Plate loading questions

[db!]

Hey guys, I have been doing quite a bit of reading over the past year regarding SE design, tube biasing, etc but I'm having trouble wrapping my head around the concept of plate loading with a choke/OPT.

From what I understand, a resistive plate load simply limits the current to the tube via resistance. Since a typical resistor is purely resistance and has no reactive component, it's impedance is always constant.

However, when it comes to an inductive/reactive load(OPT/choke), it's impedance varies with the frequency; so how is it expected to maintain a certain impedance in the absence of a signal/when the amp is idling? Aside from the DCR, it's really no more than a wire (at idle); so how does this keep the plate loaded?

Since we are on the topic of plate loading, I thought I'd also ask about CCS loads. For the purposes of discussion, say a CCS allows 10mA on the plate of a tube; how does this allow for current to swing when a signal is applied?

Sorry in advance if these questions sound noobish.

[Wavebourn]

Output AC current flows through the load (the next stage input that is driven by output of this one).

[Frank Berry]

Under idle conditions with no audio, the DCR of the OPT/Choke is the plate load.
The impedance will change in the presence of an audio signal.
As for CCS loads, the current doesn't swing, the voltage swings. This means that the impedance of the CCS swings.

[Wavebourn]

Originally Posted by Frank Berry

This means that the impedance of the CCS swings.

What?

[dsavitsk]

Originally Posted by db!

However, when it comes to an inductive/reactive load(OPT/choke), it's impedance varies with the frequency; so how is it expected to maintain a certain impedance in the absence of a signal/when the amp is idling? Aside from the DCR, it's really no more than a wire (at idle); so how does this keep the plate loaded?

Think of the inherent impedance due to reactance as being a ceiling. The reflected impedance then sets the actual impedance. So, imagine you have a transformer with 40H of inductance and a 25:1 winding ratio. This would translate to a 625:1 impedance ratio which is 5K:8 and 5K of impedance at 20Hz (impedance = 2*pi*Frequency*Inductance). If you load it with an 8 ohm load, at less than 20Hz the reactance is lower than the reflected impedance, so you won't be able to actually reflect 5K which means you'll lose low bass. But, at 20Hz and above, there is enough inductance that you can reflect to 5K, so the tube sees a 5K load. In the case of a plate loading choke, it is only the inductance that matters, so above a certain frequency, the load is very high.

Since we are on the topic of plate loading, I thought I'd also ask about CCS loads. For the purposes of discussion, say a CCS allows 10mA on the plate of a tube; how does this allow for current to swing when a signal is applied?

The CCS sets the DC operating point. As signal raises and lowers the grid, the tube wants to draw more or less current. This adjustment is made through the load (the next stage, or the parafeed transformer.)

This might be helpful: http://www.ecpaudio.com/pdf/parafeed_basics.pdf

[kevinkr]

Originally Posted by Frank Berry

<snip>

As for CCS loads, the current doesn't swing, the voltage swings. This means that the impedance of the CCS swings.

The CCS impedance doesn't change. The internal resistance of the device connected to the CCS is doing the changing. Think of the tube (triode usually) as a voltage controlled resistor (or more accurately transconductance) with a constant current through it. Change the resistance and the voltage dropped across it changes.

[wakibaki]

Originally Posted by db!

However, when it comes to an inductive/reactive load(OPT/choke), it's impedance varies with the frequency; so how is it expected to maintain a certain impedance in the absence of a signal/when the amp is idling? Aside from the DCR, it's really no more than a wire (at idle); so how does this keep the plate loaded?

It's true that the impedance of a choke varies with frequency, but chokes are rarely employed as loads in broadband amplifiers, they are more likely to be found in a narrowband, RF, amplifier.

At idle (DC) the current in the active device is controlled by the biassing, not the load. The load impedance is of significance only at AC (when there is a signal), the resistive component of a loudspeaker driver is multiplied back to the plate through the transformer turns ratio.

As regards the CCS, this is easiest thought of in terms of the limiting conditions. When the valve is saturated (fully ON) all the current flows from the CCS through the valve. When the valve is cut off (fully OFF) all the current flows from the CCS into the load (which is connected at the junction of the CCS and the plate).

w

It's important to realise that there's a certain amount of operator overloading taking place. Sometimes load means the device that gets the power in the end, and sometimes it means whatever arrangement of components is in the anode circuit.

[DF96]

Under idle conditions with no audio, the DCR of the OPT/Choke is the plate load.
The impedance will change in the presence of an audio signal.

No. The DC resistance stays as it is for 0Hz 'signals'. The inductive reactance is set by frequency, not signal amplitude. A 10H choke has an impedance of 6283ohms at 100Hz, whether or not there is actually any 100Hz signal present.

A choke and OPT are not really the same thing, although there are some similarities. If you think they are the same you will get confused. If you think they are completely different you will also get confused!

I think your problem is that you associate impedance with signal. Impedance is unaffected by signal. It tells you what would happen if a signal is present, but with the signal gone the impedance is still there.

[db!]

Thanks for the replies!

The plate choke idea is for parafeed, I am still exploring my options for a SE amp.

I have to admit, I'm still a bit confused; maybe I'm looking at it from the wrong way... I tend to think of biasing in terms of a load line. You choose a plate load, apply a negative voltage to the grid(that corresponds to one of the tube curves) and that should set the idle current. If at idle, the DCR of an OPT was the load, would this not cause the current to shoot up?

[dsavitsk]

with a resistor load, AC and DC load lines are the same. With a choke load, the load line is the choke in parallel with the actual load but this does not correspond to DC conditions which are set by B+ and bias. So yes, if you would use say a 300V supply with a resistor load to get 150V on the plate, then with a choke load you are going to want a ~150V supply. This is explained in the link I posted above.