Mixing Channels

[scstill]

I'm having a problem mixing two channels on this project.

Both channels work well separately - ie when the other channel is disconnected from the circuit not at the input switch. But when I connect them as in the schematic there is a drop in output and a hum is produced. Can see this in the pic, the black lead is to the pot, yellow lead from the pentode channel coupling cap, white lead from the triode channel coupling cap.

We previously talked about this in the "Phirty6" thread and it was thought they should mix well since channels are in phase.
see #9 and 11 here https://el34world.com/Forum/index.php?topic=29688.msg326827#msg326827
Thinking that there may be a voltage divider at input I tried reducing the input grid leak and Grid stop but the long leads created a noise pickup problem so I wanted to check before I soldered in new values

Both channels are grounded at the same star point being careful to have only one ground to chassis, except at the input jacks. Can't figure how any ground loop is caused. There is no hum with channels separately (ie other disconnected from circuit not via the input switch)

Can see this build here.
Its ok, the site is safe just the cert expired. Still waiting for son to fix.
https://stillampd.com/phirty6

[sluckey]

Mix with two 220K resistors. Any better?

[scstill]

Tried 220k mixing resistors between both coupling caps and the pot.
It is "better" in that the signal is not attenuated as much as without 220k and the hum is reduced,
BUT the signal is still reduced and the 220k seem to be sapping the fullness of the signal, seems thin even with the other channel disconnected at the mix. Also tried 1M and 100k with less success.

When both channel inputs are switched in, it seems to be good.
Issue when one or the other channel input is switched open.
Thinking maybe an output switch (before mixing) so that the channel not used is removed at both the input and at the mix.

[sluckey]

Thinking maybe an output switch (before mixing) so that the channel not used is removed at both the input and at the mix.

BINGO!

[HotBluePlates]

Tried 220k mixing resistors between both coupling caps and the pot.
It is "better" in that the signal is not attenuated as much ... BUT the signal is still reduced and ... seem to be sapping the fullness of the signal ...

Look at your two 220kΩ mixing resistor, and pretend you're seeing a 440kΩ linear-taper pot set halfway.

It's not "half-loud" (more like "90% as loud"), but each channel's signal is attenuated to half-voltage.  That's the nature of channel-mix resistors.

[PRR]

A passive mixer will always have loss.

In theory two input lose half each side and sum-back to same-level, but it never works that way.

...when one or the other channel input is switched open.

Well, an "open" input sucks-up the hum in the air, right?

[sluckey]

You need to wire your input switch such that it will send a ground to the UNSELECTED channel grid. This will probably help the hum issue.

Look at this pdf to see how I do that...

     http://sluckeyamps.com/phoenix/phoenix.pdf

[kagliostro]

Common Anode Mixer (any tube can be used, not strictly a 12ax7 / ecc83)




----






Franco

[scstill]

Well, an "open" input sucks-up the hum in the air, right?

You need to wire your input switch such that it will send a ground to the UNSELECTED channel grid. This will probably help the hum issue.

I had not thought of open input grounding, the switch approach is clever.

The goal is to have a triode channel and a pentode channel, individually and combined,
using these older oddish tubes.

I wonder if I might just hard wire the input channels and move the switch to the 2nd stage mix?
Or might there be more success with switching the input (with ground) and anode mixing as below? a lot more effort

Common Anode Mixer (any tube can be used, not strictly a 12ax7 / ecc83)

So if I stay with switched input and mix at the second stage,
would this anode mix approach be applied to the 75 and 6D7?
Would this eliminate the signal "thinning" and loss since it is an active mix?
Assuming mix P1 is a trim pot, is it being adjusted for channel balance or something else?

[kagliostro]

So if I stay with switched input and mix at the second stage,
would this anode mix approach be applied to the 75 and 6D7?

Yes, after the second stage

Would this eliminate the signal "thinning" and loss since it is an active mix?

Using this mixer the loss in mixing resistor is avoided, about "thinning" I don't know (BTW, English isn't my language and I don't understand exactly what you mean with "thinning")

Assuming mix P1 is a trim pot, is it being adjusted for channel balance or something else?

That pot acts like a balance on mixing the two channels each with his level

If you look to the other schematic I posted you can see that the use of that pot can be avoided

Franco

[scstill]

Using this mixer the loss in mixing resistor is avoided, about "thinning" I don't know (BTW, English isn't my language and I don't understand exactly what you mean with "thinning")

It sounds fuller (some use the word "fat") when the  channels are not mixed.
When they are mixed (as I did it) the sound is not as full so I used the word "thin"
just a weaker less pleasing sound I guess.

[AlNewman]

You would think stratocasters would be louder using 2 pickups rather than 1, but they aren't.  There are hum cancelling benefits, and also benefits of dialing in a certain sound that you can't achieve with just one pickup. They are always thicker using one or the other separately.

Maybe instead of a mixing pot, you could consider a volume or tone pot on one side or the other, independent of the other channel.

[PRR]

Common Anode Mixer

The maximum output of a plate mixer is quite limited. One tube can't drive well when dragging another tube around. In small stages this may not matter, but this plan seems to be driving a 6V6 directly from the mixer? Mix resistors may double the clean(ish) drive to the 6V6 grid.

[kagliostro]

Right, also I was considering as a source of problems the absence of a further gain stage that will drive the Power Tube

What about to swap the 6V6 to a gainy tube (El 84 - EL34 or other with a higher mA/V) instead to add a gain stage ?

I see, you use a 41 tube not a 6v6 tube (so I assume you don't want or can't swap it for one other tube)

Here attached a simple thing you can give a try and see what happen

Note that I don't know which will be the correct Grid Leak resistor value for the 41 tube so I wrote 1M but if you know that value it will be better you use the correct value for it

I've find that 500K is the Max allowable value for the Grid Leak resistor of the 41 Tube (and you must respect this value) so I corrected the schematic

I eliminated the MV and put two 1MA pots as volume on each channel, also put a 470k Lin pot as mixing unit (it is near the same of two 220K mixing resistor but allow control on the mix level)



Franco


p.s.: Those 46k - 32k - 82k Grid Stopper are indispensable ?

[scstill]

Worth a try, but for the long term this build is supposed to be a simple one volume design. I like the approach of hard wiring the mix at the input and switching the channel in or out after the 2nd stage.
BTW - I have a typo on the schematic as the installed volume is 500k not 1M, so the max grid leak for 41 is indeed honored. Thanks for pointing this out.

So why a 41 and not a 6v6. Well the Philco 926 (from which Phirty6 evolved) came with it and I wanted to build honoring its background. It is also my first build with grid top caps which I find pretty cool. Something different in a repurpose build, the 6v6 is way common and have built many of them. BTW - see the "Black and Blues" build as that is a car radio conversion using the original push pull Delco 6V6, has a Fender channel and a Gibson channel that can be patched. https://stillampd.com/black-and-blues-05 the cert is expired but site is safe....

The Philco 926 also came with two 78s, a 75 and a 6A7. As much as I tried I could not come up with a pentode design for the 6A7 pentagride (5 grids!) and so replaced it with the Sharp cutoff 6D7 ( a pretty cool tube and the only sharp cutoff pentode I could find with 7pins like the 6A7). The 78s were remote cutoff pentodes and so they were replaced with its sharp cutoff brother the 77 (although running the 78 in place seems ok for my ears). One the 77s was strapped as a triode and run into the 75 triode. Thus the Triode channel has two stage triodes and the Pentode channel has two stage pentodes. The lineup is more than loud enough for the "big" 6" that is in the cabinet and tone seems pretty decent. I suppose it will sound even bigger through the 212.

[kagliostro]

.... It is also my first build with grid top caps .....

I don't know the spec of those old  tubes you are using but .... top grid may be source of problems, you must use shielded cable and shielded top caps

One other thing, a pentode that follows one other pentode will result in a very high gain, may be it is too much, have you auto oscillations problems ?

Is there the reason for the grid stoppers about which I asked you ?

Franco

[scstill]

Sorry I didn't recognize the grid stop as a question. Each of those grid stops is calculated to pass 20khz given the tube internal capacitances (not always easy to get on these old tubes). Calculations for this amp are here - https://stillampd.com/phirty6 again cert expired site is safe. Theory is here https://www.valvewizard.co.uk/gridstopper.html

This amp is really really quiet. Thinking its because the grid top caps and thus signal lines are physically separated by metal chassis from the AC and DC voltages as well as the use of shielded signal lines (caps are not shielded).

Remember the problem is not hum it is a loss of signal and thin tone when the two channels are combined. Individually they sound really good. The extra hum when mixing was created by open input channel, which when grounded (as Sluckey suggested in this thread) the hum was gone. 

Surprisingly, the Pentode channel has less gain (not as loud) than the triode channel. Thinking of adjusting the plate voltage on the pentodes to get a little more. All of my load lines are also included at the link above.

[PRR]

...you use a 41 tube not a 6v6 tube...

Thanks for catching that. My mind said "type 41 is like a 6V6" (more a 6F6?) but my fingers went right to "6V6". They are not the same but not too very different. (And of course '41 uses the "old" base/sockets which are already in the old chassis, not Octal.)

[kagliostro]

The extra hum when mixing was created by open input channel

This I can't understand

If you are mixing the channels at the input how can you have an open input channel ?

Did you forgot to join the inputs when you join the channels at the output ?

To join only outputs without the inputs didn't seems to have sense

Franco

[sluckey]

The extra hum when mixing was created by open input channel

This I can't understand

If you are mixing the channels at the input how can you have an open input channel ?

Look at his schematic. The 1M on the input jack provides the ground reference for the selected channel. But the switch breaks this ground reference for the unselected channel leaving the grid with no ground reference. The grid is just floating and that causes the hum.

[scstill]

in testing having an open channel when mixing/switching after the 2nd stage does not cause any additional hum. But is it still best/good practice to ground the open channel at this point? Plan to use the clever Phoenix grounded switch approach at the 2nd stage mix just have two in and one out VS the Phoenix which had one in and two out...

Indeed the 41 was already in the radio with its 6pin socket (predecessor to a 6K6, 42 is predecessor to 6F6 to 6V6). As well I tried to keep the 7pin socket of the 6A7, which was difficult to discover the 6D7 as 6A7 replacement (like finding hen's teeth). Project goal to repurpose as much of the radio into new life as a great sounding guitar amp. What's good in a project like this is the historical knowledge gained in the research is rewarding and fun. Another purpose for me in this hobby.

BTW -  I read that the  41 and 42 are similar (more filament current in 42) and may be substituted for each other. Might try that. I do have a half completed project with 42 in PP using a Aetna 19w66a. Discussed earlier in this forum as "Prew42" with details here https://stillampd.com/prewar-42-guitar-amplifier cert expired but site is safe. I will likely use the mixing concepts learned here in the Prew42 project.

[sluckey]

in testing having an open channel when mixing/switching after the 2nd stage does not cause any additional hum. But is it still best/good practice to ground the open channel at this point? Plan to use the clever Phoenix grounded switch approach at the 2nd stage mix just have two in and one out VS the Phoenix which had one in and two out...

You probably don't need to ground the unselected channel but it can't hurt anything to do so. I did ground the unselected channel in my Dual 50 but I'm not convinced it was necessary. I never built the Dual 50 so can't say for sure.

     http://sluckeyamps.com/misc/Marshall_Dual_50.pdf

[scstill]

I never built the Dual 50 so can't say for sure.
http://sluckeyamps.com/misc/Marshall_Dual_50.pdf

My son has always wanted a Marshall.
This might be a cool approach, never had a EL34 project.
Double the EL34s to 100w for alien melting and zombie zapping...

[shooter]

I've tried the input switching like you're doing, once I built the "plexi" style pre, blending the 2 "channels" that became my standard, neither has a WOW factor stand-alone, blend the Vol's and WOW happens.


it mates well with PP or SE PA sections