18W Stout input channels

[terminalgs]

I'm building Doug's 18W Stout with a 2nd EF86 channel connected to what is normally the ground PI grid. 


One constraint of the build is that I'm building this with a 60's Concord amp faceplate (built in NZ, somehow a faceplate ended up in the US).  The faceplate has 3 channels, each with a volume control, and a single tone control pot.  I also want to position the extra 12ax7 triode as inline gain stage selectable by which input jack is used (like a MB MK.I),  rather than a relay or toggle switch.


The faceplate constraint is why the tone pot is a dual ganged pot (as apposed to two separate pots..).  Ch.1 is the normally aspirated 18W input channel. Ch.3 is the EF86 channel.  Ch.2 is extra gain stage that (right now) feeds both inputs.  R11 and R12 are yet to be determined.  I'm not sure if they will be needed, but I'll leave room for them and use a jumper if they turn to not be necessary.  R11 would form a voltage divider with R10, and R12 with R8..


R18 and P3 will be tweaked/sized.  R18 might not be needed.  I'm not completely sold on the EF86's channel's tone control BTW.


I'm posting for feedback/thoughts, so please let me know what you think!

[sluckey]

Have you considered this?

When you plug into CH 2 you will be feeding ch1 pre and ch3 pre. Those signals will be in phase when they reach the grids of the PI. But the PI will subtract those signals causing cancellation of the combined signals on the PI plates. This cancellation will probably result in a very thin sound.

[terminalgs]

Have you considered this?

When you plug into CH 2 you will be feeding ch1 pre and ch3 pre. Those signals will be in phase when they reach the grids of the PI. But the PI will subtract those signals causing cancellation of the combined signals on the PI plates. This cancellation will probably result in a very thin sound.

I had not considered that at all!  originally, my plan was to run the extra 12ax7 gain  stage into the ef86 only,  but then I thought "run it into both for free"... 'not free' as it turns out, I'll revise to run into one or the other..


excellent observation! thank you!

[sluckey]

You could use a couple mixing resistors to feed only one side of the PI?

[terminalgs]

You could use a couple mixing resistors to feed only one side of the PI?

I was doing it this way to avoid the mixing resistor network's signal attenuation.

[DummyLoad]

You could use a couple mixing resistors to feed only one side of the PI?

I was doing it this way to avoid the mixing resistor network's signal attenuation.

or simply switch between the two - no loss then. can't play them both (channels) at the same time anyhow.


--pete

[sluckey]

I was doing it this way to avoid the mixing resistor network's signal attenuation.

The input impedance of that LTP PI is verrrry high. The attenuation won't be nearly as much as you would think.

I like the possibilities of tone variety from running the channels parallel. I think it's worth experimenting some. It may sound like crap but I'd give it a try.

Another idea would be to connect the output of the EF86 to the second input of the PI as you've shown, but change the input such that when you plug into ch2 jack you connect directly to the EF86. This would make the phasing correct on  the PI output. That EF86 can sound pretty gritty without having an extra triode in front of it. Just a thought. Easy to try.
 

[terminalgs]

 

build update:
I finished the amp this weekend and fired it up (finished meaning I got it to the point that I begin to tweak!).  It sounds really great.  This truly feels like a lead guitar amp.   The EF86 channel sounds fantastic. I'm still working on the details of Ch.2 extra triode stage.   I have it built just like the schematic above except R11 and R12 were omitted, and I installed a toggle to switch Ch.1's switch input to ground or to the Ch.2's volume pot output (to play with that phase cancellation).


speaking of which, the phase cancellation is obviously there..  when I plug into Ch.2 only, and play with the volume controls P1 and P3,  there is a point of maximum cancellation when you turn up one channel and down the other,   you get that dead spot: first, one channel is "winning", then its a "tie" (dead spot), and then that channel is "losing".   If I fiddle with the 12AX7's channels tone control (P4a),  I can get different results out of that dead spot (i.e. turn P4a to the bass sounding end, then turn up the EF86 channel until those bass frequencies are cancelled ,, leaving the top end).   Still playing with it.   If in the hands of someone who doesn't know the details, my guess is they would assume something is wrong with the amp.


I installed a master volume with a dual-ganged 500K pot in place of the two 470K grid leak resistors. While handy for "bedroom volumes" (and for lowering the volume while testing the amp!),  MV doesn't seem like it'd be too beneficial in  a stage or studio environment.


I also added a Vox style "cut" knob with a 270K pot and  a .0047uf cap between the PI outputs..  I haven't played with it much other to see that it worked.


 
I still gotta tweak these consecutive input stages and the switch jacks.  It sounds really awesome with either 12AX7->EF86 or  12AX7->12AX7,  but it feels like it'll require too much "owners manual" kind of knowledge of the amp AS-IS (If I drop it off at a studio,  I suspect I'll get an email "hey, guitar player loves the amp, but Ch.2 doesn't work if turn down Ch.3...")..  If I had stereo switchcraft jacks with both tip and ring switched, I might be able to be more creative...


Oh, and I put a 470K resistor between C1 coupling cap and the 1M P2 volume control to attenuate the output sent to the EF86 (or other 12AX7 1/2).  I might swap the 1M to a 500K pot.

[terminalgs]

here's a couple of pics (they looked more in focus on my phone!):





I used Doug's tag/eyelet board because I figured there might be a fair amount of experimenting.  The board sits on 1/2" spacers.  there is a little bit of goofiness at the far end with the preamp cathode components, but that is because I wanted to have flexibility to give that extra 12AX7 triode it's own B+ node if needed.  Right now, all the preamp stages share that 22uf B+.
 







noise-wise, the amp is very quiet. I used shielded cables on long runs that carry signal, a ground lift (ala  a modern vox), and the filament CT is raised to the cathode...

[EL34]

Looks good, nice job

[Willabe]

Looks nice!


         Brad   

[terminalgs]

I think I'm done tinkering with the circuit for now.  It's buttoned up. 


I decided to change the circuit that connects the extra channel triode to either the ef86 or the 12ax7 via a switch.  The phase cancellation Sluckey pointed out was too overwhelming when both channel's volumes were turned up.  I could fiddle with the tone, to get the opposite effect and produce interesting sounds, but in the hands of anyone other than me, they'd probably take the amp in to be serviced...


So, S1 flip flops each channel between a 1M grid leak and the extra gain stage.  You can see on the schematic I tweaked R8 & R12 for the 12AX7 channel and R42 & R10 on the EF86 channel to adjust the gain coming from the extra stage.  That arrangement of the switch resistor was something I found on early Mesa-Boogie amps with an extra gain and switched input jack. 100K for R42 and R12 seem to be low enough to keep the channel quite if nothing is plugged in anywhere.  If you don't use the extra gain stage,  turning the volume down on it quiets it down.  I think the MB's used a double switched jack, or a stereo jack with both tip and ring switched.  anyhow, I'm happy with what I got for now.





the blue Sturdite vinyl cloth is actually a lot darker under interior lighting.  the Sun light brings out the blue.