In an effort to improve the grounding schemes in my builds I have been attempting to understand more the AC & DC current paths within basic tube amplifier circuits, the main resource being the Tetsu Kimura article translated here: https://ampgarage.com/forum/viewtopic.php?t=220
The image links are obviously all broken on that page, but I have had some success piecing the article back together with the images still found in the original link: http://www.op316.com/tubes/tips/b410.htm
Anyway, my hope is that someone here might be willing to check the correctness of the diagrams I've put together below, which constitute the substance of what I've been able to grasp from those pages. Apologies if this is information easily found elsewhere - I haven't been able to uncover it.
Fig 1. Single stage basic amplifier
This is essentially a duplication of the first example in Kimura's article. Hopefully my photoshop work doesn't prove too unreadable.
The pink highlighting shows the DC current path from the power supply, through the plate resistor RP, triode V1, cathode resistor RK to ground and back to the power supply.
The green highlight shows the AC signal paths. First from the source, through grid leak resistor RG to ground and back to the source. It is my understanding that no current flows in grid stopper RS or the ground between RS & RK, but feel free to correct me on that point.
Further, current flow through RG excites AC current in the triodes grid(?), forming another AC current loop from the plate through output capacitor CO, the load, back to the cathode through cathode bypass capacitor CK. Again, I suspect there's no current flow between RK & CK, but if the circuit does not include a bypassing cap then the current would flow through cathode resistor RK instead (highlighted yellow).
Fig 2. Two stage basic amplifier
Please excuse the slightly unconventional routing of the ground here. It is largely an attempt to indicate how I would intend to make the ground connections if building this in real life. Otherwise this image just demonstrates the two stages and their respective AC current loops.
Fig 3. An SRPP amplifier
With a rudimentary understanding of the SRPP circuit, I hope I have correctly identified the second AC current path here. From the plate of V2 to the grid of V1, exiting its cathode and the usual path after that. That being said, what is happening with respect to RK1, I'm not completely certain.
Fig 4. An Aikido amplifier
Here is where I become unstuck (if indeed I was ever the opposite), which is unfortunate because I am currently working on an Aikido project.
The pink DC current paths are accompanied here by the blue highlighted AC current from the power supply (ripple). CM, RM1, RM2, RS3 & V4 all form the power supply noise cancellation element of the Aikido circuit. Capacitor CM blocks the power supply DC, only allowing the AC ripple through the voltage divider formed by RM1 & RM2. A portion of the ripple is injected into the grid of V4 and out the plate, in anti-phase, so any ripple present in the amplified signal output is then cancelled out. Again, if I'm wrong please let me know.
However, I'm not sure of the current path of the Aikido's noise cancelling AC, or does it just disappear when it's job is done?
Similarly I have found it difficult to map the signal path. I'm not even sure if the portion of the green highlight I've half completed is correct. Also not sure how the connection between the plate of V2 and grid of V1 interact, nor where the current flows after the load.
Can anybody help me out?
The image links are obviously all broken on that page, but I have had some success piecing the article back together with the images still found in the original link: http://www.op316.com/tubes/tips/b410.htm
Anyway, my hope is that someone here might be willing to check the correctness of the diagrams I've put together below, which constitute the substance of what I've been able to grasp from those pages. Apologies if this is information easily found elsewhere - I haven't been able to uncover it.
Fig 1. Single stage basic amplifier
This is essentially a duplication of the first example in Kimura's article. Hopefully my photoshop work doesn't prove too unreadable.
The pink highlighting shows the DC current path from the power supply, through the plate resistor RP, triode V1, cathode resistor RK to ground and back to the power supply.
The green highlight shows the AC signal paths. First from the source, through grid leak resistor RG to ground and back to the source. It is my understanding that no current flows in grid stopper RS or the ground between RS & RK, but feel free to correct me on that point.
Further, current flow through RG excites AC current in the triodes grid(?), forming another AC current loop from the plate through output capacitor CO, the load, back to the cathode through cathode bypass capacitor CK. Again, I suspect there's no current flow between RK & CK, but if the circuit does not include a bypassing cap then the current would flow through cathode resistor RK instead (highlighted yellow).
Fig 2. Two stage basic amplifier
Please excuse the slightly unconventional routing of the ground here. It is largely an attempt to indicate how I would intend to make the ground connections if building this in real life. Otherwise this image just demonstrates the two stages and their respective AC current loops.
Fig 3. An SRPP amplifier
With a rudimentary understanding of the SRPP circuit, I hope I have correctly identified the second AC current path here. From the plate of V2 to the grid of V1, exiting its cathode and the usual path after that. That being said, what is happening with respect to RK1, I'm not completely certain.
Fig 4. An Aikido amplifier
Here is where I become unstuck (if indeed I was ever the opposite), which is unfortunate because I am currently working on an Aikido project.
The pink DC current paths are accompanied here by the blue highlighted AC current from the power supply (ripple). CM, RM1, RM2, RS3 & V4 all form the power supply noise cancellation element of the Aikido circuit. Capacitor CM blocks the power supply DC, only allowing the AC ripple through the voltage divider formed by RM1 & RM2. A portion of the ripple is injected into the grid of V4 and out the plate, in anti-phase, so any ripple present in the amplified signal output is then cancelled out. Again, if I'm wrong please let me know.
However, I'm not sure of the current path of the Aikido's noise cancelling AC, or does it just disappear when it's job is done?
Similarly I have found it difficult to map the signal path. I'm not even sure if the portion of the green highlight I've half completed is correct. Also not sure how the connection between the plate of V2 and grid of V1 interact, nor where the current flows after the load.
Can anybody help me out?
I think it would be a mistake to attempt to understand a circuit with current alone. Both voltage and current are accepted "fictions" that are really just simplifications in our large world, of fields in the smaller world of electricity, but they're only interesting together.
Many folk get into conceptual trouble in "grounding" schemes by thinking only in voltage, but you're thinking out of the box in a new way! I'd only caution that it's not useful as an only path. It takes both voltage and current to understand a circuit.
All good fortune,
Chris
Many folk get into conceptual trouble in "grounding" schemes by thinking only in voltage, but you're thinking out of the box in a new way! I'd only caution that it's not useful as an only path. It takes both voltage and current to understand a circuit.
All good fortune,
Chris
Here is an example where Voltage must be introduced. "Current" is already flowing through the valve, being pulled along by anode voltage. It is deterred by signal voltage as it tries to modulate current flow on to the anode - slightly more negative, slightly less current is allowed to pass, and vice-versa. But only voltage (pressure) need be applied to the grid; grid current is unnecessary to the valve's operation. Does that make any sense?
All good fortune,
Chris
Thanks Chris, you are right of course. It is easy to get in the habit of considering one without the other, and I have been guilty of that with regard to voltage in the quite recent past.
My fixation on current at this time is playing catch-up as it were, and that it seems more vitally linked to ground connections and where to place them.
Its simple, you have voltage in a circuit then you get current.dont over complicate things,leads to confusion and design errors and misconceptions with may lead to smoke.
A lot of things in a circuit tend to “connect” to the zero volt reference which we call “ground”, with the “ground” symbol being used as a shortcut. In order to actually understand a circuit and where the pitfalls of implementing it are, the schematic SHOULD be drawn with the connections explicitly shown. And analyze it with the understanding that each “wire” or connection between nodes has a voltage difference at each end due to the current flowing through it (DC and AC/signal). Real wires/busses/planes are physically small resistances and inductances. That will enable you to visualize the current paths in the “ground” and how things should be physically arranged when you build it.
The analysis you're doing is extremely useful when you get down to the point where you need to build something, as this can help you prevent rather obvious wiring errors. If you add a power supply to one of these circuits and trace where the AC noise currents are, they are in loops around the power supply caps, with the magnitude of AC noise current diminishing as you perform more filtering. Often when I get a noisy homebrew amp on the bench, the incoming signal ground from the RCA jack will be landed on one of those noisy power supply nodes, and that noise current gets mixed with the audio signal current and things get noisy in a hurry. When thinking only in terms of voltage, it's all 0V AC/DC along the ground bus, so on those merits alone you wouldn't expect this problem to occur.
On the Aikido, I would build one and put a 60Hz signal into it, then probe around with a scope to see what you can see. If you put a 1 ohm resistor between V1's plate and B+, you're going to see signal on that plate. Similarly in the sense of just thinking about current, you'll have a very hard time measuring current (AC or DC) through RS2, but there will be plenty of signal current through CK.
On the Aikido, I would build one and put a 60Hz signal into it, then probe around with a scope to see what you can see. If you put a 1 ohm resistor between V1's plate and B+, you're going to see signal on that plate. Similarly in the sense of just thinking about current, you'll have a very hard time measuring current (AC or DC) through RS2, but there will be plenty of signal current through CK.
Perhaps return you attention to the initial simplest circuit and reassess it. For example, replace the power supply with a local supply decoupling capacitor to show the idle dc current flow arising from static valve voltages. Then use two circuits to show how signal current flows into, and out of the coupling capacitor, due to the AC voltage generated at the anode node (ie. the positive and negative voltage deviation away from the static idle anode voltage). If you can show and explain that in text then imho that would be a better start for you.
I would also avoid trying to describe grid current flow as a tangible current that needs to be accounted for in operation. Much better and easier to use just input grid voltage as the dominant influence.
I would also avoid trying to describe grid current flow as a tangible current that needs to be accounted for in operation. Much better and easier to use just input grid voltage as the dominant influence.
Indeed, this is pretty much the object of my concern and the reason I wanted to ask the question. So I can most effectively connect the "ground" points of my circuit for lowest noise. Knowing where the current paths route and their magnitude surely is an important part of that process.
Power supplies I'm more confident with, but amplifier circuits I still find rather baffling a lot of the time.
That isn't a bad idea at all. I feel a bit silly not having thought of it, since I've already built an Aikido from one of Broskie's PCBs, and a CCDA from a PCB of my own design. Neither uses bypass caps across the cathode resistor, but that shouldn't matter in the greater scheme of things.
In a perfect world there is no grid current. All tubes have a grid leak resistor and the value of that resistor is usually stated as a max value. There is the odd election that gets caught by the grid and the vacume inside the tube is way from an absolute vacume.
I shouldn't have referred to grid current. It was a mistake in terminology rather then lack of understanding (although there's obviously plenty of that too).
I found this Merlin Blencowe article very helpful and use it in all my builds. Grounding and Grounding Schemes.
So do I 🙂 and agree, it's a great document, and have used a combo of the 'improved bus' and 'local star' grounding in previous builds as well as the hum-loop block network.
Thanks for reminding me of it. It's about time for a re-read.
Thank you, I skim-read an older tubecad article https://www.tubecad.com/may2000/page2.html but had misunderstood it in doing so.
Having re-read it this appears to be the salient point from the article:
"At idle we know there is zero difference between top and bottom tube current flow, thus no current is delivered into the load resistance. But if a positive pulse is applied to the bottom tube's grid, its current conduction might increase from 10 mA to 15 mA, while the top triode's current conduction decreases from 10 mA to 5 mA because of the bottom tube's greater conduction through resistor Rak, which forces the top tube's grid negative. The difference between these two currents is 10 mA, which is delivered into the load resistance."
...
"When the top tube conducts more current than the bottom tube, the difference in current flows through the load resistance into the cathode of the top tube. When the bottom tube conducts more current than the top tube, the difference in current flows through the bottom tube's plate into resistor Rak and then into load resistance."
Also of interest are Merlins articles for Audio Xpress on the SRPP http://www.valvewizard.co.uk/SRPP_Blencowe.pdf, which are a little technical for myself, but contains this nugget:
"If audio electronics magazines are anything to go by, the operation of the SRPP seems to be routinely misunderstood as something involving a cathode follower."
That was pretty much what I was doing.
I've been thinking about this since you posted it, but only in the last few hours has it become clear what you were asking and what it means.
The decoupling capacitor appears to have been the missing element (for me at least) to start fully understanding the operation of a valve circuit and the current flowing within. At first I had misunderstood the function of the local decoupling cap, dismissing it as an extra filter for the power supply, or if situated before the first amplification stage, considered it more appropriately as the final stage of the power supply itself. What else could a cap linking B+ & ground be?
It shouldn't have been a mystery. The Kimura article states quite plainly:
"What exactly is the purpose of the de-coupling capacitor? Removing ripple from the power supply? No. The purpose of the de-coupling capacitor is to complete the AC plate (signal) current loop in fig.3 which travels from [the anode, through the decoupling cap and back to the cathode via its bypass capacitor]. 100% of the AC signal current flows through this capacitor. Without it, the amplification circuit is not complete."
Consequently, my basic interpretation of the decoupling capacitor's purpose is to short the amplified AC signal to ground, providing a path for that current back to the cathode - plus it prevents said AC from wandering back to the power supply. I imagine without a local decoupling cap the last filter capacitor in the power supply becomes the de facto recipient of that AC current.
And so, back is the basic valve amplifier again:
Pink = DC current flow | Green = AC current flow
When an AC voltage is applied to the triode grid, the flow of electrons from cathode to anode produces an amplified 'sympathetic' AC voltage at the anode. (And an unamplified one at the cathode I assume. Otherwise how could a cathode follower work?)
Apparently it was foolish to think that the signal current would only follow the path through coupling cap CO to the load and back. Rather, at the plate the current path would appear to split, flowing through either the load or the decoupling cap, and re-combining at the base of CK (where Kimura insists the connection is made). Kirchhoff's law in action.
My expectation is that the amount of current in each pathway is governed by the current requirements of the load? i.e. if the current draw of the load increases, less current flows through the decoupling cap, and vice versa.
Also, to semi-answer my initial question - where does current flow in a valve amplifier circuit - it would seem that answer is: almost everywhere.
How am I doing?
Perhaps use two diagrams - one showing what happens as signal current flow into C0, and the other for what happens when signal current flows out of C0. That may further assist your understanding.
I appreciate it may mess up the simplicity - but Rk is a significant path for signal current too.
A few old amp schematics showed how each valve stage should have a star gnd wired - where Rg, Rk, Ck connect at one point. Sadly, modern schematics typically don't show that representation, as it allows the designer to inform the constructor how to localise signal currents so they don't interact with other stages. Eg.
http://www.r-type.org/articles/art-003f.htm
I appreciate it may mess up the simplicity - but Rk is a significant path for signal current too.
A few old amp schematics showed how each valve stage should have a star gnd wired - where Rg, Rk, Ck connect at one point. Sadly, modern schematics typically don't show that representation, as it allows the designer to inform the constructor how to localise signal currents so they don't interact with other stages. Eg.
http://www.r-type.org/articles/art-003f.htm
I suspect another mental stumbling block is that I have been thinking of AC current as being directional.
I was thinking about the cathode bypass capacitor, which may be polarised, and how the current couldn't flow through it to the cathode in the 'wrong' direction. Would it be fair to say that AC current flows through CK for half a cycle and through RK the other half? Or am a digging a hole to nowhere?
AC being bi-directional does seem to make its grounding considerations a bit simpler.
I was thinking about the cathode bypass capacitor, which may be polarised, and how the current couldn't flow through it to the cathode in the 'wrong' direction. Would it be fair to say that AC current flows through CK for half a cycle and through RK the other half? Or am a digging a hole to nowhere?
AC being bi-directional does seem to make its grounding considerations a bit simpler.
Yes, digging a hole 😎
Perhaps google for AC circuit tutorials etc, and read as widely as practical to build up a better understanding. It's not uncommon for people with no initial exposure to circuitry to get an interest later on - it then just takes interest and incentive. I know someone who was a carpenter by trade, only to then flourish as a circuit practitioner and transformer guru.
Perhaps google for AC circuit tutorials etc, and read as widely as practical to build up a better understanding. It's not uncommon for people with no initial exposure to circuitry to get an interest later on - it then just takes interest and incentive. I know someone who was a carpenter by trade, only to then flourish as a circuit practitioner and transformer guru.