3stage + CF options

[John_P_WI]

Looks good smoke.

The 470n shunts the diode hash noise (switching) and other line noise coupled through the PT. The small caps do not have the high ESR that the larger elytics have and more effectively filter this type of noise.

[Smokebreak]

Thanks John. So how important is value selection here? It would seem the largest non elytic I have, with appropriate V rating, would be be most effective, in passing noise to ground?

[John_P_WI]

I have always used 100n, 630v polypropylene film with good success. Any polypropylene with a 630v minimum rating and a capacitance 100n to 470n will work assuming your secondary voltages are not gonzo high.

[Smokebreak]

I'm a little confused about the tail resistor on LTPI, generally speaking, but specifically as it relates to Martin's Presonance.

I see designs with, say, the standard Marshall 10K tail>presence, and then some with the same 10K tail, but then another shunt(?) say 4K7 in series with the tail>ground. As I understand, it's best to introduce NFB 1/3 of the way up the tail. So, why do some designs not have this extra tail?

For the Presonance, Martin's calculations are 1-2 * tail resistor, to get the pot value, and I believe he is referring to the bottom/shunt tail resistor to use for that calculation, i.e. 10K-20K pot....Well, for the sake of discussion, the 6V6 Plexi doesn't have the "extra" tail...should I introduce it in my amp? Wouldn't that also be a factor determining the size of the pot in a regular presence circuit, much less Presonance?
I guess this is a roundabout way of asking when you would use the extra tail and when you wouldn't..

[martin manning]

I'm a little confused about the tail resistor on LTPI, generally speaking, but specifically as it relates to Martin's Presonance.

I see designs with, say, the standard Marshall 10K tail>presence, and then some with the same 10K tail, but then another shunt(?) say 4K7 in series with the tail>ground. As I understand, it's best to introduce NFB 1/3 of the way up the tail. So, why do some designs not have this extra tail?

If you don't have any GNFB, then there would be only two resistors in series from the cathodes of a LTP to ground- one for the bias, and another to make up the total tail resistance desired. The cap from the second grid would go straight to ground, too.

If there is NFB, then the lower resistor is split into two, with the bottom one forming the shunt leg of a voltage divider that reduces the signal coming from the speaker output by 90-95%. The cap from the second grid now has to go to the top of that last and now third resistor in the tail, because the (reduced) FB signal has to be fed into the second grid to make the FB negative (the speaker out is in phase with the PI's input signal).

If you want to add presence control, then the most common way is to parallel a pot/cap series combination with the bottom resistor to dump some of the FB signal to ground. The old-school alternative (which IMO actually works better, except for the shhhh when you move the pot) is to use the presence pot's element for the bottom resistor, and put the cap from the wiper to ground, a la the 5F6-A.

For the Presonance, Martin's calculations are 1-2 * tail resistor, to get the pot value, and I believe he is referring to the bottom/shunt tail resistor to use for that calculation, i.e. 10K-20K pot....Well, for the sake of discussion, the 6V6 Plexi doesn't have the "extra" tail...should I introduce it in my amp? Wouldn't that also be a factor determining the size of the pot in a regular presence circuit, much less Presonance?

Yes. In the presence circuit using the parallel pot/cap, the pot should be about five times the value of the bottom resistor in order that the presence can be more-or-less fully turned off.

[Reeltarded]

Pay close attention to what Martin saids about the 5k pot older version of pres. That is the one you want, always. The later version is a waste of a knob.

[Smokebreak]

I'm having a blast with this amp. Super fun to play. I'll gush about it later, but right now the only issue I'm having is a hum that occurs with our without input, and increases as volumes go up. My PT is putting out 7V for filaments, rated 3A, so I'm not sure why it hates me. I'm gonna go get a 6V lamp battery and see what happens. I used the chupacabra ground scheme and have reflowed everything. I even split pre/power grounds to no avail(chupa uses bus/star with everything at PT. The only issue I'm considering is that with the capcan, the PI cap is grounded with power. I noticed the chupacabra uses elevated DC heaters? I need to read about that.
Anyway, any thoughts or suggestions would be appreciated. I'm about tapped out on solutions.

EDIT : I just noticed that in standby, the transformer has a mechanical hum that is the exact same pitch as the amplified hum-Bb/B on A string. I believe that's 120hz. What is up with the PT hum?

[Smokebreak]

Cancel all that noise. I must have gone through 10 12ax7 swaps, so I was convinced that was not the issue.
I just got some nice sylvanias today, figured I'd plop one in, and hum is entirely gone. Ill leave this up as an embarrassing reminder, and publicly proclaim that I will never buy another new production preamp tube.

[Smokebreak]

Just wanted to update and say thanks a bunch for all the help! I put this amp down for a few weeks and went back to it today with fresh ears. i dig it a lot.
It's quite angry and loud and has that bordering on meltdown feel
I ended up with variable NFB(22K/100K), the peakers on a switch("sizzle"-470p,1200p) and a switch("texture") for differnt Ck for stage before CF-one big, one small. With the big one also a 500p from grid to ground gets switched in. The differences are subtle here but useful at high gain settings. I took out the presonance early on, but only because I didn't quite understand it, as you get 1 or the other. Now i think I do-the presence side is useful when not in high gain, and the resonance side is useful in high gain, or at least more useful than presence there. I had fixed resonance for a minute, but couldn't really dial it in. I may give all that another shot- the presonance is a really nice circuit

I used 22V and 5V for clippers and they sound good for bedroom volumes.

I'm still swapping around pre tubes to find a balance between microphonics and noise, which is actually quite low at this point.

Oh I moved around the dividers so that I could get 68Ks on added stage, and 470Ks on plexi 2nd stage.

Of course after all the modding I was left with extra space at the end of the board and not enough at the input side. of course I used a leftover piece of board for the bias, which is why it is so ridiculously huge

edit:just put 470p across stage 1 Rp and 500p across 3rd stage

Here's pics and a short clip I did when my neighbor left and I could dime. I didn't really dial it in, just cranked all 4 volumes and everything else at noon(actually the CF master backed off just a bit-sounds best). no clippers. I think I switched the "texture" at some point and had a lot of NFB. Crappy 15$ Best Buy mic.
https://soundcloud.com/jeremyslemenda/3stage-cf-dimed

[Reeltarded]

Rawk!! It came out great!

[Smokebreak]

Well, now I'm having fizz problems. I didn't really notice it until I started listening to the straight plexi channel. I guess it was kinda covered up by all the gain. So, I hardwired it today w/o the added 2nd stage, and without the CF master, to eliminate those as possible problems. Still there. It's like a whole nother set of tones that is worse in the lower octaves. It comes on with the gain above noon, and I can dial it out with my guitar volume. I've tried everything I can think of and can't make it go away. I've tried up to ridiculously high values for Rp bypass, PI plate cap, lowered my Rp to 100K. 5K6 on output grids, 22n output couplers, resistance between tonestack and master, all tube swap, new caps, grid resistors everywhere. All that stuff is effective in killing the tone, but the fizz is still there. It's really bad on low notes and chords...kinda splatty and harsh too...you can really hear it when notes fade ... and this is just with simple 2stages+CF. Nasty.
Any ideas? Here's a clip from earlier before I broke it down, but it's the same : https://soundcloud.com/jeremyslemenda/fizz

[Reeltarded]

hmm so maybe the fizz is a physical thing at a ground on a cathode or something? Do you think it's a tweak or a fault?

You mean the sustaining grit? Do you have the clipper in? turn the master to like 4-5 and the gain to around 2. still?

[Smokebreak]

hmm so maybe the fizz is a physical thing at a ground on a cathode or something? Do you think it's a tweak or a fault?

Tweak or fault in terms of fixing it? I'm not sure but it doesn't "feel" like any tweak is gonna help. But the amp is super quiet at idle, which somehow makes me think connections etc are solid. Ran it through a gold, a vintage 30 and some crap stock Fender speaker and its actually a little worse with the gold.

[Smokebreak]

You mean the sustaining grit? Do you have the clipper in? turn the master to like 4-5 and the gain to around 2. still?

Sorry I missed this part. Yea, that's the grit. In person its all over the notes from the start. The clippers are out. I took out the whole CF Master circuit to troubleshoot. It's basically 1987 w/ normal master right now.
Clean sounds are great and fizzfree . Master can go anywhere but gain can only go till noon or so..where it starts to breakup. With the breakp comes the fizz. I can hear the good tone along with the bad

[Reeltarded]

dime the normal master and the amp should remain cleanish on bright until about 4 on the dial with no bright cap at all.. ish.