Using both inputs on a LTPI for a parallel preamp design

[guitarboy58]

I've recently built two amps where two different preamp configurations are blended into a single input to the PI.  I've seen similar designs in this forum.  I have two great sounding designs, one with parallel preamps made from 12AX7's and the other using parallel EF-86's.  Now I'd like to experiment with an EF-86 paired with a 12AX7 or 12AY7 configuration.  I've seen several designs where this tube configuration is blended dirrectly and designs where they are switched, but they always feed the same input to the PI. 
Having said all this, my question is: why don't I see designs where the two preamp configurations feed different sides of the phase inverter?   I think this would eliminate the loading concerns for the EF-86, allowing you to have different tone control designs for each preamp.  I've seen schematics of Bad Cat amps where half of a 12AX7 feeds into the second input of the LTPI and I have to assume someone in the world has run an A/B box to this amp with both channels active without issues.
I would think you need to pay attention to the phase relationship between the two signals, but other than that, are there concerns, drawbacks, etc to this configuration?  Is there a reason not to pursue this idea?  I'm looking for feedback about the electrical implications and the sonic quality. 
Thanks for sharing your knowledge and experience.

[sluckey]

This works quite well...

[guitarboy58]

Wow.  That's almost exactly what I was envisioning.  I've not seen that schematic before.  Thanks for sharing it.  Let me ask a follow up question; would adding a cathode follower to the EF-86 channel to drive a more complex tone circuit alter the behavior of the PI with signal going to both inputs? In otherwords, will I get any cancellation of signal if the phases are not in a specific orientation?  It may be a stupid question and I should probably know the answer, but I just can't seem to convince myself that it will not create some weird signal cancellation if things are not in phase.  Time to get out the text books I guess.....

[punkykatt]

Sluckey, any chance you might have a layout  for that amp?

[SILVERGUN]

Nice schematic sluckey  

I have to assume someone in the world has run an A/B box to this amp with both channels active without issues.

Anyone have any thoughts on this comment/question?

Am I seeing this correctly (see pic)....and does it show us that "yes, you can do that, because the phase relationship works" (in this case)

And the fact that the cathode follower is non-inverting, means it won't add to the dilemna of phase relationship.....right?

Sorry guitarboy,,,I'm learning too,,,,,good questions

[sluckey]

Works fine on my amp.

[Boots Deville]

As sluckey's Matchless schematic above is drawn, you could plug you guitar into the 12AX7 channel, run a cable from the preamp out  into the input of the EF86 channel with the result of running the channels cascaded.  Optionally you could run some effects in between using the second channel as recovery.

Anyone ever experiment with anything like that?

[sluckey]

Am I seeing this correctly (see pic)....and does it show us that "yes, you can do that, because the phase relationship works" (in this case)

Yes. Since the PI is really a differential amp (meaning one input is inverting and the other input is non-inverting) you will need an even number of inversions in one preamp and an odd number of inversions in the other preamp. This only applies if you will feed the same signal into both inputs simultaneously. However, if you will play one guitar thru one channel and another guitar thru the other channel, the number of phase inversions inside the amp are meaningless because there is no phase relationship between the two guitars. They are totally non-coherent.

[sluckey]

Sluckey, any chance you might have a layout  for that amp?

I do but I doubt you would like it. This amp began as a November amp (18W plexi). Then I replaced the board with a Lightning board. I kept the same size board so it would fit the original amp perfectly. There's some wasted space on the main board. After playing the Lightning for a while, I realized I didn't really care for the sound, so I added the EF86 preamp since this was a popular mod for an 18W amp. I built the EF86 preamp in a small Bud box and bolted the box onto the top of the main chassis. Hardwired filaments, B+, and a signal cable to the PI.

So, I have a layout but it is scattered. I can pull it all together if you still want to see it.

The Trinity TC-15 is almost identical to my amp. Everything is on one board so it looks better.

[punkykatt]

thanks sluckey, i will check out the Trinity TC-15.

[sluckey]

The Trinity Amps site is well done. There are some great sound clips for the amps...

http://www.trinityamps.com/Product_Trin15.htm

There is also a nice forum. If you join you'll have access to schematics, layouts, etc. for the TC-15 as well as many others via the Resources forum.

[HotBluePlates]

... my question is: why don't I see designs where the two preamp configurations feed different sides of the phase inverter?  ...  I've seen schematics of Bad Cat amps where half of a 12AX7 feeds into the second input of the LTPI and I have to assume someone in the world has run an A/B box to this amp with both channels active without issues.
I would think you need to pay attention to the phase relationship between the two signals, but other than that, are there concerns, drawbacks, etc to this configuration?  ...

There's one main reason you don't see it on most production amps: output stage feedback.

Typically, the output stage has negative feedback running from OT secondary back to the "other input" of the long-tail inverter. That precludes that side from also being used the way you're wanting, as the feedback and the guitar signal at that side would be in phase and cause the amp to howl.

Therefore you usually only see channel signal to each side of the long-tail inverter in no-negative-feedback designs, generally in Vox amps or those inspired by Vox (like Matchless, Badcat).

[sluckey]

The very popular Marshall 18 Watt amp feeds one preamp to one PI input and feeds the other preamp to the other PI input.

[guitarboy58]

Thanks to everyone who responded.  While I rarely comment, I read this forum every day.  I've learned a lot over the years.  So again, many thanks.
If I've followed the comments correctly, my original concerns about phase orientation are correct; the two preamps need to be opposite phase polarities to avoid issues with the PI inverting and non-inverting inputs.
And you have to give up the negative feedback loop. 

[tubeswell]

Am I seeing this correctly (see pic)....and does it show us that "yes, you can do that, because the phase relationship works" (in this case)

Yes. Since the PI is really a differential amp (meaning one input is inverting and the other input is non-inverting) you will need an even number of inversions in one preamp and an odd number of inversions in the other preamp. This only applies if you will feed the same signal into both inputs simultaneously. However, if you will play one guitar thru one channel and another guitar thru the other channel, the number of phase inversions inside the amp are meaningless because there is no phase relationship between the two guitars. They are totally non-coherent.

And so the 'product' of the output of the LTP is the sum of what is occurring at both its inputs. i.e.; both sets of input signals get amplified and split into opposing phases for the PP output stage. (To answer Silvergun's question further)

[HotBluePlates]

The very popular Marshall 18 Watt amp feeds one preamp to one PI input and feeds the other preamp to the other PI input.

Good observation!

I'd contend this one amp is unlike anything else in the Marshall lineup, while many Vox amps have long-tail inverters and no negative feedback around the output stage.

[sluckey]

I'd contend this one amp is unlike anything else in the Marshall lineup

Yep. I think Marshall copied the Watkins Dominator 15 for that one. Most of the other stuff was based on the 5F6A.