designed for a 6.6K Ohms impedance source.

[Steveo]

When trying to get the right OT for an amp should we take into account “sag”voltage? I read that term now and then, i’m assuming its the b+ voltage at the plates and screens of the power tubes when the amp is being pushed pretty hard? This post got me thinking maybe my amps OT 3.2k primary impedance might be a little low for its plate voltage of around 480vdc with two 6CA7. It still sounds good though and no red plating. I’ll have to learn how to draw the load lines for power tubes. Ive only done it for preamp tubes.

[pdf64]

When trying to get the right OT for an amp should we take into account “sag”voltage? I read that term now and then, i’m assuming its the b+ voltage at the plates and screens of the power tubes when the amp is being pushed pretty hard?

Yes, pretty much, but I suggest not to attempt to measure the anode voltage of an output valve that’s being pushed hard, especially into an inductive load, as the back emf voltages may damage the meter; a 100:1 scope probe reveals such voltages may be well over 2kV peak. The OT CT HT node voltage is fine.

This post got me thinking maybe my amps OT 3.2k primary impedance might be a little low for its plate voltage of around 480vdc with two 6CA7…

I suppose 480V is at idle? It’s unlikely to stay that high when high current is being drawn, around 400V is typical.

[Steveo]

Yes, thats at idle. I feel a bit better now, thanks.

[tschwarz]

A link to this highly informative text by O.H.Schade was just shared on another forum (TDPRI):
https://mennovanderveen.nl/cms/images/o ... _Tubes.pdf

Here's the page relevant to this thread's subject showing how OT load and distortion correlate:

[Steveo]

Very informative, thanks. Ive been listening to some videos of 2nd and 3rd harmonic distortion examples. The 2nd has a very warm sound alone where the 3rd almost adds like a clear presence to it. Maybe ken fischer was trying to find that balance between the two with his different impedances?

[martin manning]

Note that the page posted above is a Class A example, not AB push-pull.

[Steveo]

I did notice that, im not educated enough to know if there would be a difference with AB.

[tschwarz]

Note that the page posted above is a Class A example, not AB push-pull.

Yes. The P/P configuration will cancel out even harmonics. As those are more favorable they need to be reintroduced by deliberately throwing the P/P couple out of fully symmetric operation - eg. asymmetric driving from the PI or asymmetric biasing. Peavey even filed a patent for this - I believe.

The fundamental operation of the pentode along its load line is the same though, for class A and A/B...

[Addendum - please ignore if you already were aware of this resource... ]

There's this phenomenal online tool by Giuseppe Amato: https://www.vtadiy.com/loadline-calcula ... alculator/ that lets you interactively play around with all the variables in a P/P stage (HT, bias, screen voltage, OT impedance, ... and observe the effect on the load line.

[Steveo]

Ive been playing around a lot with this load line calculator. But im not sure what I should use for a correct reactive load resistance. I select EL34 pp pentode, for V+ ill enter my plate voltage 480. For Iq, ill enter my bias point 32Ma, but for reactive load Im not sure. Do they want the primary impedance of the OT? When i enter 3200 my load line goes way above the 25watt curve.

[pdf64]

I suggest you use the sagged HT voltage, rather than the level at idle.

[tschwarz]

Ive been playing around a lot with this load line calculator. But im not sure what I should use for a correct reactive load resistance. I select EL34 pp pentode, for V+ ill enter my plate voltage 480. For Iq, ill enter my bias point 32Ma, but for reactive load Im not sure. Do they want the primary impedance of the OT? When i enter 3200 my load line goes way above the 25watt curve.

OK, did as pdf64 suggested and set HT to 400V, stayed slightly below that for screen voltage (voltage screen to cathode, I'm assuming cathode biasing) and used the OT's 3.2k reactive load.

loadline_EL34.png

Looks perfect to me. The intersection with the 0V grid voltage curve is right in the knee. Since this tube only conducts during half the cycle (push-pull) and your bias current at 400V stays way below 60mA there's no problem with the B-part of the load line running above the pmax curve...

[Steveo]

Ok, that looks a lot better. Thanks very much. With changing speaker impedances and voltages while playing im imagine this line is moving a lot. I know i like my amps tap set at 8ohm into the 8ohm speaker load (3200ohms) better than when i set my amps tap to 4ohm into the same 8ohm speaker load. If im correct when i do this my amp is seeing double the primary impedance, so it would be roughly 6400 ohms. The amp gets darker and quieter with the 6400ohm load. When i went the opposite to the 16 ohm tap of my amp to the same 8ohm speaker load just the opposite happens. It gets brighter and louder, again if im correct that would be around 1600ohms reflected back to the amp.

[martin manning]

The fundamental operation of the pentode along its load line is the same though, for class A and A/B...

I don't agree with this statement. p-p load line actually looks like the composite picture below.

The P/P configuration will cancel out even harmonics. As those are more favorable they need to be reintroduced by deliberately throwing the P/P couple out of fully symmetric operation

Not necessarily. There are some p-p guitar amps where measures are taken to minimize the second harmonic. There is lots of discussion around that in the Dumble section of the forum.

[tschwarz]

Hi Martin.
Thank you very much for engaging in this discussion!

The fundamental operation of the pentode along its load line is the same though, for class A and A/B...

I don't agree with this statement. p-p load line actually looks like the composite picture below.

But the upper tube will shut down while the lower tube swings towards full conduction and vice versa. How quickly this shut down happens, i.e. how much overlap there is, depends on the slope of the load line, i.e. the OT load. Smaller load = steeper load line = less overlap. The overlap is out of phase by ~180° so the signals cancel out at least partially - in the OT. Also, the load each tube actually sees changes when the idle one no longer runs current through its half of the primary.

In the context of the original discussion about optimal load versus HT versus distortion each tube still follows the same fundamental physics as in SE class A - no ? The mechanics of tube distortion still apply - per system. The upper system's behavior is not directly changed by the presence of the lower system. Mangling of both signal halves with cancellation of even harmonics and generation of crossover distortion occurs in the OT ?

[martin manning]

In Class A the grid curves and therefore the plate current vs. grid voltage expands on the positive swing and compresses on the negative swing, which gives rise to the distortion vs. plate load curves shown in the diagram above. In p-p the grid lines are expanding on both the positive and negative swings, so one would expect different distortion vs. plate load characteristics. My point is that those curves apply to Class A, not to AB.

AB load lines involve A to AB transition and alternating cut-off on each side as mentioned, so I would say in operation they are quite different.