PP output transformer wiring

[dscottguitars]

Hi, is there any way to tell which of the two primary wires goes to which output tube?  I am building a pp 6AQ5 PP amp, no complete schematic used, and using the phase inverter similar to the old bassman amp.  Pins 1 and 6 are the outputs from that to the two power tubes.  The OT has blue and brown wires for the power tubes and rather than hook them up and take the 50/50 chance they are in phase, I am wondering if there is a way to know which goes to which side?

[eleventeen]

Short answer: No. Hook them up one way on a "tack" soldering basis and see if the amp motorboats. If so, flip' em. Probably, if you were replacing a Fender amp OT with the same part number, sure, blue stays blue and brown stays brown. But on a build, not predictable. (Well, I suppose it is, but it's easier to just tack them in place and check it)

[PRR]

No.

I understand the theory. I know how to predict. On my last build I was very careful. Nevertheless, it freaked-out when I connected the NFB. There's just too many +/-/+/- flips for a human mind to keep track of.

Beside which, few OT makers note the phasing and sometimes the actual part in hand did not read the document.

[stratele52]

Short answer: No. Hook them up one way on a "tack" soldering basis and see if the amp motorboats. If so, flip' em. Probably, if you were replacing a Fender amp OT with the same part number, sure, blue stays blue and brown stays brown. But on a build, not predictable. (Well, I suppose it is, but it's easier to just tack them in place and check it)

+1 , this is the only way to do .

[HotBluePlates]

Beside which, few OT makers note the phasing and sometimes the actual part in hand did not read the document.

DScott: You may miss the subtle point here, but to me it's the core problem.

Transformers can invert phase, or not. Complete spec diagrams for transformers include phasing dots, which tell you if you have a positive-going signal  at on point of the primary where the positive-going signal will be on the secondary. In a nutshell, they tell you whether the transformer inverts polarity or not.

But just as soon as you think you've got it figured, you'll get the odd part where the wrong color-code was used on that end of the primary, and the phasing dots on the diagram now lie to you (at least for the part in-hand).

Just leave a little extra lead length on the primary, tack the wires in place and try. If you have it right, shorten as needed and make your final solder connection; if not, flip and make your final connection.

[dscottguitars]

Thanks for all your responses...

It was an old transformer so I have no way of knowing any spec about it, other than it came from a 6v6 amp that was 8 ohm. 

[eleventeen]

The other thing is...if you used NASA-grade computers to predict which way it should be phased...then built the amp...and it motorboated....what would you do? Argue with it?    (not trying to be snide, but OK, maybe some)

[dscottguitars]

Now that leads me to more questions...

My understanding of the OT:

The primary windings start, go around and around and finish.  These are the 'usual' blue and brown wires going to the tube plates.  There is a center tap wire coming from the middle of the windings, if there are 20 turns, the center would be at 10 turns.  This is the 'usual' red wire that goes to the B+ power supply.

The secondary windings do the same but at a ratio that drops the impedance to what it's designed for, like and 8 ohm speaker; the 'usual' green and black wires.

How does the transformer 'know' what the phase is??  And why??  Is the start to the center the 'push' part and then the center to the end is the 'pull'??  If DC current is flowing from the cathode through the tube into the windings out to the B+ and back to ground, how or why should it matter that two AC signals from the phase inverter can be affected in this way??

( I may have my flow direction backwards, I can't remember off the top of my head which it is, but you understand my meaning...)

[guitardude57]

The transformer does not have any idea of phase.
The issue is when you connect the nfb for a presence control, that
connection is or isn't in phase with the PI.  If backwards, it howls.
You simply reverse the wires.
If there isn't a nfb/presence circuit, it may not matter which wires go where.

IN a P-P amp one tube does the negative part of the AC wave, the other... the positive part.

[eleventeen]

"The primary windings start, go around and around and finish....."

Fair enough, but do they wind around the core clockwise or counter clockwise? Do they "start" at the bottom of the transformer on Fenders but start at the top on Stancors and Hammonds? That right there is enough to foil your attempts to predict phase if you are subbing trannies or doing a new build. There is no practical way to predict this!

This is all kind of a moot discussion, because if the amp motorboats when you wire it up and fire it up, you have to flip the primary OT leads no matter what the heck else you have going on. So (to me) it's not worth thinking about. You just wire it up one way and it's either right or it's not, and most amps will let you know in no uncertain terms right away.

[dscottguitars]

Thanks guitardude... That makes more sense of what's going on now.

Eleven,  I got it.  One can't predict it.  That's why I wondered what was going on and why it should matter.  Guitardude answered that for me... And to me it is worth thinking about because I want to know why, not just what.

[HotBluePlates]

How does the transformer 'know' what the phase is?? 

The other descriptions made this more confusing that it should be.

Make it easy; imagine you're looking down on a transformer core that is a straight iron bar, and we're starting the winding of both primary and secondary from that end of the core.

If we wind both the primary and secondary clockwise from that end, then when a positive-going signal is applied to the primary (as judged from the start of that primary winding), then there will be a positive-going output on the secondary (as judged from the start of the secondary, which is at the same end as the start of the primary).

If you wound the primary clockwise, but the secondary counterclockwise, you could change the apparent polarity of the secondary output.

There are many other winding complications that could come into play; if you really want an easy-to-read understanding of how transformers work, review NEETS Module 2.

[dscottguitars]

I'll be more specific:

I can change the leads on a speaker without producing squealing out of control.  The in/out of phase from the primary to the secondary only changes the direction the speaker moves, in or out first.  So it should not matter what primary is hooked up to what tube, since they are alternating the signal, one amplifying the top half of the wave and the other the bottom half. 

But I can 'see' how hooking up feedback, can make it negative-canceling or positive-increasing to infinity or squealing out of control.

Maybe using the term phase was not so correct as feeding back.  That is why I could not understand what the transformer had to do with the squeal and which lead to which tube.  But if I'm taking negative feedback and feeding it to the negative half of the sine wave I can understand why the result is positive.

With this, one could switch the NFB line to the other side of the phase inverter to stop the squeal; YES?

[HotBluePlates]

Essentially yes.

Look at the AA864 Bassman schematic. Start at the phase inverter input (from the preamp) and count polarity switches moving through the circuit.

From the coupling cap to the phase inverter input, I count 1 inversion from at the phase inverter plate, another at the "top" output tube plate. That means the top 6L6 output is in-phase with the input to the phase inverter.

We don't know whether the output transformer inverts polarity, so let's assume it doesn't. Therefore, a positive input at the "top" of the transformer results in a positive output at the "top" of the secondary, as drawn. So the top wire of the secondary would be in-phase with the input to the phase inverter if our assumption is correct.

The negative feedback in this amp runs from the top wire of the secondary back to the tail of the phase inverter, with a 0.1uF cap feeding the "other input" of the phase inverter. We know for an output, the input to this side of the phase inverter would have to be out of phase with the other input of the phase inverter (the one fed by the preamp).

Because we know this is negative feedback and doesn't result in additional signal input to the phase inverter, we can figure that the OT does not invert polarity, and the top wire of the secondary is the same phase as the input to the phase inverter. That's why it needs to be connected to the other phase inverter input to result in negative feedback.

The schematic in the Hoffman Library doesn't include a layout, but by comparison to the AB763 Super Reverb which does include a layout, we find out the 6L6 fed from the input side of the phase inverter is conencted to the blue OT primary wire. We could conclude that the blue OT primary wire should connect to the output tube plate which is fed from the phase inverter section fed by the preamp, with the green OT secondary wire connected to speaker + and the black wire to speaker - and ground; that also tells us that the signal polarity on the blue primary wire is the same as the green secondary wire in the stock transformer.

But that assumes no manufacturer making copies of these transformers gets the wire color coding wrong. It also doesn't account for the AB165 Bassman which we might start with before modding to AA864 specs.

If you look closely at the AB165 schematic, there's a feedback loop from 6L6 plates to grids, and the feedback from the speaker output is connected to the "input side" of the phase inverter, opposite that found in the AA864 Bassman. Again, no layout is included. Knowing what we know from our previous analysis, the brown OT primary wire must be connected to the top 6L6 fed from the "input side" of the phase inverter when using the same transformer color code and phasing found in the AB763 Super Reverb, and that we assumed for the AA864 Bassman.

No matter what the actual colors turn out to be, I can tell you from experience modding an AB165 Bassman that after I changed the injection point of the feedback, I got a full-power howl. I forgot to look at the way the feedback circuit had been changed, and that the different injection point turned the previous negative feedback into positive feedback. Swapping OT primary wires solved the howl (of course, I could have swapped secondary wires instead and gotten the same result).

But most vintage schematics don't include a layout and OT color codes, so you still don't know what the phasing is between any given primary and secondary wire without doing some kind of test, or just trying and seeing if you get positive feedback.

[PRR]

> start at the top on Stancors

Top of the winding? Put the outside turns on first, then the inside turns?

Actually: the winder turns one way. The leads *can* be labeled Start, Tap, and Finish. Put the dot on each Start lead. This gives polarity.

Unless the job is done on two winders and the bobbin gets flipped between.

Or the operator gets careless and mis-marks the ends.

Interestingly, in Power Distribution (the big tranny on the pole at the end of your driveway and larger lumps at the edge of town), the polarity IS standardized and dependable. Sometimes you need to connect two or three transformers in parallel or series. At 25,000-1,000,000 Watts nominal (50 times that in a fault), mistakes are very unfortunate, expensive, maybe fatal.

But small-signal winding is more casual.

> one could switch the NFB line to the other side of the phase inverter

Yes. But most NFB inputs are asymmetrical, signal and NFB. In the simplest, signal to grid and NFB to a cathode. Flip it, and a high-impedance signal is trying to drive a low-impedance cathode, signal is very weak. (The NFB is from a 10W-100W Power source, and can easily drive a low Z or a hi-Z.)

The long-tail does offer symmetry. Not only can you flip the grids, you can cross the coupling caps from driver plates to power grids. This actually gives you more ways to go wrong.

*Usually* it works out best to wire for beauty (clean debuggable layout) leaving the OT leads long enough to switch on test.

There's other ways, using a 'scope or careful summing networks. While this works in principle, in practice it is even more ways to get mixed-up.

> I could have swapped secondary wires

If one-impedance or a 4/16 winding, yes.

With 4/8/16 or similar tapping, you can't swap and still access all the impedances and still be grounded.