Did some doodling the other night. Digging through the stash, I initially was after the box of 6BQ6GTB's, but I found a shoebox full of NOS 7C5 and some sockets for them. I don't need 30-40W so clearly the 7C5 is the way forward here.
Couldn't find a good operating point for the 12AT7 LTP for the phase inverter, but 12AU7 biases up nice and puts up a clean operating point. 11mA per section, op point should be OK.
12AX7 is the clear choice for input gain and tone stack gain recovery, but I still should take a second look at that and make sure it'll work the way I want. R18 is a typo, and should be a 1k grid stopper, not 100k
Borrowed from the 5F6 and some Marshall stuff for the inputs. Should work OK.
Couldn't find a good operating point for the 12AT7 LTP for the phase inverter, but 12AU7 biases up nice and puts up a clean operating point. 11mA per section, op point should be OK.
12AX7 is the clear choice for input gain and tone stack gain recovery, but I still should take a second look at that and make sure it'll work the way I want. R18 is a typo, and should be a 1k grid stopper, not 100k
Borrowed from the 5F6 and some Marshall stuff for the inputs. Should work OK.
I don't get the volume control. If I read part numbers correctly, it looks like a dual pot. SO both channels go up and down at the same time. WHy would you not want separate volume controls? Also, if separated, then buck's point becomes especially important. If you turn to zero on one, it grounds the grid so the other channel cannot pass signal through.
I assume you intended a ground between the two power tube cathode resistors.
You seem to be counting on the bass control to be the grid return resistance for the phase inverter input grid. That alone is scary, but also, that grid then will not have the elevated voltage reference like the other grid has from the tail resistor. MAy I suggest another 470k for the left grid, and add a cap between the grid and the tone stack
I assume you intended a ground between the two power tube cathode resistors.
You seem to be counting on the bass control to be the grid return resistance for the phase inverter input grid. That alone is scary, but also, that grid then will not have the elevated voltage reference like the other grid has from the tail resistor. MAy I suggest another 470k for the left grid, and add a cap between the grid and the tone stack
Some calculations, partly to underline what Enzo already wrote.
If the current per triode in the LTP would be 11 mA, than the voltage at the top of the tail, so at the cathodes, would be 0.022 x (4K7 + 250) = 108.9 V. The grid of the left side triode however is at ground potential which would give a bias of - 108.9 V which doesn't make sense.
The bias of the right side triode would be 0.022 x 250 = - 5.5 V. The voltage over the anode resistors would be 0.011 x 9.1K = 100.1 V. The voltage between anode and cathode would be 280 - (100.1 + 108.9) = 71 V. But if you look in the datasheets for the 12AU7 you will see that the anode current would almost be 0 mA at Va = 71 V and Vg = - 5.5 V.
If the current per triode in the LTP would be 11 mA, than the voltage at the top of the tail, so at the cathodes, would be 0.022 x (4K7 + 250) = 108.9 V. The grid of the left side triode however is at ground potential which would give a bias of - 108.9 V which doesn't make sense.
The bias of the right side triode would be 0.022 x 250 = - 5.5 V. The voltage over the anode resistors would be 0.011 x 9.1K = 100.1 V. The voltage between anode and cathode would be 280 - (100.1 + 108.9) = 71 V. But if you look in the datasheets for the 12AU7 you will see that the anode current would almost be 0 mA at Va = 71 V and Vg = - 5.5 V.
Yes, you are correct. I've goofed up here as originally the B+ was higher, I missed that in the last go-round when I finally settled on the 7C5 output tubes.
This is why I post these things-in my mind, "yep I already did the figures on that, they're right here" and I don't remember to go back and fix the changes.
As to the volume controls-originally they were separate, I joined them together out of ignorance. I don't even play guitar, yet, so it went completely over my head that they wouldn't be joined together. 270K should be enough for the decoupling resistors, I'll rework that.
Yes, I am missing a ground at the 7C5 cathodes.
Thanks, gents. I'll make these corrections and run it by you again, before I start punching holes.
Worked on this a bit more. By using the -50V bias supply for the LTP tail, I was able to throw some more voltage to the LTP, although I had to reduce the current some. Now right around 8mA flows through each triode.
I went back and did some more reading and was able to apply a very commonly used gain/cathode follower stage used by MANY guitar amp makers. Now V5, the 12AX7 gain stage and buffer to feed the tone stack, should give a gain of ~90 and an output impedance of ~620 ohms. I have been advised to add a grid-cathode protection diode and resistor to stop the grid arcing to the cathode on startup.
Also, since I have a single input for each channel instead of two, I change the 68k grid stopper (originally there were two in parallel, one for each input) to a single 37K, as there's only one input now.
It's not Hi-Fi, but it isn't supposed to be. (And I have not implemented a feedback loop yet.) The LTP will be around 3.5ish % THD, at maximum drive for the power tubes.
I went back and did some more reading and was able to apply a very commonly used gain/cathode follower stage used by MANY guitar amp makers. Now V5, the 12AX7 gain stage and buffer to feed the tone stack, should give a gain of ~90 and an output impedance of ~620 ohms. I have been advised to add a grid-cathode protection diode and resistor to stop the grid arcing to the cathode on startup.
Also, since I have a single input for each channel instead of two, I change the 68k grid stopper (originally there were two in parallel, one for each input) to a single 37K, as there's only one input now.
It's not Hi-Fi, but it isn't supposed to be. (And I have not implemented a feedback loop yet.) The LTP will be around 3.5ish % THD, at maximum drive for the power tubes.
There's still something not right around the LTP.
According to the schematic the voltage drop over the tail is 54.25 V, so the current throught the tail would than be 54.25 / (3K3 + 250) = 15.3 mA.
But according to the schematic the voltage drops over R5 and R10 are 280 - 150 = 130 V. So the current through one of these resistors would be 130 / 12K = 10.8 mA.
The current through R5 and R6 (2 x 10.8 mA = 21.6 mA) should be the same as the current through the tail (15.3 mA) but they are not.
I did check if perhaps the "V3 Cathodes at 4.25V" in the schematic means the voltage difference between the grid and the cathodes (so the voltage drop over R12) but that would give a current of 4.25 / 250 = 17 mA which also doesn't match with the calculated currents through R5 and R10.
According to the schematic the voltage drop over the tail is 54.25 V, so the current throught the tail would than be 54.25 / (3K3 + 250) = 15.3 mA.
But according to the schematic the voltage drops over R5 and R10 are 280 - 150 = 130 V. So the current through one of these resistors would be 130 / 12K = 10.8 mA.
The current through R5 and R6 (2 x 10.8 mA = 21.6 mA) should be the same as the current through the tail (15.3 mA) but they are not.
I did check if perhaps the "V3 Cathodes at 4.25V" in the schematic means the voltage difference between the grid and the cathodes (so the voltage drop over R12) but that would give a current of 4.25 / 250 = 17 mA which also doesn't match with the calculated currents through R5 and R10.
Yep. Schematic Plate voltage is wrong, but the current is correct now at 8mA per triode. Plate voltage should be around 150V.
Going to the -50V line instead of ground gave 330V for the LTP to work in, I just hope that the 3.3K tail is enough. I've read that sometimes the anode resistors are trimmed to make up for a too-short tail?
Going to the -50V line instead of ground gave 330V for the LTP to work in, I just hope that the 3.3K tail is enough. I've read that sometimes the anode resistors are trimmed to make up for a too-short tail?
With 8 mA per triode the bias would be - 4 V. The voltage between anodes and cathodes would be 127.2 V. The datasheets for the 12AU7 shows about 5 mA per triode at these voltages.
Yes, if the tail is short(er), balance can be improved by choosing the value of the anode resistor at the right hand side a bit higher than the left hand side. You can see this for instance on page 4 of the attached datasheet for the ECC40 (left side 110K, right side 120K) with very good results distortion wise. But the tail there is not very short to begin with.
Why do you want to run the 12AU7 run at this relatively high current of 8 mA per section? Because of it, the value of the anode resistors gets rather low, which in view of gain and distortion is not so good. I think it would be better to run the 12AU7 at a lower current (like 1 to 2 mA per section). Or maybe use a double triode with a higher μ since they tend to make better LTP's.
Yes, if the tail is short(er), balance can be improved by choosing the value of the anode resistor at the right hand side a bit higher than the left hand side. You can see this for instance on page 4 of the attached datasheet for the ECC40 (left side 110K, right side 120K) with very good results distortion wise. But the tail there is not very short to begin with.
Why do you want to run the 12AU7 run at this relatively high current of 8 mA per section? Because of it, the value of the anode resistors gets rather low, which in view of gain and distortion is not so good. I think it would be better to run the 12AU7 at a lower current (like 1 to 2 mA per section). Or maybe use a double triode with a higher μ since they tend to make better LTP's.
Last edited:
Guitar/Instrument usage. 12AX7 and 12AU7 are available everywhere, forever. 6V6 is always available. 7C5 is a simple socket change or socket adaptor from 6V6, so no worries in the future. Might even dual-socket the amp just for that reason. (I have lots of NOS 7C5 but not many 6V6 anymore.)
Also when going through the loadlines, 12AU7, 12AY7, and 12AX7 all are within 1 percent of each other on distortion given the operating conditions, asking for 48V of total output swing. Having that extra current through the triode should help push the 7C5's harder. I am told phase inverter overdrive distortion is not one of the desirable modes with guitar amplifiers and to make sure to avoid it if possible. I like 12AY7's operating point more, but I don't have any yet.
I have no idea what kind of distortion signature I'll get or want from this thing but I am trying to stay inside the rails from the player's perspective, without getting lost into the audiophoolery that comes with guitarists and "amp voodoo".
Also you are correct, I did not update the plate voltage for the LTP in the schematic notes. It should be ~154v with 8mA through each triode.
I am tempted to just go back to the cathodyne that I am quite familiar with, but this should work well once I get the details ironed out and the math errors eliminated.
Also when going through the loadlines, 12AU7, 12AY7, and 12AX7 all are within 1 percent of each other on distortion given the operating conditions, asking for 48V of total output swing. Having that extra current through the triode should help push the 7C5's harder. I am told phase inverter overdrive distortion is not one of the desirable modes with guitar amplifiers and to make sure to avoid it if possible. I like 12AY7's operating point more, but I don't have any yet.
I have no idea what kind of distortion signature I'll get or want from this thing but I am trying to stay inside the rails from the player's perspective, without getting lost into the audiophoolery that comes with guitarists and "amp voodoo".
Also you are correct, I did not update the plate voltage for the LTP in the schematic notes. It should be ~154v with 8mA through each triode.
I am tempted to just go back to the cathodyne that I am quite familiar with, but this should work well once I get the details ironed out and the math errors eliminated.
Last edited:
Never forget, this is a guitar amp, not a hifi. SO often, in online discussions about making an amp "better" the point being to make it more hifi. Sometimes distortion is what you WANT.
These things are subject to personal preference, of course, but most of the classic "holy grail" circuits I've looked at seem to drive the phase inverter pretty hard. The tweed Bassman (and 4-input Marshall, by extension) PI exhibits lots of bias excursion in simulations I've done.
Cathodyne PIs can sound unpleasant in some cases, but this problem can be mitigated by using a large grid stopper.
FWIW, Merlin Blencowe (aka the Valve Wizard) has the opposite opinion, and there is a sizeable section of his valve preamp book devoted to deliberatly getting more distortion out of an LTP phase splitter, so that you can get overdriven tones without ear-bleeding SPL from the speaker.
As bmc0 says, cathodyne phase splitters do tend to have (fixable) nasty problems when overdriven hard. But you're not using a cathodyne, so that doesn't apply to this design.
A different question entirely is, do you want lots of amp distortion for a ukulele? Taimane Gardner is the only 'uke player I've heard who seems to want that. 🙂
One more comment; with the rather small-value tail resistor in your LTP, R5 & R10 need to be two different values if you want the best balance from your two output phases. Equal-value resistors here will actually result in a smaller signal through C9, than through C6, slightly unbalancing your phase-splitter.
Then again, an unbalanced LTP lets some even-harmonic distortion through, and it is the even harmonics that tend to sound sweet and bluesy. A perfectly balanced LTP only generates odd-harmonic distortion, which is more growly-sounding, and often thought more appropriate for rock music.
All told, it's nice to see a project that isn't yet another blind copy of an old Fender design!
-Gnobuddy
If you push an AC-coupled LTP into grid clipping, the bias current of one half goes down while the other goes up, causing lots of even-order harmonic distortion. Balance doesn't really matter in this condition 😉.
Last edited:
Well, yeah, sure! If you start with a balanced LTP, but then push it out of balance with heavy overdrive, sure, you'll get even harmonic distortion. 🙂bmc0 said:
IMO this is exactly the kind of thing that makes valve guitar amps sound much better than opamps with diode clippers, which produce boringly predictable waveforms, that don't change duty-cycle with input signal amplitude.
By the way, Merlin Blencowe's valve preamp book contains cautions about allowing a balanced LTP to become too far unbalanced in the wrong direction under heavy drive, as this can cause one of the two output valves to red-plate and kill itself.
This can happen because the duty-cycle of one of the waveforms from the phase splitter alters under overdrive, in such a way as to increase the average current through that output valve.
-Gnobuddy
Why the old loctal 7C5 output tube? A 6V6 would work there well and is likely to be more available and cheaper. Except for basing, 7C5 and 6V6 are electrically the same.
