hi folks
I have seen a topic in Morgan Jones' "Valve Amp" regarding the direct coupling.
Regarding the introduction of direct coupling using the pure resistive level shifter, I understand that it can use two resistors to form a potential divider where one end connected to the preceding stage's anode and the other end connected to negative power supply, the mid point of the divider connected to the grid of next stage.
This method seems very simple and easy to implement but the price paid for is attenuation of gain, almostly a couple of - db.
My problem is: I saw some design using this simple resistive level shifter, a small value of capacitor will be in parallel with the first resistor and I don't get it !
Is this arrangement mainly for the pupose of recovering the attenuated gain caused by the potential divider or some of other reason?
I have seen a topic in Morgan Jones' "Valve Amp" regarding the direct coupling.
Regarding the introduction of direct coupling using the pure resistive level shifter, I understand that it can use two resistors to form a potential divider where one end connected to the preceding stage's anode and the other end connected to negative power supply, the mid point of the divider connected to the grid of next stage.
This method seems very simple and easy to implement but the price paid for is attenuation of gain, almostly a couple of - db.
My problem is: I saw some design using this simple resistive level shifter, a small value of capacitor will be in parallel with the first resistor and I don't get it !
Is this arrangement mainly for the pupose of recovering the attenuated gain caused by the potential divider or some of other reason?
Attachments
I think that this is not a direct coupling circuit anymore. AC goes thru the 22nF capacitor just like in ordinary RC-connected circuit.
In case of semiconductors direct coupling useful and popular technique but for tube gear that is not cost effective anyway good decoupling capacitors or trafos the way to go imho. Jensen foil capacitors are extremely transparent sonically.
For tube OTL circlotron topology might be used.
For tube OTL circlotron topology might be used.
In all my experiments I've liked DC better than caps. I haven't tried transformers yet though. Best cap is no cap.
I think you should post a larger section of the schematic, specifically to the right. It would appear that the circled RC-section is not connected to the direct coupling grid circuit - is there a tube to the right not shown? The pentode that is half visible would get the direct coupling line to it's cathode, not grid.
Just guessing here without a larger schematic section, but it could be that the RC section is connected to the screen of that pentode there.
I think you should post a larger section of the schematic, specifically to the right. It would appear that the circled RC-section is not connected to the direct coupling grid circuit - is there a tube to the right not shown? The pentode that is half visible would get the direct coupling line to it's cathode, not grid.
Just guessing here without a larger schematic section, but it could be that the RC section is connected to the screen of that pentode there.
Yes
The arrangement forms a high pass (shelving) filter with a transition point of about 7Hz, if I calculate it correctly.
As artosalo quite correctly pointed out, signals above this frequency do not really see a DC coupling anymore.
The difference between this circuit and a more standard one --where you would AC couple and then bias the following stage using, e.g., a voltage divider between B+ and GND-- is that in this circuit, you get a 3dB attenuated response down to DC. A "real" AC coupled circuit forms a first-order lowpass filter while this circuit constitutes a -3dB shelving filter.
Kenneth
Thanks for Kenneth's explaination in details.
I ask this question because of a design of a 805SE which claimed itself in fully direct coupling (of course, except the output !) and also found a 845PSE of which, the driver is direct coupled to the grid of a pair of 845 and seems using pure resistive level shifter method but same as this 805SE, a series of capacitors also found in paralell with the potential divider. The circuit is designed by Patrick Turner Audio and can be referred to the following link:
Deep Space 845se55 July 2008
Go back to the piece meal circuit, in my understanding of coupling using capacitor is for the purpose to avoid the DC directly goes to next stage and affecting the bias of next stage. But price paid for is the RC constant and making phase shifted which is no one wanna to see.
Now if we implement the direct coupling but a capacitor still added in, then my thought is : will it affect the phase shifted and so making direct coupling become meaningless?
That is what I worry about !!! Let's discuss .
The following is the original circuit designed by a PRC website:
I ask this question because of a design of a 805SE which claimed itself in fully direct coupling (of course, except the output !) and also found a 845PSE of which, the driver is direct coupled to the grid of a pair of 845 and seems using pure resistive level shifter method but same as this 805SE, a series of capacitors also found in paralell with the potential divider. The circuit is designed by Patrick Turner Audio and can be referred to the following link:
Deep Space 845se55 July 2008
Go back to the piece meal circuit, in my understanding of coupling using capacitor is for the purpose to avoid the DC directly goes to next stage and affecting the bias of next stage. But price paid for is the RC constant and making phase shifted which is no one wanna to see.
Now if we implement the direct coupling but a capacitor still added in, then my thought is : will it affect the phase shifted and so making direct coupling become meaningless?
That is what I worry about !!! Let's discuss .
The following is the original circuit designed by a PRC website:
Attachments
The frame work of the circuit is very simple:
Cascode - direct couple - cathode follower by a power penthode wired as triode - direct coupled to 805 - Output tran.
Cascode - direct couple - cathode follower by a power penthode wired as triode - direct coupled to 805 - Output tran.
Hi,
Actually if you consider the phase, both cases are very similar.
* at high frequencies, the capacitor is the dominant signal path so the phase shift is 90 degrees
* at low frequencies (much below 1/(omega*RC)) the capacitor is essentially invisible so there is no phase shift
* around 1/(omega*RC) it gradually changes from zero to 90 degrees.
So I don't think the phase behavior is that much different.
The DC coupling+cap can be a nice trick if you need to lower the LF response of the amp a little bit (for example, to prevent the OPT from going into saturation, or to improve the stability of the feedback loop). In this case, you only need one extra resistor to do it: the cap and lower resistor are there anyway. I assume this is why it is used in the Deep Space amp.
P.S. correction to my previous post:
Kenneth
Actually if you consider the phase, both cases are very similar.
* at high frequencies, the capacitor is the dominant signal path so the phase shift is 90 degrees
* at low frequencies (much below 1/(omega*RC)) the capacitor is essentially invisible so there is no phase shift
* around 1/(omega*RC) it gradually changes from zero to 90 degrees.
So I don't think the phase behavior is that much different.
The DC coupling+cap can be a nice trick if you need to lower the LF response of the amp a little bit (for example, to prevent the OPT from going into saturation, or to improve the stability of the feedback loop). In this case, you only need one extra resistor to do it: the cap and lower resistor are there anyway. I assume this is why it is used in the Deep Space amp.
P.S. correction to my previous post:
I meant -6dB, we're talking voltages so a 1M+1M voltage divider causes 6dB loss, of course 🙂
Kenneth
Pretty much. One of the problems with regular RC coupled stages is that they can cause instability problems at very low frequencies when NFB is used because the loop is not closed down to dc. dc coupling at least reduces and can in some cases completely eliminate these instability problems. This little tweak gets back much of the signal loss caused by the potential divider but maintains the dc connection that helps stability.
Cheers
Ian
I'm not sure which two cases you mean: with and without the capacitor? Consider three cases:
1. normal cap coupling, 2. DC coupling without cap, 3. DC coupling with cap.
Assuming the right value cap is used then all three have the same phase at audio frequencies - 0 degrees.
At very low frequencies 1. has 90 deg, the other two are 0 deg.
At intermediate freq (low but not too low) 1. and 3. have some phase shift, 2. has none.
3. is a half-way house between 1. and 2. Because it has no phase shift at very low frequencies it can help maintain LF stability, as ruffrecords says.
- Status
- Not open for further replies.