What's happening in this plate circuit?

[Groover]

Forgive me tubeswell, but the point is that a) there is a transformer coil in parallel with the plate resistor, and b) I don't know how to account for that when considering the total resistance/impedance in the plate circuit.

[Steve Conner]

We've been here before, didn't I explain it at the start of the thread?

There's no easy answer to "what resistance should I use for my load line". To a first approximation, the resistance is just the 15k resistor. But the transformer is a reactive load, so the load line would be an ellipse at lower frequencies, and you probably don't know how to deal with that. (I certainly don't.)

The circuit would still work with the 15k resistor removed, but you don't know how to analyze it in that condition. You need to know transformer core losses, magnetizing inductance, self-capacitance and so on, because these now determine what load the plate sees at various frequencies. Fender added the resistor to swamp these non-ideal quantities. In a way, it's valid to say that the resistor was put there to make the circuit easier to design.

However, a 12AU7 has a low plate resistance in its own right. If it's lower than 15k, the circuit might work just fine without the damping resistor, it might even sound better.

Short answer: Just hook it up and see how it sounds! You can't change anything anyway without buying a new interstage transformer.

[Groover]

Yes, thank you Steve. I just re-read the beginning of this thread, and it makes more sense to me now. We've really been around the block on this one!

I'll get to work on wiring the amp for the 12DW7! Stay tuned.

[octal]

A couple of thoughts.

1.) You will need to adjust the cathode resistor value to find the bias centerpoint. It will obviously need to be a different value with a 12AU7 type tube section vs a 12AX7.

2.) You're not married to passing DC through the PI transformer. You can design a conventional R/C amplifier with the 12AU7 and capacitor couple to to the PI transformer. I.e. the plate of the 12AU7 goes to B+ through a chunky resistor, and also is connected to a coupling cap which goes to the primary of the PI transformer. The other end of the PI primary can be grounded.

I think you said that you wanted to use this as a clean amp. If so, the setup in #2 will almost definitely be more hi-fi.

Cheers,

Nathan

[Groover]

Thanks Octal

1.) You will need to adjust the cathode resistor value to find the bias centerpoint. It will obviously need to be a different value with a 12AU7 type tube section vs a 12AX7.

Right, I was planning on doing that, hence my load line question above.

2.) You're not married to passing DC through the PI transformer. You can design a conventional R/C amplifier with the 12AU7 and capacitor couple to to the PI transformer. I.e. the plate of the 12AU7 goes to B+ through a chunky resistor, and also is connected to a coupling cap which goes to the primary of the PI transformer. The other end of the PI primary can be grounded.

I think you said that you wanted to use this as a clean amp. If so, the setup in #2 will almost definitely be more hi-fi.

Of course clean in guitar amp terms doesn't necessarily mean hi-fi clean as far as my needs go.

[Steve Conner]

The setup in 2 is what hi-fi guys call "parafeed".

On the one hand, it keeps DC out of the transformer. But the advantage of this is debatable, it's still a SE transformer with an air gap, so it'll still have low inductance.

On the other hand, it cuts the available signal swing in half because the plate is now loaded by a resistor, not a transformer/choke.

So I think in this case parafeed would give no advantage and might even make things worse. It really only comes into its own when you use it with an ungapped OT and a separate plate choke, IMO.

[octal]

Steve,

I think you're right about the drawbacks of parafeed. I was just thinking it might make it easier to calculate load lines.

Groover,

The Gibson Minuteman uses a 12AU7 driver with about 240V on the plate. The cathode resistor is 1K, bypassed with 5uf. That's probably a good starting point.


Cheers,

Nathan

[Groover]

I switched my tube socket around and installed the 12DW7. Trying different values of Rk I got these results:

If Rk=1k5 Vk=11
If Rk=1k Vk=10
If Rk=835 Vk=8.5

It also pulls the preamp B+ down by about 10v when this stage is biased at Vgk=8.5v

So it looks like I am running about 10mA quiescent current. I did load lines for the 12AU7 based on a Ra=15k, and based on that I was going for a Vgk of ~7.5v, but if I plot the measured quiescent current value on the load line graph it looks like the load line must be quite a bit steeper than the one I drew, so I guess there is a lot of current through the PIT primary. If Va=~255v the load line crosses the plate dissipation curve just above the bias point. That seems dicey to me.

As far as how it sounds I was hot-swapping between the 12AX7 and the 12DW7 (With Rk=820ohms Vgk=1.6v) and there isn't much appreciable difference! I actually think I prefer the 12AX7 if I hear any difference at all.

Suggestions at this point?

FWIW the Musicmaster Bass uses 470R for Rk on V1b biasing the 12AX7 at 1.2v per schematic.

[Groover]

more WTF! for ya

Out of curiosity I took some current readings with my DVM. These figures were obtained connecting the DVM in parallel, I did not insert the DVM in the circuit. These measurements are with the 12DW7, but the readings with the 12AX7 were similar.

Iak=~52ma, Ik-ground=~25ma, Iag=~25ma

So unless I can't accurately measure current flow in the circuit using the DVM in parallel I have ~25ma current flowing between the grid and the anode of V1b.

Maybe it's me, but that seems like too much. I should probably just go practice!

[octal]

Try the 12DW7 with the 15K resistor across the primary of the transformer disconnected? (I'm just proposing Gibson-fying it slightly.)

What exactly is your sonic goal with these mods? More gain?

Nathan

[Groover]

Octal; I suggest you read this full thread if you haven't already to fully understand my goal with this amp. My goal is more gain, not overdrive tone gain, just normal BF gain. The amp just seems wimpy compared to when I had a 5F2a preamp with this power amp. As is it is clean almost up to the top of the volume control range, certainly to about 7. That is very un-Fenderish IMHO.

The other concurrent goal is learning more about amps.

[octal]

What Rk did you do those measurements with? The Gibson schematic shows 8.3 volts across the 1K cathode resistor, so 8.3mA, which is reasonable.

You can't use the "shunt" method with an ammeter in this circuit connected across the primary of the transformer because the transformer has a high DC resistance (you will shift the bias point by essentially shorting across it with the ammeter) and you can't shunt the cathode resistor b/c it sets the bias point. Measure the voltage across the cathode resistor and do an ohm's law calculation to determine the current in the circuit.

[Groover]

What Rk did you do those measurements with? The Gibson schematic shows 8.3 volts across the 1K cathode resistor, so 8.3mA, which is reasonable.

You can't use the "shunt" method with an ammeter in this circuit connected across the primary of the transformer because the transformer has a high DC resistance (you will shift the bias point by essentially shorting across it with the ammeter) and you can't shunt the cathode resistor b/c it sets the bias point. Measure the voltage across the cathode resistor and do an ohm's law calculation to determine the current in the circuit.

I did those measurements with the Rk=835R. I connected the DVM shunt from the plate to the cathode to measure Iak, from the cathode to ground to measure IkR, and from plate to pin 2 to measure Iag.

None of these measurements had the DVM connected across the primary of the PIT. I actually took the measurements not from the pins (except for the grid), but from the eyelets on the board that those pins are connected to, but that shouldn't matter, nothing but wire between points.

So I guess you are saying in this instance the shunt method of reading current won't give me accurate data? Maybe I don't have 25mA grid current!

[octal]

Groover,

Yes, that's correct. Generally, the shunt method is used only in output stages, where there's sometimes a.) no cathode resistor to measure voltage across if it's a fixed bias amp b.) you want to be able to measure the plate current independent of the screen grid current.

With preamp/driver triodes, there's always a cathode resistor (unless we get back to that battery bias thread ) and there's no screen current to worry about... so measure your voltage drop across your cathode resistor & calculate or put an ammeter in series with the cathode.

Cheers,

Nathan

[Groover]

Now Rk=510R and Vk=~7.3 - close enough for me to what I see on the 12AU7 graph.

The amp is still what I call clean up to about 7 or 8. And that is with the mid control which is > 25k ohm full on for the most signal.