NFB is shunted somewhere elevated from where the brilliance cap would be grounded to sort of equalize the two switches affect on each other.

[Dave H]

I love that you're always game with a spice simulation

I already have a few circuits drawn up in LTSpice so it's usually just a case of adding a few components. I only had to add two resistors and a cap to an existing SE 6V6 circuit for the above plots. I can do real frequency response plots too. I have a program for my Picoscope USB scope which does it in a few seconds, like the one below.