gain stage question

[iknowjohnny]

I went back to 22uf/500pf cathode/coupling. It just sounds so much tighter in the lows. So I put it back and found some changes to compensate for the things i didn't like about it and i pretty much have it where i want it. Well, close at least. It;ll do but i will likely brainstorm for a while and hope a lightbulb goes of that will lead me to a tiny bit more complexity that it had before i changed from the marshall values.

[Roe]

on my JTMs I've been using a shared 820/330uf cathode with 5nf or 3n3 coupling caps. Perhaps I should try 25uf cathode cap instead and bigger coupling caps?

[pdf64]

330uF is bigger than necessary in our application, the cathodes will already be fully bypassed by 22uF.
Have a play with https://www.ampbooks.com/mobile/amplifi ... alculator/
Click 'launch me'

[Roe]

yes, I just followed the schematic and rolled off bass with the coupling cap. I may try a 3.3uf film cap as cathode cap instead

[JMFahey]

In the old days (think the 40´s), construction was sloppy (and that´s an understatement) and it was common to believe that chassis was actually a good ground, so much so that filament current was returned through it, by soldering one end of each filament straight to chassis just by the socket, lots of amps and most all radios, used half wave rectification, etc ..... "Hum City" .
I guess they were used to it
Yet some cared and did things better, including our now normal center tapped filament windings.
Yet some AC induction interference "jumped" from filment to cathode .... specially if it was unbalanced as I mentioned before, so tube manuals of the day (and some datasheets) , suggested using "large" capacitor decoupling caps, not for Audio reasons, but to AC ground cathodes and kill induced hum.

In fact, som datasheets includedsome phrase sounding somewhat like: "for minimum hum use a large cathode bypass cap, 10uF or more preferred" .

Today it is not mandatory, since we use balanced 6.3VAC , or elevate heaters , or plain use DC.

Fender who was an earlier designer basically followed what datasheets said, with a few notble exceptions, while Marshall who came much later could afford to choose such caps to tweak sound.

The main goal was to cut deep bass for a tighter sound, very important in a distorting amp.

[tubeswell]

...
Yet some AC induction interference "jumped" from filment to cathode .... specially if it was unbalanced as I mentioned before, so tube manuals of the day (and some datasheets) , suggested using "large" capacitor decoupling caps, not for Audio reasons, but to AC ground cathodes and kill induced hum...

And the reason these were applied to V1, is because that is where the S:N ratio is greatest.

Same principle in the original 6G15 circuit for the reverb recovery stage.

[RJ Guitars]

In Merlin's book he talks about the need for using a 22uf bypass on V1 cap rather than a lesser value as a critical factor in reducing hum... I am slaughtering my quote of the reference here but the point is that you potentially allow some excess low end through as you kill some unwanted hum. I take the lesson here as being it is ultimately easier to throw away some of that excess low frequency after the fact than it is to get rid of hum. I typically use a 22uF cap on the first stage bypass but I would be curious if 1) I understood things correctly and/or 2) if anyone has ever heard a little extra hum when using a small bypass cap on V1.

[andresound]

Dumble typical V1 cathode bypass is 5uf. VERY quiet amps! My uneducated 2c worth.

[TUBEDUDE]

On two similar builds, I went from 20uf Ck on the first amp, to using 4 or 5 uf's. Both are extremely quiet, but paradoxically, amp number 2 is a tad quieter.

[Malcolm Irving]

Although the fundamental frequency of the low E on a guitar is 82 Hz, I’ve noticed that much lower frequencies are also present in the signal (associated with percussive effects such as palm muting). I discovered this by playing guitar into a big PA sub-woofer, just to see what happened!
I wonder if a very low pass frequency at V1 might influence things by allowing the low frequency stuff to block V2 or a later stage. Just a thought!

[Mant]

I am a fan, after much experimentation,
Which is also subjective, of 4.7ufd for first cathode bypass.

[Ken Moon]

I generally center bias the input stage. With a typical preamp B+ of between 250V and 300V, this means a plate resistor of 100k and a cathode resistor of 1.5k.

I like to partially bypass the input stage with a smaller value cathode bypass cap, usually 1uF. The frequencies that are not attenuated with this setup are approximately 100Hz and above. This design eliminates boominess and reduces note smearing, making both single notes and chords sound better than with the typical Fender setup with a large bypass cap.

Don't try to adjust the input stage to affect distorted tones, as there is not enough gain there to accomplish anything. You're just trying to boost the input signal enough that the following stages have a large enough input to work with, and trying to keep the noise level low (using a smallish 10k grid resistor on the input stage reduces the resistor noise, which the leading contributor to noise throughout the signal chain).

I also like to use very small coupling caps, usually 4.7nF or even 2.2nF, on all succeeding gain stages. If the input stage's output goes to a Gain pot or voltage divider, a small bright cap, between 180pF and 270pF, adds a little sparkle, and can easily be made switchable to adjust for differing user preference and guitar brightness or lack thereof.

Following stages that alternate between cold and warm bias add a nice-sounding complexity to the distorted tone.

[iknowjohnny]

Hmmm...i didn't realize that. I was using 100k plate, 1.5k cathode but i thought i read somewhere 2.7k was center bias so i tried that and liked it on V1A. Rigged up a switch and A/B'd them and the 2.7k just sounded cleaner so i am using that. Tried it on V1B too but there i didn't like it. On the input stage it seems to be a tad more touch responsive, more dynamic. Back to a .68uf too. Just sounds more balanced in this circuit even tho you'd figure it would be thinner.

[tubeswell]

Compare load lines with differences in biasing

e.g.: 12AY7 using 5F6A V1 values vs 12AY7 with 2k7 cathode resistor

Note the difference where the cathode load line (green) crosses the plate load line (blue) in each top graph, vs the difference in frequency response (each bottom graph. The bias point shifts colder (obviously) with Rk = 2k7, and the freq roll-off shifts marginally to the right

[Ken Moon]

Thanks for the graphs, tubeswell

iknowjohnny,

I've experimented with this quite a bit with several amps, using a signal generator and scope, with a 5k pot in place of the cathode resistor, looking for the spot where grid current limiting and cutoff clipping occur at the same time as the input signal is slowly increased.

The value I repeatedly come up with for Rk with a preamp B+ of 300V and a 100k plate resistor is about 1.7k.

So 1.5k could be said to be very slightly warm biased, and 2.7k is slightly cold biased.

1.8k is the closest standard resistor value for true center bias under these conditions, but 1.5k imparts a slight bluesiness and musicality that I find to be very pleasant in the input stage, even in a high-gain amp.

Going to a colder bias using 2.7k imparts an edgier (crunchier?) tone which is also harder-edged and more brittle sounding to my ears, but this may be desirable to some players, and may explain why this value is often found in hard rock/metal amps.

As you've already done, the best idea is to try different values like 1.5k, 1.8k, 2.2k and 2.7k, and go with what sounds best to you. After all, it's your amp, so make yourself happy