Experimental 30-Watt All-Tube Guitar-Amp
By H.K. Richter
This is a theoretical design for an all-tube guitar-amp. But since it has not been built, it is considered experimental. However, it is based on circuits and design information in the 1965 RCA Receiving Tube Manual, has been thoroughly analyzed, and should work fine once final testing is done in the event someone decides to actually build it.
We start with the Preamp, which uses a 12AX7A dual-triode. This circuit is of my own design, though based on long established and proven topologies, to which it can readily be compared. Here’s a link to schematics for many of the most popular guitar-amps.
Guitar Amp Schematics - www.thetubestore.com
The input signal goes directly to the grid of the first triode, V1a, which is set to provide a gain of 68, and whose output is sent to a Fender style tone-stack, as used in the 5F6-A Bassman, except I changed the part-values to make the response flatter than that of the original stack. Should anyone wish to build this circuit, other part-values can be used in accordance with the calculator linked here. Yet Another Tonestack Calculator
And the original Bassman schematic is linked here. Fender-Bassman-5F6A-Schematic.pdf
The output of the tone-stack goes to a 1M volume-control whose value whose element is required by V1a should all tone controls be turned all the way up, but which also serves to bias the grid of V1b, itself set to provide a gain of 73. That volume control, labeled Pre, will allow V1b to be overdriven, if that is desired. However, I recommend that it is kept low, say between 1/4 to 1/3 turn clockwise, to ensure good note definition up front.
The output of V1b goes to a Master Volume control whose signal goes next to the grid of the first triode in the Driver Stage, which uses yet another 12AX7A. That triode is set to perform as another gain stage, with a gain of 68, and whose output is fed directly to the second triode of the tube, itself set for unity gain, so the Master Volume control can be used to get distortion from the Driver and from the push-pull Output Stage involving two 6L6GC beam-power tubes with maximum plate dissipation at 30W electrical each (thus giving 30W RMS signal-power into the speaker via the OT).
Note: There are errors in this diagram. Changes have been made according to advice from posts below. See Post #14 for latest updated version.
This circuit is based on a 30W amp in the 1965 RCA Receiving Tube Manual, page 535, though I made a number of changes, including part-value changes to get more gain out of the Driver Stage, adding a Feedback control, and specifying a different transformer (since the original ceased being made long ago). The original also used a 7199 tube for the Driver circuit, which has a pentode, connected as its input, and a triode, connected as its output. That tube is still available as new old stock (NOS) and was designed as a low-noise preamp tube for hi-fi applications. So, it would work, but I replaced it with another 12AX7A instead, using the first triode in a basic gain stage, as said, and the second as the driver tube. And since V1b, V2a, and V2b can be overdriven, if desired, along with V1a if a stomp-box with extra gain is used on the amp’s input, then the output tubes should easily be overdriven if distortion is wanted from the output stage.
At only 30 W RMS into an 8-ohm speaker, this would make a nice personal amp for use at home or in a recording studio, or in small-room live venues like dinner-clubs, certain nightclubs, etc., but could be miked for use in larger buildings or outdoors. However, after recently updating my diagrams, I am of the opinion that 30W should be considered a minimum. If it works, the amp might turn out to be louder than that.
This is the power supply circuit. It was changed from the original as follows. A modern power transformer has been specified. And only the transformer’s voltages were noted on the original diagram, so I calculated the others based on what I wanted for the gain stages, driver, and output tubes, then redesigned the filters accordingly, although the rectification circuit topology was retained, with the exception that the heater-cathode type of dual-diode tube originally specified is not easily obtained and was replaced by an equivalent (the 5U4GB) which remains in wide usage and is thus readily available. Also, this is an updated diagram. The previous one was redone based on advice from forum members below. Checking a number of proven guitar-amps, I saw the need for negative bias voltage for the input grids of the 6L6GCs. This schematic has replaced the previous drawing.
As noted, this is experimental circuitry. It has not been built and tested, so anyone who builds it must accept all responsibility for the outcome. However, if someone indeed builds and tests it, I would appreciate an email on the result. hkurtrichter@gmail.com
EOF
By H.K. Richter
This is a theoretical design for an all-tube guitar-amp. But since it has not been built, it is considered experimental. However, it is based on circuits and design information in the 1965 RCA Receiving Tube Manual, has been thoroughly analyzed, and should work fine once final testing is done in the event someone decides to actually build it.
We start with the Preamp, which uses a 12AX7A dual-triode. This circuit is of my own design, though based on long established and proven topologies, to which it can readily be compared. Here’s a link to schematics for many of the most popular guitar-amps.
Guitar Amp Schematics - www.thetubestore.com
The input signal goes directly to the grid of the first triode, V1a, which is set to provide a gain of 68, and whose output is sent to a Fender style tone-stack, as used in the 5F6-A Bassman, except I changed the part-values to make the response flatter than that of the original stack. Should anyone wish to build this circuit, other part-values can be used in accordance with the calculator linked here. Yet Another Tonestack Calculator
And the original Bassman schematic is linked here. Fender-Bassman-5F6A-Schematic.pdf
The output of the tone-stack goes to a 1M volume-control whose value whose element is required by V1a should all tone controls be turned all the way up, but which also serves to bias the grid of V1b, itself set to provide a gain of 73. That volume control, labeled Pre, will allow V1b to be overdriven, if that is desired. However, I recommend that it is kept low, say between 1/4 to 1/3 turn clockwise, to ensure good note definition up front.
The output of V1b goes to a Master Volume control whose signal goes next to the grid of the first triode in the Driver Stage, which uses yet another 12AX7A. That triode is set to perform as another gain stage, with a gain of 68, and whose output is fed directly to the second triode of the tube, itself set for unity gain, so the Master Volume control can be used to get distortion from the Driver and from the push-pull Output Stage involving two 6L6GC beam-power tubes with maximum plate dissipation at 30W electrical each (thus giving 30W RMS signal-power into the speaker via the OT).
Note: There are errors in this diagram. Changes have been made according to advice from posts below. See Post #14 for latest updated version.
This circuit is based on a 30W amp in the 1965 RCA Receiving Tube Manual, page 535, though I made a number of changes, including part-value changes to get more gain out of the Driver Stage, adding a Feedback control, and specifying a different transformer (since the original ceased being made long ago). The original also used a 7199 tube for the Driver circuit, which has a pentode, connected as its input, and a triode, connected as its output. That tube is still available as new old stock (NOS) and was designed as a low-noise preamp tube for hi-fi applications. So, it would work, but I replaced it with another 12AX7A instead, using the first triode in a basic gain stage, as said, and the second as the driver tube. And since V1b, V2a, and V2b can be overdriven, if desired, along with V1a if a stomp-box with extra gain is used on the amp’s input, then the output tubes should easily be overdriven if distortion is wanted from the output stage.
At only 30 W RMS into an 8-ohm speaker, this would make a nice personal amp for use at home or in a recording studio, or in small-room live venues like dinner-clubs, certain nightclubs, etc., but could be miked for use in larger buildings or outdoors. However, after recently updating my diagrams, I am of the opinion that 30W should be considered a minimum. If it works, the amp might turn out to be louder than that.
This is the power supply circuit. It was changed from the original as follows. A modern power transformer has been specified. And only the transformer’s voltages were noted on the original diagram, so I calculated the others based on what I wanted for the gain stages, driver, and output tubes, then redesigned the filters accordingly, although the rectification circuit topology was retained, with the exception that the heater-cathode type of dual-diode tube originally specified is not easily obtained and was replaced by an equivalent (the 5U4GB) which remains in wide usage and is thus readily available. Also, this is an updated diagram. The previous one was redone based on advice from forum members below. Checking a number of proven guitar-amps, I saw the need for negative bias voltage for the input grids of the 6L6GCs. This schematic has replaced the previous drawing.
As noted, this is experimental circuitry. It has not been built and tested, so anyone who builds it must accept all responsibility for the outcome. However, if someone indeed builds and tests it, I would appreciate an email on the result. hkurtrichter@gmail.com
EOF
Attachments
Last edited:
The phase splitter resistors are way off.
Hello Kurtus,
I think there are some issues with the design.
1) The bias of the 6L6GC's. The bias in your design is derived from the voltage that developes over the 120 Ohm resistor in the power supply. That voltage will be the total current the amplifier draws times 120 Ohm.
With 450 V on the screengrids of the 6L6GC's, you would need a bias of at least -45 V, and probably higher, for class AB1 push-pull operation. But to get to -45 V with the 120 Ohm resistor would mean that the amplifier has to draw a total of 45/120 = 375 mA (!). This of course is not possible because it would violate the maximum current a pair of 6L6GC's and the specified power transformer could withstand.
2) Besides the value of 120 Ohm being way too low, it is not a good idea to derive the bias in this way for a class AB1 power stage. When the power stage gets driven, the current goes up, which in turn makes the bias voltage go up. So, your bias voltage will not be stable.
3) The value of the grid resistors of the 6L6GC's in your schematic is 1M. That value is way too high. The datasheets specify 100K as a maximum for fixed bias operation, although in practice 220K is being used by Fender and others. The higher the value of the grid resistor, the higher the risk of the power tubes 'running away' (they start to draw more and more current).
4) The value of the cathode and plate resistor of V2b (a 'cathodyne' a.k.a. 'split-load inverter') should be equal for poper operation. With only 5.1K for the lower resistor and 470K for the upper, the drive of the lower 6L6GC will be almost non-existent.
I did not study/calculate all aspects of the design so there might be more issues.
I think there are some issues with the design.
1) The bias of the 6L6GC's. The bias in your design is derived from the voltage that developes over the 120 Ohm resistor in the power supply. That voltage will be the total current the amplifier draws times 120 Ohm.
With 450 V on the screengrids of the 6L6GC's, you would need a bias of at least -45 V, and probably higher, for class AB1 push-pull operation. But to get to -45 V with the 120 Ohm resistor would mean that the amplifier has to draw a total of 45/120 = 375 mA (!). This of course is not possible because it would violate the maximum current a pair of 6L6GC's and the specified power transformer could withstand.
2) Besides the value of 120 Ohm being way too low, it is not a good idea to derive the bias in this way for a class AB1 power stage. When the power stage gets driven, the current goes up, which in turn makes the bias voltage go up. So, your bias voltage will not be stable.
3) The value of the grid resistors of the 6L6GC's in your schematic is 1M. That value is way too high. The datasheets specify 100K as a maximum for fixed bias operation, although in practice 220K is being used by Fender and others. The higher the value of the grid resistor, the higher the risk of the power tubes 'running away' (they start to draw more and more current).
4) The value of the cathode and plate resistor of V2b (a 'cathodyne' a.k.a. 'split-load inverter') should be equal for poper operation. With only 5.1K for the lower resistor and 470K for the upper, the drive of the lower 6L6GC will be almost non-existent.
I did not study/calculate all aspects of the design so there might be more issues.
Bias filter cap has wrong polarity.
Phase splitter V2B will be biased into full conduction with grid current due to high grid voltage (DC coupling) and low value cathode resistor. So can't process a signal.
Generally a cathodyne cannot be used as a gain stage.
Also largely increasing gain within the NFB loop would likely cause HF oscillation.
Phase splitter V2B will be biased into full conduction with grid current due to high grid voltage (DC coupling) and low value cathode resistor. So can't process a signal.
Generally a cathodyne cannot be used as a gain stage.
Also largely increasing gain within the NFB loop would likely cause HF oscillation.
Last edited:
Thank you all for your interest in my design. I will be making the changes you suggest and replacing the diagrams with updated ones as soon as I can.
It's great to get such instruction. My whole purpose for posting was to come up with a viable design. Your help is greatly appreciated.
It's great to get such instruction. My whole purpose for posting was to come up with a viable design. Your help is greatly appreciated.
To get the phase splitter right, look for example at the fender 5e6 bassman schematic. By the way, 30 watts into an efficient guitar speaker wil be very, very loud at home!
Yes, I have done that, and also rechecked other amps, including Ampegs, Garnets, etc. Very helpful review.
And I have just completely reworking my design, based on renewed study of such amps, and re-crunching some numbers.
Will be posting the updated schematics shortly. But I won't be surprised to learn that the design is still not totally viable.
Thanks.
And I have just completely reworking my design, based on renewed study of such amps, and re-crunching some numbers.
Will be posting the updated schematics shortly. But I won't be surprised to learn that the design is still not totally viable.
Thanks.
Sorry for the typo. I have just completed reworking my design and changed many things.
And I have just replaced the previous diagrams with updated versions.
Maybe the project is getting close to completion now.
And thanks again.
And I have just replaced the previous diagrams with updated versions.
Maybe the project is getting close to completion now.
And thanks again.
Hello Kurtus,
In the revised schematic of the "Output Section" in post #1 the 1M grid resistor of the second setion of V2 is connected to ground. This is not right.
Because the 1M resistor is connected to ground, the grid will sit at 0 Vdc. That will cause the current through the second triode section to be way too low to create the voltage swing needed to drive the power tubes.
You want the grid to sit at something like 80 Vdc, so the cathode will settle at something like 81.5 Vdc.
I advise you to study some proven designs in which the preceding stage is coupled to a "splitload inverter" by means of a coupling capacitor, like in your schematic. You then will see that an extra bias resistor is needed at the cathode, and that the 1M grid resistor has to be connected to the bottom of this extra resistor.
In the revised schematic of the "Output Section" in post #1 the 1M grid resistor of the second setion of V2 is connected to ground. This is not right.
Because the 1M resistor is connected to ground, the grid will sit at 0 Vdc. That will cause the current through the second triode section to be way too low to create the voltage swing needed to drive the power tubes.
You want the grid to sit at something like 80 Vdc, so the cathode will settle at something like 81.5 Vdc.
I advise you to study some proven designs in which the preceding stage is coupled to a "splitload inverter" by means of a coupling capacitor, like in your schematic. You then will see that an extra bias resistor is needed at the cathode, and that the 1M grid resistor has to be connected to the bottom of this extra resistor.
Additional:
What is the reason for the value of only 3 nF for the coupling capacitor between the first and second triode section of V2?
What is the reason for the value of only 3 nF for the coupling capacitor between the first and second triode section of V2?
To Robert H. Gribnau
Here is another diagram for the Output Section. It's more like the original RCA schematic and eliminates the said 1M resistor to ground on the driver tube. Truthfully, I don't remember where that 1M resistor came from. The main change would be to replace the given V2b circuit with this one and re-label the B connection below with the C label of the first post while leaving the rest of the first diagram unchanged. I don't recall why I posted an unfinished diagram to start with, before correcting its flaws. My bad. Given time I will post a full but corrected schematic instead of replacing the first one with updated versions.
Here is another diagram for the Output Section. It's more like the original RCA schematic and eliminates the said 1M resistor to ground on the driver tube. Truthfully, I don't remember where that 1M resistor came from. The main change would be to replace the given V2b circuit with this one and re-label the B connection below with the C label of the first post while leaving the rest of the first diagram unchanged. I don't recall why I posted an unfinished diagram to start with, before correcting its flaws. My bad. Given time I will post a full but corrected schematic instead of replacing the first one with updated versions.
Seems the post #12 phase splitter tube's cathode resistor is a bit off, and the plate resistor as well.
The circuit cannot work as given. (The schematic in post #1 could not have worked either.)
Generally speaking, circuit component values must be calculated from design equations that were derived
from circuit analysis of the desired topology. The circuit component values given were not found in any tube manual,
could not have been calculated from any circuit model, and a Concertina phase splitter does not have any voltage gain.
Perhaps AI is not ready for circuit design, or even copying someone's personal circuit design. These schematics do not qualify
to be called experimental circuitry, rather they are simply wrong, or perhaps as W. Pauli said, they are not even wrong.
The circuit cannot work as given. (The schematic in post #1 could not have worked either.)
Generally speaking, circuit component values must be calculated from design equations that were derived
from circuit analysis of the desired topology. The circuit component values given were not found in any tube manual,
could not have been calculated from any circuit model, and a Concertina phase splitter does not have any voltage gain.
Perhaps AI is not ready for circuit design, or even copying someone's personal circuit design. These schematics do not qualify
to be called experimental circuitry, rather they are simply wrong, or perhaps as W. Pauli said, they are not even wrong.
To Rayma
The diagram of Post #12 was not from the RCA manual but from various sources. Anyway, here is the updated version of the diagram of Post #1.
Just so you know, this is a work in progress. Hence the notes saying "Experimental Circuitry". But your observations are valid. And feel free to check the diagram below for flaws. I am here to get it right.
The diagram of Post #12 was not from the RCA manual but from various sources. Anyway, here is the updated version of the diagram of Post #1.
Just so you know, this is a work in progress. Hence the notes saying "Experimental Circuitry". But your observations are valid. And feel free to check the diagram below for flaws. I am here to get it right.
Generally speaking, beginners should build (but not modify) an existing, well documented, and proven design.
By proven, I mean that the design has been previously built, measured, and tested in actual use, by an experienced person.
The design process only comes after one first learns the basics, followed by learning analysis, and then synthesis.
Unfortunately, the Socratic method does not work for electronic circuit design.
By proven, I mean that the design has been previously built, measured, and tested in actual use, by an experienced person.
The design process only comes after one first learns the basics, followed by learning analysis, and then synthesis.
Unfortunately, the Socratic method does not work for electronic circuit design.
Please
Could you please share your calculation, knowing that for ac signals the cathode resistor and plate resistor (both 130K) are in parallel with the 1M resistor?
In the latest updated version, Post #14, the 1M resistor has been eliminated. And the only part-value changes I otherwise made to the original RCA schematic on V2b was to go from 150k cath / plate resistors to 130k, and the coupling capacitors from 0.25uF to 0.1uF. I have done numerous calculations over and over while confirming results using online calculators, including checking load-lines, operating points for each tube, frequencies from different coupling caps, gains for different cath / plate resistors for each tube. and so on. But that work is too tedious to post to this thread. It would take many pages and I'm not in the mood. If you are interested in building this project, you can look into the math on your own. But I have confirmed that most online calculators work well. Look them up by simply searching the type of calculator you seek.
Examples:
https://www.ampbooks.com/mobile/amplifier-calculators/output-impedance/
https://www.v-cap.com/coupling-capacitor-calculator.php
https://calculator.academy/bridge-rectifier-output-voltage-calculator/
https://www.ampbooks.com/mobile/vacuum-tubes/12AX7/
https://www.amplifiedparts.com/tech-articles/tube-amplifier-bias-calculator
https://www.ampbooks.com/mobile/amplifier-calculators/
Examples:
https://www.ampbooks.com/mobile/amplifier-calculators/output-impedance/
https://www.v-cap.com/coupling-capacitor-calculator.php
https://calculator.academy/bridge-rectifier-output-voltage-calculator/
https://www.ampbooks.com/mobile/vacuum-tubes/12AX7/
https://www.amplifiedparts.com/tech-articles/tube-amplifier-bias-calculator
https://www.ampbooks.com/mobile/amplifier-calculators/
Pretty simple. Since the load (1M) is much larger than the source (output impedance of the first stage, which is well under 220k),
just ignore the source impedance for an estimate (with around 20% accuracy) of the pole frequency.
Then f = 1 / ( 2Pi x Rgrid x C)
and f = 1 / (6.28 x 1M x 3nF ) = 53Hz
But the output impedance of the first stage is actually in series with the 1M grid resistor, and attenuates the signal.
However, the 130k resistors are not connected to the 1M resistor in any way.
Of course, the first stage output impedance can be calculated, but the frequency within 20% is good enough.
And that circuit is defective and cannot work, regardless.
just ignore the source impedance for an estimate (with around 20% accuracy) of the pole frequency.
Then f = 1 / ( 2Pi x Rgrid x C)
and f = 1 / (6.28 x 1M x 3nF ) = 53Hz
But the output impedance of the first stage is actually in series with the 1M grid resistor, and attenuates the signal.
However, the 130k resistors are not connected to the 1M resistor in any way.
Of course, the first stage output impedance can be calculated, but the frequency within 20% is good enough.
And that circuit is defective and cannot work, regardless.
By the way, I cannot afford to build the projects I post about. But I have been studying audio electronics as a hobby for decades, but clearly still have much to learn. I would like to build and test the circuits I come up with, starting with kits, but I'm just a poor retiree on a fixed income and the only way I can be in on the audio engineering scene is from a circuit-design standpoint. Then again, I know much more about solid-state electronics than tube circuits. But I would like to know tube-amps too. And since this is a DIY website, I thought I might get some help with my tube-amp projects here.