Voltage and tone and current and feel and amp design

[dehughes]

So yeah, I don't know what I'm talking about, BUT, I do know that on this last amp I built, the bridge pickup of my Tele just seems to be on the harsh side, especially on middle-of-the-fretboard notes on the G-B-E strings. The amp is Brown Deluxe/Komet Constellation preamps run in parallel, into a Voxy PI and a Tweed-ish 6V6 power section. MM trannies, G12H30 speaker kinda broken in. High quality components throughout. My first instinct is to re-examine how I'm running all the tubes...

Would high (maybe too high?) power section dissipation cause harshness in the midrange? I'm running it at essentially 90% as-is...cathode biased Class A, basically.

Would low (maybe too low?) voltages in the preamp cause harshness in the midrange? Both sides exhibit this midrange irritation....and when I bumped up the voltage to the 6SN7 (Constellation side) it warmed up/filled out that side significantly. So, I'm wondering if that's the ticket...

I've tried different speakers and cabinets and EQ settings, but its just weird, as my Tele is AWESOME and always sounds good (rebuilt Am. Std., with Lollar Alnico V, Glendale bridge/saddles, etc...) through just about any amp, so I'm wondering if my design of the amp is suspect. Or, maybe I've fallen off the deep end of tonal OCD...

I will say, though, that when I crank up the Brown Deluxe side, I'm loooooving it....but it's still harsh at lower volumes, as is the 6SN7 side.

All thoughts/suggestions are appreciated.

[dehughes]

Here is perhaps a question more easily answered:

What's the best way of reducing voltage in an amp? Different recto tube? Zener diode on the CT? I ask because these transformers are Mercury Magnetics repros of old school Brown Deluxe transformers, so they're running at a higher voltage than might be necessary due to today's wall voltages.

Of course, then, in theory, if I had a real Brown Deluxe, it too should be running at higher voltages as well...but, my main point is that I think I'd like to drop all voltage in the amp about 5-10%, but don't know how best to do so.

[Andy Le Blanc]

Back bias on the center tap, balance/ buffer resisters before the rectifier
isolation transformer at the wall, VVR.

It could be high gloss turd shine, or you have an amp that finally lets you
hear your tele the way it really is. Not everything tube sounds good.

There are many ways to address a harsh amp:

I'd re-think how and were I had applied by-pass caps, one thing mentioned
in the trainwreck pages, in conjunction with this, low value caps across the
plate resistor. A low value cap from the plate to the cathode. Or simply a
resistor between the plate and coupling cap.

Changing supply voltage with out redesign will affect the gain structure.
You can get away with it, tubes are 20% device, but I'd be looking at the
published tube data and R-C amp data at the same time. A few times in the
past, Ive suffer'd my own assumptions, and got the plate volts in the pre
too high with negative results.

[dehughes]

Back bias on the center tap, balance/ buffer resisters before the rectifier
isolation transformer at the wall, VVR.

It could be high gloss turd shine, or you have an amp that finally lets you
hear your tele the way it really is. Not everything tube sounds good.

There are many ways to address a harsh amp:

I'd re-think how and were I had applied by-pass caps, one thing mentioned
in the trainwreck pages, in conjunction with this, low value caps across the
plate resistor. A low value cap from the plate to the cathode. Or simply a
resistor between the plate and coupling cap.

Changing supply voltage with out redesign will affect the gain structure.
You can get away with it, tubes are 20% device, but I'd be looking at the
published tube data and R-C amp data at the same time. A few times in the
past, Ive suffer'd my own assumptions, and got the plate volts in the pre
too high with negative results.

Awesome. So true... I'll go through this line by line and see what comes of it. Thanks!

[dehughes]

So, I'm thinking that it's something in the PI that's causing my harshness. I can dial it out with the Cut knob somewhat, but you can still tell it's there. Voltages seem to be in line throughout the amp...falling just shy of the Brown Deluxe voltages on the 6G3 schematic by about 3-5%.

I'll keep tinkering...but I'm almost certain it's coming from the PI or power section, as both of my (very dissimilar) preamp channels exhibit this same annoying harshness...

[Andy Le Blanc]

One thing to check with the PI is DC balance, put a pot on one side and go
for a close to zero difference between the plates. With a cathode coupled
long tail you can't easily balance AC and DC at the same time. A bad imbalance
can effect tone. One more thing to look at.

[dehughes]

One thing to check with the PI is DC balance, put a pot on one side and go
for a close to zero difference between the plates. With a cathode coupled
long tail you can't easily balance AC and DC at the same time. A bad imbalance
can effect tone. One more thing to look at.

Huh. 100k on each plate gives me 225v on one side and 228v on the other.

I used to have a 6G3 Deluxe PI in there (1k5 under a 6k8, grid caps to input/top of 1k5, 1M to grids, 82k/100k on plates), but changed the PI from stock 6G3 to more of a Voxy PI (47k underneath a 1k2, grid caps to input/ground, 1M to grids, 100k/100k on plates).

I'm beginning to think my issue is somewhere in the PI/Powersection for sure, as I've tinkered with the voltages/cathode caps/etc. in the preamp and still find this uniform midrange "quickness/harshness" regardless. I can't remember if it was present when I had the PI at the stock 6G3 values.

What exactly am I doing to the PI going from these two sets of values? I'm running it cooler with the "Voxy" PI, right? That might make it a bit harsher, yes? Does putting the unused grid cap on top of a resistor vs. straight to ground change things, tonally?

Thanks for your help...I'm slooowly getting it...

[Andy Le Blanc]

Are you using feedback? Like the 6g3 but with the vox values?

The 1.2k in the vox, and the 820 in the deluxe provide bias and set the operation point.
The "tail", the 47k in the vox, 6.8k in the fender provide feedback by themselves
without the addition of global feedback. The fender implements a ratio of feedback
in conjunction with the tail, the vox does not.

[dehughes]

Are you using feedback? Like the 6g3 but with the vox values?

The 1.2k in the vox, and the 820 in the deluxe provide bias and set the operation point.
The "tail", the 47k in the vox, 6.8k in the fender provide feedback by themselves
without the addition of global feedback. The fender implements a ratio of feedback
in conjunction with the tail, the vox does not.

No feedback (r.e., 56k resistor from speaker jack to PI tail)...the PI goes off to a cathode biased pair of 6V6s (300R, 25uf), with a Cut control (a la Vox) instead of global feedback.

I've read that putting in global feedback in addition to a cathode biased power section is moot......something about the cathode resistor negating anything you'd hear from feedback.

I'm almost positive my issue is somewhere in here...

[Andy Le Blanc]

OK, I'd look at how the channels are mixed and what the inverters input "sees",
and maybe try a 100 to 500p cap on the plate resistor of the inverters
input stage. I used to describe it as an overtone, that annoying character
you cant seem to resolve. What value are your grid resistors?

[dehughes]

OK, I'd look at how the channels are mixed and what the inverters input "sees",
and maybe try a 100 to 500p cap on the plate resistor of the inverters
input stage. I used to describe it as an overtone, that annoying character
you cant seem to resolve. What value are your grid resistors?

Exactly...a sharpness to the tone. Odd...

I'll just have to re-draw the schematic and post voltages. Here's the deal thusfar:

Channel 1: 6SN7, 115/118v per plate, 4-ish volts on the cathodes, 220k plate resistors with 2k7/25uf on the cathodes. Triodes run in series with a volume pot only (500k) in between.

Channel 2: 12AX7 - basically dead-on Brown Deluxe, 150-ish volts per plate, 220k plate resistor on first triode, 1k5/25uf and 2k7/25uf on the cathodes, 15k/100k plate resistor on second triode with a .01uf cap to the PI. Volume/tone in between the triodes, .01uf to ground on the tone (1M), with a 500pf to the volume (1M)

Channels mixed with 220k resistors into a .01uf cap and off to PI

PI: 12AX7 - 100k/100k plate resistors with 100pf across the plates, 1M grid resistors joining under a 1k2, which sits over a 47k to ground. .01uf on the input side, .1uf on the unused grid side right to ground.

.1uf coupling caps into a Cut control (a la AC30) into 220k resistors from grids to ground, 1k5 resistors into 6V6 grids

Power Section: 6V6 pair, 300R/25uf on the cathodes, 352v on plates, 22v on cathode

Recto: 5AR4, about 360 off of the plates

Does that help? Anything suspicious? I'll re-draw a schematic ASAP...

[Andy Le Blanc]

The description is good. I'd try a 220k to ground at the junction of the mixing
resistors. Vox scheme have the have the top boost and normal channel mixed
thru 220k, when the volume of the normal channel turned down the inverter
see's a 220k to ground. In many fenders the normal channel see's 270k to
ground thru the verb/vibro channels mixing resistor and trem's intensity control.
You have a lot range to explore with the grid resistors. You can also
try lifting the cut control out, I haven't favored it myself, its worth a try if you haven't.

[smzinno]

dehughes - you might also try lowering the cathode resistor and tail values in your PI to something like a 470 cathode resistor with a 22k tail. I've had good luck warming up tone with these values, which are basically from the blackface design.

[dehughes]

The description is good. I'd try a 220k to ground at the junction of the mixing
resistors. Vox scheme have the have the top boost and normal channel mixed
thru 220k, when the volume of the normal channel turned down the inverter
see's a 220k to ground. In many fenders the normal channel see's 270k to
ground thru the verb/vibro channels mixing resistor and trem's intensity control.
You have a lot range to explore with the grid resistors. You can also
try lifting the cut control out, I haven't favored it myself, its worth a try if you haven't.

Thank you, Andy. You've given me lots to look into. I greatly appreciate your time and effort put into my "issue".

Also, could I put a 50k pot in place of the "Cut" control and have that be a "Presence" control, providing feedback off of the speaker jack back into the PI? I'm not sure if that'd be moot with a cathode bias power section (have heard it negates the effects of such feedback).

[dehughes]

dehughes - you might also try lowering the cathode resistor and tail values in your PI to something like a 470 cathode resistor with a 22k tail. I've had good luck warming up tone with these values, which are basically from the blackface design.

Thanks man. Good suggestion. So, wouldn't those values be for a 12AT7...which would be a higher current/lower voltage tube? I mean, it'd probably work, but I'm curious as to what that'd do with a 12AX7 in there...probably run it hotter, kinda like having a 900ohm Rk on each triode, huh.