Hello, in very few diagrams I have seen this capacitor from the input grid to the cathode (C9 from 22pf to 330pF). Since, to my surprise, in the simulations it widens and flattens the frequency response (if it were grounded it would cut it), I wonder if it is part of the feedback network (negative or positive? maybe part of GNF?) and what the possible benefits and values suggested are. Thank you for your possible clarifications.
Attachments
Yes it is a form of feedback. Not GNF - where would the global loop be?
Since the grid and cathode move with the same phase, it is positive feedback.
Positive feedback, in itself, will promote instability.
In contrast, the capacitance between anode and grid is negative feedback, because the anode and grid move in opposite phase (one goes up, the other goes down, and vice versa).
It lowers the gain, because decrease in anode voltage works, through Cag, against the increase in grid voltage that caused the anode drop in the first place.
Jan
Since the grid and cathode move with the same phase, it is positive feedback.
Positive feedback, in itself, will promote instability.
In contrast, the capacitance between anode and grid is negative feedback, because the anode and grid move in opposite phase (one goes up, the other goes down, and vice versa).
It lowers the gain, because decrease in anode voltage works, through Cag, against the increase in grid voltage that caused the anode drop in the first place.
Jan
Not GNF - where would the global loop be?
GNF is applied from the secondary of the output tranformer (via a RC network not shown in the schema) and between Rk1-Rk4.
Rk4 is shorted by the 220uF cap, hence C9 could be connected at any of both sides of Rk4 without visible effect in the simulation.
GNF is applied from the secondary of the output tranformer (via a RC network not shown in the schema) and between Rk1-Rk4.
Rk4 is shorted by the 220uF cap, hence C9 could be connected at any of both sides of Rk4 without visible effect in the simulation.
It is local positive feedback ( higher gain ) at high frequency , because the cap is small . The cap should have just the minimum capacitance to compensate for other loses in high frequency response . Probably it is used just because 6n2p is not a good choice in the first place for a SRPP and any load would kill high frequency fast .
I am potentially incorrect but, I think the responses thus far have been off the mark.
I have not LT-Spiced this, but my brain spice tells me that this cap will roll off high frequencies. This cap connects grid to cathode at high frequencies. Therefore, there will be no signal to drive the Gm of the tube at high frequencies. The tube functions as a constant current source at high frequencies.
I believe this cap is used to block out of band signals from being demodulated by the tube. Think RF noise, switching power supplies, cell phones, noisy mains etc.
Please correct me gently if I have this wrong. I may need to update my brain spice software.
I have not LT-Spiced this, but my brain spice tells me that this cap will roll off high frequencies. This cap connects grid to cathode at high frequencies. Therefore, there will be no signal to drive the Gm of the tube at high frequencies. The tube functions as a constant current source at high frequencies.
I believe this cap is used to block out of band signals from being demodulated by the tube. Think RF noise, switching power supplies, cell phones, noisy mains etc.
Please correct me gently if I have this wrong. I may need to update my brain spice software.
Good point. It depends on the impedances.
As I see it, it will work as positive feedback if there's some impedance at the cathode where the cap can inject some signal.
In this circuit, the cathode is decoupled to the feedback return (impedance unknown) so any signal coming in through Cgk will still increase Vk. So I still think it is feedback.
If the cathode had been totally decoupled, and acted as a ground for AC, then I would fully agree with you.
But good catch anyway,
Jan
As I see it, it will work as positive feedback if there's some impedance at the cathode where the cap can inject some signal.
In this circuit, the cathode is decoupled to the feedback return (impedance unknown) so any signal coming in through Cgk will still increase Vk. So I still think it is feedback.
If the cathode had been totally decoupled, and acted as a ground for AC, then I would fully agree with you.
But good catch anyway,
Jan
Normal RC filter that cut high frequency is connected to ground , to cathode is positive feedback , but is rarely used and could create confusion .
In fact , as I said , the only reason they used it is because 6n2p like ECC83 is not good for this stage ...
In fact , as I said , the only reason they used it is because 6n2p like ECC83 is not good for this stage ...
That's quite clever, it will reduce open loop gain and could be used as the dominate pole, but much less effect on frequency response than if placed across the input to ground. It protects both RF on the input and on the speaker cables by preventing rectification of the RF in the first tube. Could also be the beginnings of an oscillator too. One to remember.
That's another way to look at it, yes.
That would have the effect that the input impedance is increased, but looking at the circuit, I doubt that was the intention.
Jan
That would have the effect that the input impedance is increased, but looking at the circuit, I doubt that was the intention.
Jan
I do not understand😕. Post #11 talk about greater stability, others talk about positive feedback and potential oscillator.
You did not give the complete and accurate schematic of the amplifier.
That is important whenever global negative feedback is used.
All the poles in the amplifier need to be considered and analyzed.
Perhaps C9 is a high frequency pole that makes up for the rest of the gain phase of the amplifier, including the output transformer.
Often, there is a dominant pole you can easily spot:
Look at the Dyna Stereo 70.
The input pentode's plate has a series RC network to ground. If I remember correctly, that pole is at 7kHz.
Wow! Who wants the amplifier to roll off frequencies above 7kHz?
Well, the stability of the amplifier Demands it, so that the sum of gain phase through the amplifier does not make the amplifier into a high frequency power oscillator.
And, another good thing to look at on the Stereo 70 is a capacitor that connects from the 'Pull' EL34's screen (and UL Tap) back to the input stage.
Interesting!
Making an amplifier that uses global negative feedback to behave is an art and a science; it requires creativity.
That is important whenever global negative feedback is used.
All the poles in the amplifier need to be considered and analyzed.
Perhaps C9 is a high frequency pole that makes up for the rest of the gain phase of the amplifier, including the output transformer.
Often, there is a dominant pole you can easily spot:
Look at the Dyna Stereo 70.
The input pentode's plate has a series RC network to ground. If I remember correctly, that pole is at 7kHz.
Wow! Who wants the amplifier to roll off frequencies above 7kHz?
Well, the stability of the amplifier Demands it, so that the sum of gain phase through the amplifier does not make the amplifier into a high frequency power oscillator.
And, another good thing to look at on the Stereo 70 is a capacitor that connects from the 'Pull' EL34's screen (and UL Tap) back to the input stage.
Interesting!
Making an amplifier that uses global negative feedback to behave is an art and a science; it requires creativity.
It all has to do with phase margin. If the phase of the feedback is far from 0 degrees at frequencies where the cap is effective, the feedback through the cap to the grid can be negative rather than positive.
I think there a two possible explanations leading to the same result:
a) The cap lowers effective feedback at HF (negative feedback minus positive feedback).
b) The cap lowers open loop gain at HF, also resulting in reduced feedback ratio.
Can anyone falsify one of the theories?
a) The cap lowers effective feedback at HF (negative feedback minus positive feedback).
b) The cap lowers open loop gain at HF, also resulting in reduced feedback ratio.
Can anyone falsify one of the theories?
1) This cap effectively shorts out the input to the amplifier at high frequencies, this is true for both the feedback (-input) and the signal input (+input).
2) Roll off at high frequencies does produce phase lag at high frequencies. This can affect the stability of the closed loop stability. Depending on the rest of the amplifier design, this can either enhance or degrade the response. If the pole occurs at a high enough frequency, it will have little affect at all.
3) There is no net positive feedback.
4) This location for the cap is a good one to block HF noise from entering the amp from either the feedback (speaker terminals) or the main input.
2) Roll off at high frequencies does produce phase lag at high frequencies. This can affect the stability of the closed loop stability. Depending on the rest of the amplifier design, this can either enhance or degrade the response. If the pole occurs at a high enough frequency, it will have little affect at all.
3) There is no net positive feedback.
4) This location for the cap is a good one to block HF noise from entering the amp from either the feedback (speaker terminals) or the main input.
Yes.
I think my 2 alternative explanations above are equivalent and result in lag compensation to increase stability.
I think my 2 alternative explanations above are equivalent and result in lag compensation to increase stability.
Both true, but in addition the input signal is also attenuated. Acts much like an additional pole in the open loop response.
Agreed, though the cap will look rather small to the input signal due to bootstrapping.