preamp/PI ideas/design for 300ish watt amp

[black_labb]

i've been thinking about the preamp for the big amp i'm going to make. i am going to use a hiwatt style tone stack, and probably use the presence control and the PI as well.

the first question: the front end and driving the tone stack. i'm thinking i might try running a cathode follower into the tone stack, but i'm not sure if it will benefit the sound. anyone have any ideas what difference in sound i could expect? or will i have to try it myself.
i'm thinking i will have 2 inputs, 1 into a 6bl8 pentode section, 2 into a 12ax7 triode. they will have a vol control after both of them and then feed into the 6bl8's triode (think 12au7 triode) as a gain stage. that will feed into the second triode of the 12ax7 as a cathode follower then into the tone stack.

i plan on putting a switch in to make it possible to link the 2 input gain stages in parallel. i'm assuming that a pentode will have a phase change of 180deg just like a triode?

Q 2: i often see cathode followers on each half of the signal on large amps after the phase inverter. why is this? is it because so many tubes in parallel means there is more demand on current from the previous stage? this may be the way to go with 8 tubes if this is the case.

Q3: in this schematic http://www.mhuss.com/Hiwatt/Schemati...e4Input_v2.pdf (and some other variations of the hiwatt preamps) what does the 2nd half of v3 do? it looks like a cathode follower, but it doesnt have a signal input. i cant figure out what it is and keep thinking its missing something, but its the same in many drawings.

Q4: when using 2 diodes in series for extra voltage handling, whats the best way to ensure the work is spread evenly? a cap in parallel with each of them?
not related, but its been bugging me for a while. how do diodes (ie rectifier diodes) have a ground reference? how do they know what is positive and what is negative? i originally though they must just cut the peak/bottom of an ac signal, but if that was the case then you could put a bunch of diodes in series for dc smoothing.

[black_labb]

and another

is there any way to create a mid shifting control, to alter the frequency that the mid control effects? would a different cap value for the 1n cap to the left of the mid control pot (same schematic as above) change it? i might put a pot and a cap in series parallel with the 1n cap, or a switch to change cap values.

thanks for any help

[defaced]

1) I don't know about the sound you'll get, but could I'd think you might want to try using the second half of the 6bl8 as the CF and the 12AX7 as the gain stage sing the second half of the 6lb8 can deliver more current.

2) Yes, CF after PI is for increased current to the tubes. There's an old thread about running in AB2 that discusses this in more detail.

3) I have no clue.

4) In the power supplies I have seen there has not been anything to balance the voltage drop across each resistor.

They don't cut the peak, they just simply don't conduct in one direction below a certain voltage because one side of the diode is made of P material and the other side is made of N material. N material can only emit electrons and P material can only sink electrons.

For a tranny with a CT, the ground ref is the center tap. For a FW bridge, the ground ref is the negative end of the output.

Yes, turn the series resistor into a pot. I just tried this with a Marshall tone stack and didn't like it, but it definitely worked.

[Don Moose]

Your question 3 -

The nub of explanation I got was that V3b some how regulates the bias to the PI giving the HiWatt a more 'HiFi' power amp than the typical LTP.

The 'input' is the voltage divider from plate to GND across the grid and the output is the resulting DC voltage, across that 220k Rk. The previous stage then modulates that bias, just like the PI output to a biased power tube.

The details got away a long time ago, and I don't remember who explained it to me. Possible candidates: SteveA, Steven Conner, R.G.Keen, Gil Ayan, Randall Aiken or Duncan Munro. With any luck one of them, or one of the other smart guys, will wander in and spill the (correct) details.

[black_labb]

thanks for the help.

i wasnt sure wether to use the 6bl8 triode or the 12ax7 for the CF. the reason i had the 12ax7 was because i didnt want too much gain distorting the triode gain stage, but i guess the cf will distort more with the 12ax7 there. i wanted some distortion when close to full but clean most of the way through.

the diode direction thing makes sense now, i knew it had to be simple like that, just couldnt figure it out in my head. just wasnt thinking about it in the right way.

re the mid shift, i assume you mean a pot and a cap in parallel with another cap? i did it on a marshall ts as well on this amp http://music-electronics-forum.com/s...ad.php?t=10743 . i found it useful, but i tend to keep my mids on or near 10, so there's very little cut, and varying the frequency doesnt do much. i've heard that the hiwatt gives alot of mids, so that should suit me nicely, and being able to have the mids higher than the highs and lows would mean that a shifting mid would be useful. i might try a switch with different values for the cap.

[black_labb]

Your question 3 -

The nub of explanation I got was that V3b some how regulates the bias to the PI giving the HiWatt a more 'HiFi' power amp than the typical LTP.

The 'input' is the voltage divider from plate to GND across the grid and the output is the resulting DC voltage, across that 220k Rk. The previous stage then modulates that bias, just like the PI output to a biased power tube.

The details got away a long time ago, and I don't remember who explained it to me. Possible candidates: SteveA, Steven Conner, R.G.Keen, Gil Ayan, Randall Aiken or Duncan Munro. With any luck one of them, or one of the other smart guys, will wander in and spill the (correct) details.

thanks, that makes sense. i thought maybe regulation, but i thought that it must be something else.

[defaced]

In that schem, the 47k resistor is part of the circuit that provides the knee of the frequency cut off for both the bass and the mid pot. So from the node where the resistor ties to the trebel pot, give the bass pot the resistor and give the mid pot a resistor in series with a pot wired as a variable resistor - that pot will give you a mid shift. I found that with my setup there was a very small window of usable sounds, and it was right near the resistor value.

Edit: ignore the values and look at how this is wired. This is the setup I was playing with. You could also look at a parametric EQ for ideas.

[black_labb]

In that schem, the 47k resistor is part of the circuit that provides the knee of the frequency cut off for both the bass and the mid pot. So from the node where the resistor ties to the trebel pot, give the bass pot the resistor and give the mid pot a resistor in series with a pot wired as a variable resistor - that pot will give you a mid shift. I found that with my setup there was a very small window of usable sounds, and it was right near the resistor value.

ahh, if you look at the tone stack on my schematic, there is a 25k pot near the (unmarked) mid control pot. that controls the shift on mine by effectively changing the value of the capacitor for the mid control.

edit: i see what you've done. makes sense, separate the resistor leading to the bass and mid, and replace the resistor with a potso it is variable

[defaced]

I made an edit and attached a schem.

[black_labb]

I made an edit and attached a schem.

thanks, makes sense. now i just need to apply one of these methods to the hiwatt TS

[Steve Conner]

Hi all,

I don't know what the V3 does, but the explanation that it somehow smooths the bias to the PI seems to make sense.

In amps with parallel output tubes, the cathode followers are to stop the bias feed resistors from loading down the PI. (The more output tubes you have, the lower the bias feed resistance the PI must drive, since the maximum bias feed resistance given in the tube datasheet is per tube.) They may also help HF stability margins in the NFB loop, by pushing up the pole caused by the power tubes' grid and wiring capacitance.

Diodes conduct when their | end is more positive than their >| end, and don't conduct otherwise. That's all they know, and all they need to know in order to function. Modern diode types can share PIV by themselves when connected in series, they don't need any equalizing components.

[black_labb]

thanks steve. i'll have a look at the bias feed resistor data on the datasheets.

[teemuk]

3) It's a simple voltage regulator. The grid is tied to a DC reference, the cathode follower tries to maintain a corresponding DC voltage at the cathode and can provide a decent amount of current to do that.

If you take a look, some Hiwatts do not have this circuit but directly couple the PI to the preceding cathode follower. The outcome is pretty much the same thing (PI's grids are tied to a proper DC bias potential). Since in this circuit the preceding stage is a common cathode (not a cathode follower) the plate's DC potential would like be too much for direct coupling (in comparison to cathode follower).

[black_labb]

3) It's a simple voltage regulator. The grid is tied to a DC reference, the cathode follower tries to maintain a corresponding DC voltage at the cathode and can provide a decent amount of current to do that.

If you take a look, some Hiwatts do not have this circuit but directly couple the PI to the preceding cathode follower. The outcome is pretty much the same thing (PI's grids are tied to a proper DC bias potential). Since in this circuit the preceding stage is a common cathode (not a cathode follower) the plate's DC potential would like be too much for direct coupling (in comparison to cathode follower).

thanks teemuk. still not sure if i will use that specific PI, but i may