Selecting an OT for a 1 watt design.

[jazbo8]

Here is a mock-up load line using the Hammond 125 with the highest RL of 27k, one side of the PP shown only:

[Mick Bailey]

I use an 12AU7 in PP for the power amp. B+ is 230v obtained from a 555, IRF740 and home-made inductor. The amp powers off a 12vdc laptop supply. Preamp is a 12AX7. Tone is astonishing, and plenty of volume through a G12H. Not much clean headroom.

Transformer part number is 210-6475 from RS components (UK) don't know if there's an equivalent in the US. There's a lot of flexibility in transformer choice with small amps running off a 12AU7. Mismatches don't have too much an effect on tone or volume. I've also used small mains transformers for OPTs and they can work fine as well.

[tubeamptech]

Jazbo8, not sure how you derived this.

[jazbo8]

Jazbo8, not sure how you derived this.

Which part? It is the load line for Class AB operation, with Ep=275V, Eg=-14V, Class A load line = 27k/2 = 13.5k (thin blue line), Class B load line = 27k/4 = 6.75k (turquoise line), the purple line represents the composite Class AB load line.

[Tooboob]

I have a Hammond 125c, and 125E on the way. 8W and 15W respectively. They have multiple primary taps available. I still have to figure out which tap to use. I know it's based on the load resistance of the tube. For instance, the 6L6GC data sheet gives you a load resistance in a push pull AB config averaging 5K, depending on the B+voltage. That lines up fairly well with a typical Fender style OT.

The problem is, the 12A_7 family of tubes don't give you that value. They only give you the plate resistance spec. A 12AX7 is listed as 62.5K. My gut tells me that is going to be too high, even when paired in a PP config. The 12 AU7 is listed as 7.7K. That sounds about the right value, after I do the calculation, which I haven't figured out yet. I know the Marshall 1 watts series are using the 12AU7.

You've been confused. Tubes are made of mechanical conductive elements with contacts for external connections, and there's a bunch of electrons flying through space in there, sort of with a mind of their own, and you want to bend them to your will.

In order to wrap our brains around it, we can employ a simplistic model for the tube's function that is vaguely related to the tube designer's intentions. We can imagine a voltage source, and pretend that there is a constant, mu, where grid voltage times mu is output voltage. It would be really cool if that were true. We'd have a very pure gain element. For small changes in grid voltage, it kind of works, but it is more accurate if we put a resistor, usually called the plate resistance or Rp, between the voltage generator and the plate. It's the "plate resistance" in the tube spec. This covers the fact that the tube can't suck infinite current, or we'd replace it. This is the small signal voltage model for a triode. We can also do a current model. We can imagine a current source, controlled by grid voltage, and imagine a constant, gm, called transconductance, that is the ratio of the change in current for a change in grid voltage. Once again, this model is very innacurate, even for small changes in grid voltage, without a resistance to the cathode from the current source output, and it's the same darned Rp.

This yields a current model and a voltage model for the triode. The models are about as primitive as you can make them, and they're only accurate for small signals for a set of stated conditions. It needs great refinement to yield accurate results in other applications, and there's lots of nerdy discussion on the web, since it would be very nice to be able to accurately simulate tubes. You can see how mu and gm vary in the last graph of my previous suggested link, the 12AX7 datasheet at drtube.

So a datasheet will tell you that a 12AX7 has a mu (transconductance of 100 and a plate resistance (Rp) of 62.5K. They show a circuit with a 1K cathode resistor and a 100K resistor from the plate to the supply (that's not Rp. it's part of the load). They tell you that gm is 100, and the small signal voltage gain is maybe 60. Why isn't it 100? It's because of the Rp. It's dropping voltage from the current. It's in series with the 100K resistor on the plate, and it's dropping 62.5K/(100K+62.5K) of the voltage.

So don't confuse the plate resistance in the spec with the resistance of the load. 12AX7s have a comparatively high plate resistance, so to get good performance, you need a light load, 100K+. That other 40% of the voltage gain that the plate resistance accounts for is dissipated as heat in the tube. 12AT7, AU7, AY7, etc. aren't just crummy 12AX7s they renumbered to sell in spite of their lower gm. Plate resistance is lower (sometimes maybe 4 times lower) in these tubes, and for a heavier load, they can actually provide more gain and more power, and more of the total power dissipated always gets to the load.