Using OT "polarity" to determine destination of blue and brown primary wire

[pjd3]

Hello,

So, while I get that it is totally feasible to spark up the new amp build (Sluckey's 6V6 plexi - two in one chassis stereo amp) and have a 50% chance of being correct on which OT wire, blue and brown goes to which power tube so that the NFB ends up being just that, negative feedback, the alternative being positive feedback creating howling and turning your newest amp build into a marvelous oscillator.

But I decided to see if I could make a thought excersise out of this and see about predetermining the correct wiring by logic and speculation.

While I am still partially mystified by all the workings of the long tailed phase inverter, I did determine that knowing whats happening with the output transformer with respect to "primary to secondary polarity" would be an essential key in determining the right destination of the Blue and Brown OT primary leads.

What I discovered was that the speaker taps are in phase with the blue OT primary wire by way of injecting a sine wave into the blue wire, scoping that on one channel of the scope and looking at the 16ohm tap on scope channel 2. In phase sine waves.

Looking at the schematic, it appears that when say, a signal enters the phase inverter (the one with the 82K plate resistor) its polarity gets flipped twice on the way to the OT, once at the PI and then again at the power tube. That tells me that the phase of the signal arriving at the OT is the same as the signal initially coming into the PI. So with that, plus knowing what the phase relationship is of my OT primary to secondary, that should provide the necessary info to determine what wire, either brown or blue goes to the power tube associated with that signal path (PI input and its plate going to the grid of that power tube.

Looking at the schematic, I'm struggling just a bit to understand what the phase of the NFB signal should be with respect to the PI input signal phase. Once I get that straight in my head, I should be able to accurately determine the answer to which power tubes the brown and blue OT primary wires should go to.

Any insight would sure be appreciated. This is actually a good education for me, get tthese phase relations sorted out withing the PI, and power sectioin.

Thank you!
Best,
Phil D. pjd3

[martin manning]

See here: https://ampgarage.com/forum/viewtopic.p ... 94#p463994

I’ve posted a couple of times on this, but the above link describes testing with the OT installed and wired in.

[LOUDthud]

What most builders get wrong is that the feedback to the "100K" side of the PI is actually in phase with the input signal that is going to the "82K"side.

But that's not NEGATIVE feedback you say. It is because a positive input to the "100K" side SUBTRACTS from the positive signal on the "82K" side. The PI's input is the difference between the grid on the "82K" side and the grid on the "100K" side. (forget the signal going to the tail of the PI for a minute.) If the feedback to the "100K" side was the opposite phase compared to the input of the "82K" side, that would make the input to the PI bigger, that's positive feedback.

So in your case, the Blue wire goes to the tube fed by the "82K" side of the PI.

[Helmholtz]

OT phase is correct when disconnecting the feedback increases gain/output signal.

[pdf64]

The power amp feedback loop block is set up as a non-inverting operational amplifier, signal input to the non-inverting input, feedback signal (ie a sample of the loop output) to the inverting input.
See https://www.electronics-tutorials.ws/opamp/opamp_3.html

[sluckey]

Seems that Martin's test method works fine for the typical LTP PI. But what about a cathodyne or paraphase inverter?

I just plan to be wrong with the NFB phase so I don't trim the primary plate leads until I'm absolutely sure the phase is correct. Even if I got it right the first time I will still reverse the plate leads to be absolutely sure. Once I'm sure, I will trim the OT plate leads. Much faster than doing any tests, at least for me.

[martin manning]

Seems that Martin's test method works fine for the typical LTP PI. But what about a cathodyne or paraphase inverter?

That's left for the student ;^) You just have to work out the phase required at the feedback injection point. The essential step of determining which is the primary lead that is in-phase with the secondary "hot" is the same, and then you can determine which primary lead goes where. It's pretty quick.

[maxkracht]

Even if I got it right the first time I will still reverse the plate leads to be absolutely sure.

This is a wise move. I worked on an amp a while ago with a, mostly, stable positive feedback loop. Actually sounded pretty good with the bit of added gain. I can't remember the details, but think it was working just fine except under extreme settings. Always good to double check.

[pjd3]

Thanks everyone for inputting on this.

While I was initially ready to go forward with simply doing the "50-50" chance approach, I did make an intentional decision to see this through in an analytical approach, just to challenge my little brain and expand my understanding of what is really happening to the signals as they relate to the NFB. I have a pretty good idea of how an amp sounds with or without NFB as I installed a variable NFB on my last amp build, which I like having. I find that fairly useful. (but not on this stereo amp - too much squeezed into one chassis already).

I am likely to try out all the methods recommended here to see how they play out.

That is very helpful to see the PI as a non inverting OP amp. That actually makes the figuring out of this far easier. That tells me that the signal feeding into the 100K side should be in phase with the signal arriving at the 82K side for there to be real NFB. I think I'm getting that right, please correct me if I'm wrong.

And this is really where I wanted to be with this. And I just might reverse the brown/blue wire just to see/hear what happens to the audio. After all, thats where the rubber meets the road on this I believe.

This thread has been very helpful. Thank you all for your generous shares here. I'll be looking at this more tonight on the bench!
Best,
Phil D

[martin manning]

If you use the measurement technique linked above, you know for sure and you get it right the first time ;^)

[pjd3]

Hi Martin,

Yes, I had planned to make that the very first test I do tonight, and the tenet of that test makes alot of sense - just another method for showing the relevant conditions of OT status!

Thank you,
Best,
Phil D.