Impedance matching in valve circuit?

[Paul Mathew]

I'm curious to understand the impedance matching valve directly coupled to the phase inverter valve in the attached diagram. I would think there should be a coupling capacitor between the HT of the first 12AU7 and the second. Thanks in advance.

[wd40addict]

The concertina type phase splitter is operating its cathode at around 80v, giving plenty of headroom for swing, so it gets its DC bias straight from the previous stage.

[Radio Wrangler]

I don't see any impedance matching valve in that circuit.

It's just an audio amplifier. Each stage has to have adequate capabilities to drive its following stage. It needs to be able to swing the necessary voltage to drive the next stage fully, it needs to be able to source/sink the amount of current needed to be able to drive the next stage.

Rarely does this amount to impedance matching.

Matching is a very special term for when the output impedance of a driving stage is exactly equal to the input impedance of the driven stage. It's used in radio frequency systems where frequencies are high enough that stray reactances have to be considered and also line lengths that are becoming significant fractions of a wavelength.

There is a theorem, mathematically provable, that a matched load will extract the maximum amount of power that can be extracted from a given source. People have latched onto this and think matching must be wonderful and efficient, too. It is certainly special, and very desirable in the right circumstances, but the danger is that it can cause trouble under other (and common) circumstances. It's a theoretical maximum based purely on the source impedance. It takes no account of whether the amount of extracted power will damage the source.

Tirade over! Just be very careful talking about matching, it comes with a lot of baggage and conditions.

So the circuit is a simple valve amplifier. Tere is a triode input stage, with the overall feedback loop fed into its cathode circuit, so that the feedback opposes the input signal, giving negative feedback. Triodes are quieter than pentodes, more linear than pentodes if used wisely, but give less gain. Various designers take different choices here.

The second stage is the phase splitter. It is the 'concertina' type and uses another triode. If the grid is swung positive, the cathode voltage will follow it, giving a non-inverting gain of 1 driving the grid of the lower output valve.

In doing so, the current in the phase splitter triode must increase because it's driving more voltage across R733, it's cathode bias resistor. The increased current in this valve, pulls more current through R732, pulling the anode voltage lower and feeding a phase inverted output to the grid of the upper output valve.

There are a few flaws in this circuit. It can't drive the power valves into any grid current because its output is a bit feeble, so it would need cathode followers to drive aggressive class AB2 output stages for an amp where power was everything.

Also the output impedance of the bottom side drive is quite a bit lower than the output impedance of the top side. This is significant when you consider that it has to drive stray capacitances to fairly high frequencies. This causes issues with loop stability, ring, along with an imbalance between the bandwidth capabilities of the two paths reaching the output transformer, as explored in a recent thread.

If you want to push an amplifier to get the best distortion, flatness etc performance, these things need to be gone into, but on the other hand many thousands of people have had satisfaction from equipment not so thoroughly designed.

David

[Paul Mathew]

Thanks for the long and short of it. Very interesting.

[kalee20]

Originally Posted by Radio Wrangler

The second stage is the phase splitter. It is the 'concertina' type and uses another triode. If the grid is swung positive, the cathode voltage will follow it, giving a non-inverting gain of 1 driving the grid of the lower output valve.

In doing so, the current in the phase splitter triode must increase because it's driving more voltage across R733, it's cathode bias resistor. The increased current in this valve, pulls more current through R732, pulling the anode voltage lower and feeding a phase inverted output to the grid of the upper output valve.

There are a few flaws in this circuit.... the output impedance of the bottom side drive is quite a bit lower than the output impedance of the top side. This is significant when you consider that it has to drive stray capacitances to fairly high frequencies.

That's true! But rarely a problem... if both sides are loaded with the same amount of capacitance, things even out. Loading the top side causes a droop to the top side alone. Loading the bottom side with capacitance makes hardly any difference to the bottom side, but it does cause a rise to the top side. If both capacitances are equal, things cancel out.

I mention this because, although probably not what the OP intended, it is relevant to the thread title! If the impedances loading each output of the concertina are matched to each other - equal R, equal C - everything is rosy. It's probably one of the few cases at AF where impedance matching is important!

[Sideband]

Originally Posted by wd40addict

The concertina type phase splitter is operating its cathode at around 80v, giving plenty of headroom for swing, so it gets its DC bias straight from the previous stage.

Yes...notice the cathode of the ECC82 is at 80V and the grid directly connected to the anode of the previous stage is at 75V making the grid of the ECC82 5V negative with respect to it's cathode.....As far as the valve is concerned it has a nice negative voltage on its grid irrespective of the fact that it's grid is 75V positive (with respect to chassis).

[Radio Wrangler]

Originally Posted by Sideband

....As far as the valve is concerned it has a nice negative voltage on its grid irrespective of the fact that it's grid is 75V positive (with respect to chassis).

This works just fine, but be careful to check allowable k-h voltage in the spec sheet of any other valve type for this application.

David