Advice on this little mod I found for the cathode bypass cap 5F2-A

[tubedood]

Hi again!
I am looking thru some ideas on my little 5F2-A build and I came across an interesting "voicing" switch I guess we can call it on the 1st stage of the 12axy. It has 3 choices, one is just the 1.5K resistor, and the other two are either a 10UF or a .047UF each parallel with the resistor. Tonally what can I expect from the .047? Will it be a nice chimey treble sound? I don't think Ive seen a bypass of that value before... Sorry for so many questions I do appreciate your efforts

Dale

[galaxiex]

Small value bypass cap on the cathode resistor cut bass.

Here's a nice calculator for that.

https://www.ampbooks.com/mobile/ampl...ode-capacitor/

I've used .68uF for cathode bypass on the first stage
for an amp that had waaaay too much bass response for the speaker that was in it.

[The Dude]

IMO, the best thing to do is tack a .047UF in there and see if you like it. Then, if you do, you can leave it, make it switchable, etc.

[Enzo]

Do you have a 0.047uf cap? Tack solder it across the 1.5k resistor or even just use clip wires. Now you tell us how it sounds. No 0.047? Try a 0.1uf instead.

[The Dude]

You're scaring me, Enzo.

[tubedood]

Cool good thinking I'll try it for a few days and see how it sounds. The old "Prince o Wails" amp ( I posted schematic on here a while back) I was wanting to try again had a switchable .047 on V1B and he labeled it as "Brilliance" it really added a nice sound but it also had an inductor thing going on in addition to it. Wasn't sure how it would sound just itself and the resistor .

Dale

[R.G.]

Enzo is correct - as always! - about the simple thing being to temporarily solder on a cap and see if you like it. It's quick and simple, and lets you know if you like it.

On the other hand, I personally can't leave it at that. My personal approach to things like this is that I need to understand what a mod is doing to get effective control of it.

In the case of a triode gain stage, the gain will always be somewhat less than the ratio of the plate resistor to the cathode resistor. If you think about it, this must be true, because the same current flows in the plate and cathode, so the voltages across them simply must be in the ratio of the resistances. The gain is less because the tube's internal resistances eat some of the voltage change. But the gain is the plate resistor plus some internal resistances divided by the cathode resistor.
a
Actually - cathode impedance. Bypassing the cathode resistor with a cap changes things. The cap is an open circuit at DC, and its impedance decreases linearly with frequency. At some frequency, the cap's impedance to signal gets low enough that the cap is controlling the gain, not the cathode resistor.

That starts to happen at the frequency where the cap's impedance equals the cathode resistor. So the gain of the tube stage at low frequencies is some number set by the tube, plate resistor, and cathode resistor. The effective gain starts ramping up at a frequency F = 1/2*pi*Rc*C), and increases linearly from there. It's a treble boost, getting bigger as frequency goes up.

At some frequency, the gain begins to be limited not by the cap and plate resistor, but the tube internal resistances, so the gain flattens out again.

If you make the cathode cap big, the boost happens way in the bass part of the audio range, so the gain is high for all audible frequencies, and it's just a high gain stage. If the cap is made smaller, the gain is lower, then starts rising, then levels off again at some frequency in the middle. The effect is a bass cut, which sounds like a treble boost.

With some understanding under my belt, I can go play with numbers.

For a 1.5K cathode resistor, the frequency of boost is F = 1/(2*pi*1500*C). For C = 0.047uF, the frequency is 2258Hz, in the high mids to low treble part of the audio range. It ramps up from there. The treble sound gets more and more gain until the tube runs out of gain, so the treble end of the audio range is emphasized, the bass appears to be cut.

Having once calculated this, converting to other caps and frequencies is easy. If you double the cap to 0.094uF - shoot, call it 0.01 - the frequency where the boost starts is lowered by half, to 1129Hz, clearly in the mid ranges. Double the cap again and it drops to 551.5Hz. If you increase 0.047 to 4.7uF, the frequency where the gain boost starts is dropped to 225.8Hz, and at 10uF, you get to 112.9Hz, right down near the fundamentals of the lowest guitar strings. This points out why 22uF for the cathode cap is a common value - it gets all the guitar frequencies inside the boosted gain range. Lowering the cap, moves the emphasized frequencies up, making the treble more prominent and the bass effectively less.

[Steve A.]

For A/B comparisons I will tack solder a switch across Rk. To my ears a 0.047uF Ck cap is like a bright switch, a 0.68uF cap has a Marshall mid-rangy sound and 10uF on up has the bassier response of a BF/SF amp. (Dan Torres sold a "Triple Killer Kit" with a center off SPDT switch, and 0.68uf and 22uF caps for V1 on BF/SF amps.)

Steve A.

[tubedood]

Thanks so much to all of you. Great info here! I really am trying to "wrap my head" around the workings of these circuits. The folks in this forum have been great and very patient with my beginner questions! The main reason I'm asking about this .047 is because I remember that value in this old schematic I tried but failed in (It worked for maybe 15 glorious minutes) due to my poor workmanship and virtually zero understanding of what I was doing. Attached is that old "Prince o Wails" modded Princeton schematic he uses the .047 as the Bright switch and has a makeshift inductor as a mid cut/boost. I remember that brilliance switch really did open the treble sound up, I liked it. Hoping this simple .047/1.5K resistor sounds similar! I will temp tack in place and see! Any feedback on the attached schematic and its workings would be appreciated as it might be my next build. You guys are helping make this little tweed Princeton a very fun and informative project. I wish I started this 20 years ago with this community

.

[cerrem]

For a 1.5K cathode resistor, the frequency of boost is F = 1/(2*pi*1500*C). For C = 0.047uF, the frequency is 2258Hz, in the high mids to low treble part of the audio range. It ramps up from there. The treble sound gets more and more gain until the tube runs out of gain, so the treble end of the audio range is emphasized, the bass appears to be cut.

That equation is not quite valid...
The R value in the time constant needs to also include the "looking in" resistance to the cathode...ie 1/gm ...
This 1/gm is in parallel to the cathode resistor... ie the SOURCE impedance is always included when figuring the time constant, unless you have an ideal source...
The online calculator results look goofy...
If you are using one resistor for biasing both halves of the pre-amp tube, then you would need 1/(gm*2) ...such as early Bassman and Marshall V1 circuits..
1/gm for 12AX7 per cathode is "roughly" 700 to 800 Ohms.. in the typical operating conditions in these types of amps..
For your 5F2-A amp....
The ,047uF would have a POLE @ 6.6kHz
The 100uF would have a POLE @ 3Hz
A .68uF would have a POLE @ 450 Hz...

[Helmholtz]

It has been discussed here:
http://music-electronics-forum.com/s...ght=half+boost

I recommend posts #1, 27, 31.

[R.G.]

Ah. Tube internal impedance trips me up again. Sorry about that. Right concept, need more numbers.

[Helmholtz]

Our previous discussion (see link above) helped me to realize that:

- As the gain-frequency-response at the output exhibits a step and a step has two "corners", the complete description requires two corner frequencies. The lower corner frequency is fixed, the upper shifts with the gain ratio.
- If the step attenuation is low (say below 6dB), corner frequencies cannot be precisely defined and it makes sense to calculate the so-called half-boost frequency in between.






The internal cathode impedance is rk=(ra+Ra)/(µ+1). This makes G1/G2 = 1+ Rk/rk and f2= f1 (1+Rk/rk) = 1/(2piCkR*) with R*= Rk//rk. (I skipped one or two trivial conversions). The fully bypassed gain is G2= Ra/rk give or take. The cathode impedance acts in parallel with Rk and both together influence the -3dB (upper corner) frequency.

Glossary:
f1: lower (gain) corner frequency
f2: upper gain corner frequency
G1:lower gain limit
G2:upper gain limit
Rk: cathode resistor
rk: tube internal cathode impedance
R*: rk in parallel with Rk
Ra=RL: anode/plate resistor in parallel with external load impedance