Push-pull OT phasing

[thorpej]

I'm a little confused as to how to think about the phasing of the Hammond 1650HA OTs I'm using in my 6L6GC ultra-linear amp. As I understand, the dots on a transformer schematic indicate the start of the winding. But I seem to recall that how the windings are oriented towards one another can also impact whether the transformer is inverting or not.

For NFB sent back to the cathode of the first state in the loop, you want the signal to be in-phase with the grid, right? (Because an inverting gain stage amplifies the difference between the grid and the cathode.)

My amplifier looks like the Mullard 5-20 or Eico HF-89... Gain stage, LTP phase inverter, P-P output stage... Viewed from the perspective of the "top half" of the output stage, it's 3 inversions... so: UP (input signal) - DOWN (output of gain stage), UP (output of "top half" of LTP), DOWN (output of "top" output tube). Given that, and the attached snippet from the 1650HA data sheet, I think I want to attach the blue / blue-yellow leads to the "top" output tube so that the output of the transformer ends up in the UP direction for the feedback tapped off of the 16 ohm secondary tap.

Sound right?

 

[gadget73]

Honestly it bends my head too much to try and work out all the phase inversions between the transformer and whatnot. I just hook stuff up and see if it screams. If it does, I reverse the plate leads.

 

[battradio]

It depends on the manufacture of the transfomer as much as anything , it's a 50-50 chance so i just check every time to make sure it's not oscillating .

 

[ferninando]

just run a 1k ac from a funct gen thru the opt and compare in to out for phase inversion.
Bobs your uncle.

 

[thorpej]

just run a 1k ac from a funct gen thru the opt and compare in to out for phase inversion.

Yah, easy enough I suppose.

 

[ferninando]

even a variac will do, at low voltage 60~ 1 volt or more.

 

[GordonW]

You don't even need a scope. An AC voltmeter will do.

Connect the primary plate leads to the output of a variac. Connect ONE of the secondary leads to the junction of one of the primary leads and the wire from the variac (doesn't matter which one).

Turn on the variac, and measure the voltage across the primary. Set to 50v or so (anything convenient).

Then, move the test lead that's connected to the junction between the variac, the primary, and the secondary, to the unconnected lead of the secondary. Leave the other test lead on the other side of the primary.

If the voltage goes UP from the primary read alone, the primary and secondary are in the SAME polarity (i.e, the (-) of the primary is connected to the (+) of the secondary, or vice versa- they're wired in series, in the same polarity).

If the voltage goes DOWN, the primary and secondary are out of polarity (you have both (+) leads or both (-) leads connected together).

Works every time.

Regards,
Gordon.

 

[GordonW]

Or, if you want to check the OPT in circuit- you do need a scope- but you can disconnect the feedback loop, and run a small AC signal into the amp. Connect one side of a dual-channel scope to that input. Connect the other channel to the output. It they are in phase on the scope, then you have connected everything OK. If they are out of phase, then reverse the OPT primary leads.

This works, because ALL amps with negative feedback to the cathode of the FIRST tube stage (i.e, the feedback goes to the FIRST tube- there's no preamps or anything before the feedback stage), ALWAYS have the output transformer in phase with the input- because input to the cathode of the input tube (where the feedback goes), is out of phase with input that is put into the input grid. If you put the same polarity into the grid and cathode, they will cancel out- which is by definition, negative feedback.

Regards,
Gordon.

 

[ferninando]

many pathes to the same end

 

[Palustris]

As I understand, the dots on a transformer schematic indicate the start of the winding.

The dots indicate the phase relationship. So, the primary dot is in phase with the secondary dot. Connect the secondary dot winding to ground and use the other end of the winding for NFB.

 

[kward]

My amplifier looks like the Mullard 5-20 or Eico HF-89... Gain stage, LTP phase inverter, P-P output stage... Viewed from the perspective of the "top half" of the output stage, it's 3 inversions... so: UP (input signal) - DOWN (output of gain stage), UP (output of "top half" of LTP), DOWN (output of "top" output tube). Given that, and the attached snippet from the 1650HA data sheet, I think I want to attach the blue / blue-yellow leads to the "top" output tube so that the output of the transformer ends up in the UP direction for the feedback tapped off of the 16 ohm secondary tap.

Sound right?

That does sound right. But the plate and UL leads are the very last thing I hookup, just in case they are wired backwards coming out of the transformer. I've never actually experienced miswired leads on a Hammond, but I have seen it on Edcor for example.

 

[trobbins]

In circuit test can be done by opening the feedback path (disconnect the feedback link/resistor from the OT secondary). For a low input signal, then measure or listen to the OT output signal voltage - keep the output level low so that it is well below max signal level. Reconnect the feedback path but insert a large extra series resistor in the feedback path, so that the feedback ratio is made very small. Use the same test input signal level, and the OT output signal will be either increased or decreased just a bit - you want it to decrease to confirm that negative feedback is being applied.

 

[ferninando]

As I learned it in the USNavy, dots indicate phase.