Peculiar Fisher circuit

[SpreadSpectrum]

The one on the right - the purpose of the 100K B+ to cathode is to act as a ~2.8 milliamp current source (basically independent of signal), widening the range of cathode swing that the input signal will incur in the course of amplifying the signal. Again, it would be instructive to see what is leading TO the grid of this stage. I'm betting it isn't the very first stage, but a subsequent mild-amplication middle stage.

I assume you mean 'left' on this. It seems like there would be less cathode swing if you use a smaller resistance and dump extra current. I am going to agree with Ketje that this was done to increase gain by decreasing degeneration (but not eliminating it).

The first two stages have resistors from B+ to cathode. These are gain stages for the tone controls and balance controls.

 

[SpreadSpectrum]

The circuit on the right kind of bothers me still since cathodynes function best when loads on the anode and cathode are equal. Obviously the loading will be unequal with a resistor from B+ to cathode.

Of course, there will always be differences in the gain of the output tubes, and Fisher's policy in their owner's manuals was for you to take it to a shop and have the tech null distortion on a distortion analyzer when adjusting these pots.

It should be noted that this resistor was eliminated in the X-101-C. Maybe it was just a case of an engineer being "too clever."

 

[GoatGuy]

I am afrait we are'nd there at all 😛
The balance of the splitter has to be adjusted so that the total resistance in the anode- an cathode-circuit is the same.The operating point is imposed by the voltage applied to the grid and comes from ??
The 330k takes some current from the cathode resistance leaving less for the tube and thereby raising the anodevoltage,giving more space for the tube.
As for the left picture,a smaler cathoderesistance doesn't widening the range.On the contrary the possebility of the cathode following the input is become smaler (needs more current).
Mona

LOL... as you said, not QUITE right.

Actually - when you figure in the dissimilarities in transconductance of real, in the socket output tubes ... the balance not only does not "need" to be exact, but should not be exact - at the phase splitter. The "exactness" should be measured at the plates of the output tubes, or, at their cathodes, using an oscilloscope. When a sinusoidal signal is symmetrically deviating in a reciprocal and equal fashion ... or as close as the phase-splitting adjustment affords ... well, then that's as good as it gets.

Sorry... this is the difference between "perfect-world" circuit design (which we ALL engage in), and real-world devices. Real-world are far, far, far from ideal, and pretty-damned-far from even "somewhat ideal". From batch-to-batch, from lot-to-lot, from year-to-year and from plant-to-plant.

I remain unbudged.

GoatGuy

 

[Ketje]

The circuit on the right kind of bothers me still since cathodynes function best when loads on the anode and cathode are equal. Obviously the loading will be unequal with a resistor from B+ to cathode.

Of course, there will always be differences in the gain of the output tubes, and Fisher's policy in their owner's manuals was for you to take it to a shop and have the tech null distortion on a distortion analyzer when adjusting these pots.

It should be noted that this resistor was eliminated in the X-101-C. Maybe it was just a case of an engineer being "too clever."

The loading unequal with the 330k ? I think not.
Anode-side 47k.Cathode (27k+Rp)//330k ,for AC, B+ and GND are the same.Adjust Rp to get 47k (almost 30k).Naturaly you can create an unbalance on purpose to compensate for an unperfect final stage.
What do we get without the 330k.(asume Rk=47k and some rounding to make calc.simple)
Vg=105V -> Vk=100V -> Ik=Ia=100/47=2mA -> Va=320-(47x2)=220V
Voltage on the tube 220-100=120V,not much for a 12AX7 .
Now we put in the 330k.Vk=still 100V but the 330k feeds (320-100)/330=2/3mA in the Rk leaving 2-2/3=1.33mA for the tube.Voltage-drop on Ra 47x1.33=65V ->Va=255V .Voltage on the tube 255-100=155V,that's better 😀.
Mona

 

[GoatGuy]

I want to meet Mona. Well done analysis - 😉

 

[SpreadSpectrum]

Mona,

You are of course correct that the pot can be set (near its limit?) so that there is no imbalance. The question is, what does the 330k resistor get us that changing the operating point and leaving the 330k resistor out can't?

In reality, they probably didn't want to increase B+ or change the operating points in other stages for whatever reason so this was probably a band-aid fix, unless I am missing some other compelling benefit this resistor gives. If it did, you would think that you would see it more in cathodynes. As I said before, the next revision of the amp eliminated this resistor.

It's an extra part and makes the solution less elegant.

 

[SpreadSpectrum]

Actually - when you figure in the dissimilarities in transconductance of real, in the socket output tubes ... the balance not only does not "need" to be exact, but should not be exact - at the phase splitter. The "exactness" should be measured at the plates of the output tubes, or, at their cathodes, using an oscilloscope. When a sinusoidal signal is symmetrically deviating in a reciprocal and equal fashion ... or as close as the phase-splitting adjustment affords ... well, then that's as good as it gets.

Sorry... this is the difference between "perfect-world" circuit design (which we ALL engage in), and real-world devices. Real-world are far, far, far from ideal, and pretty-damned-far from even "somewhat ideal". From batch-to-batch, from lot-to-lot, from year-to-year and from plant-to-plant.

Yes, systems engineering is important. Don't design the subsystem in isolation from the system.😉

 

[Ketje]

Less elegant +1
But what to do if the voltage on the inverter is to low?
Ra has to be Rk so you lose 2xVk from B+.If that's to low three solutions.
Up with B+
Down with Va previous stage
Or 330k
Mona