So I'm going to call my amp the Cobalt Twin (twin for 2 channels).
I posted an earlier revision in another thread that was on a topic about a single part of the amp. I want to use this thread to get a little help fleshing out the design. I've still got quite a few questions.
Also, since the last revision I've added reverb and tremolo. My plan is not to make those part of the initial build. The way they're set up, I should be able to add them after the fact without too much trouble. I want to focus making a working amp first and then I'll add those once I get it working well.
A few things about the design:
1> 2 channels: 6U8 pentode preamp for chan 1, 12AX7 for chan 2. A switch grounds whichever channel isn't in use.
2> The diode in the cathode follower came from the "A useful mod for cathode followers" on this page.
3> The paralleled 12AU7s (with the LED biasing) came from the "Blues Preamp" on the AX84.com site. It's supposed to "thicken" the sound.
Specific Questions:
1> I wasn't sure how to calculate the plate voltages for U3a and U3b because of the fact that they're parallel and biased separately. Can someone explain that to me?
2> I posted a separate topic asking how to calculate the plate load impedance for U1a. I'm not sure how to do that with the mixing resistor. I need to know it to calculate the gain for the stage.
3> Are my gain calculations correct for the ones I have? I assume U1b is unity?
4> I assume I need to attenuate the voltage between stages. How do you do that? I would think creating a voltage divider between the grid stopper and grid leak resistors, but I haven't really read anything that explains it. Any pointers on where I could read up on that online would be awesome.
5> Are the pots on my reverb and tremolo correct? Are the Rate and Depth pots on the tremolo actually rate and depth or did I misunderstand it?
I'm not sure about my voltages. Please let me know if any look out of line.
Schematic Here
Thanks to you all for your help. Everyone here has just been incredibly helpful with my past questions. I really appreciate it!
I posted an earlier revision in another thread that was on a topic about a single part of the amp. I want to use this thread to get a little help fleshing out the design. I've still got quite a few questions.
Also, since the last revision I've added reverb and tremolo. My plan is not to make those part of the initial build. The way they're set up, I should be able to add them after the fact without too much trouble. I want to focus making a working amp first and then I'll add those once I get it working well.
A few things about the design:
1> 2 channels: 6U8 pentode preamp for chan 1, 12AX7 for chan 2. A switch grounds whichever channel isn't in use.
2> The diode in the cathode follower came from the "A useful mod for cathode followers" on this page.
3> The paralleled 12AU7s (with the LED biasing) came from the "Blues Preamp" on the AX84.com site. It's supposed to "thicken" the sound.
Specific Questions:
1> I wasn't sure how to calculate the plate voltages for U3a and U3b because of the fact that they're parallel and biased separately. Can someone explain that to me?
2> I posted a separate topic asking how to calculate the plate load impedance for U1a. I'm not sure how to do that with the mixing resistor. I need to know it to calculate the gain for the stage.
3> Are my gain calculations correct for the ones I have? I assume U1b is unity?
4> I assume I need to attenuate the voltage between stages. How do you do that? I would think creating a voltage divider between the grid stopper and grid leak resistors, but I haven't really read anything that explains it. Any pointers on where I could read up on that online would be awesome.
5> Are the pots on my reverb and tremolo correct? Are the Rate and Depth pots on the tremolo actually rate and depth or did I misunderstand it?
I'm not sure about my voltages. Please let me know if any look out of line.
Schematic Here
Thanks to you all for your help. Everyone here has just been incredibly helpful with my past questions. I really appreciate it!
Very theoretical. No grid stoppers, so could be unstable dependant on layout. The reverb tank, if a standard one is used, it will sound dreadful as the input impedance is 600R and you are feeding it with a 0.01uF cap!. Put it in the cathode and AC couple it with a 4u7F cap. That will give you some clarity. The LEDs in the bias look good but will make the sound uneven and distorted. Tremello; the depth control will alter the frequency. Remove the diode, use a 1k cathode resistor with 47uF decoupling. Place the depth control from the anode via decoupling cap of 100nF and feed the grid of your output stage and use standard grid 3 coupling. Take note from established manufacturers, Fender, Marshall, Mesa Boogie etc.
You can work out the 'gain' by reading the values from the datasheet of each valve.
Hope that helps.
You can work out the 'gain' by reading the values from the datasheet of each valve.
Hope that helps.
I'm not sure what you mean by "very theoretical." A lot of the pieces come from other places.
For example, the reverb driver comes from my current amp (an Ace Tone Elite). The identical circuit is also found in the Ace Tone Model 101.
As I mentioned before, the paralleled 12AU7 comes from the blues pre-amp on the AX84.com site of which several have been built (I've heard recordings from 2 and they sound amazing). That amp has a 7199 pentode & triode as stages 1 and cathode follower respectively. I'm using the 6U8 instead because it's cheaper and more common, but should be a reasonable replacement. There's doesn't have the 12AX7 in the second stage. I added that because 1> I had the other half of a 12AX7 to work with and 2> I figured it might make up for some of the loss to resistive mixing. I may be misunderstanding that, though.
The tremolo circuit comes from Valve Wizard's site. The only thing I changed was trying to add the Depth control. From the page: "By replacing the cathode resistor with an LED we eliminate the need for the large bypass capacitor, and obtain maximum gain and minimum output impedance at all frequencies in one fell swoop." It also has the happy side-effect of pulsing at the tremolo's frequency, so makes for a good visual indicator on the front panel. Hence my desire for that.
I'll add your changes for the depth control. Thanks!
As for gain, I wasn't talking about mu, but actual voltage gain: (mu * Rl)/(ra + RL).
Thanks for the comments. I'm continually comparing to other existing amps to fix/improve the design. For example, I originally had 2 input jacks (though only want 1) and I picked up the idea for the switch to ground one of the pre-amps from this amp.
As for the reverb, it may just be that that reverb driver works with that old original tank (I've never seen a tank like it before). If it sounds like crap when I put in the new tank, then I'll know how to fix it. Thanks again.
For example, the reverb driver comes from my current amp (an Ace Tone Elite). The identical circuit is also found in the Ace Tone Model 101.
As I mentioned before, the paralleled 12AU7 comes from the blues pre-amp on the AX84.com site of which several have been built (I've heard recordings from 2 and they sound amazing). That amp has a 7199 pentode & triode as stages 1 and cathode follower respectively. I'm using the 6U8 instead because it's cheaper and more common, but should be a reasonable replacement. There's doesn't have the 12AX7 in the second stage. I added that because 1> I had the other half of a 12AX7 to work with and 2> I figured it might make up for some of the loss to resistive mixing. I may be misunderstanding that, though.
The tremolo circuit comes from Valve Wizard's site. The only thing I changed was trying to add the Depth control. From the page: "By replacing the cathode resistor with an LED we eliminate the need for the large bypass capacitor, and obtain maximum gain and minimum output impedance at all frequencies in one fell swoop." It also has the happy side-effect of pulsing at the tremolo's frequency, so makes for a good visual indicator on the front panel. Hence my desire for that.
I'll add your changes for the depth control. Thanks!
As for gain, I wasn't talking about mu, but actual voltage gain: (mu * Rl)/(ra + RL).
Thanks for the comments. I'm continually comparing to other existing amps to fix/improve the design. For example, I originally had 2 input jacks (though only want 1) and I picked up the idea for the switch to ground one of the pre-amps from this amp.
As for the reverb, it may just be that that reverb driver works with that old original tank (I've never seen a tank like it before). If it sounds like crap when I put in the new tank, then I'll know how to fix it. Thanks again.
Last edited:
I neglected to mention (thought I had) that the reverb tank in my current amp died last year (mechanical failure). It sounded fine before it died. I have a replacement tank (8FB3C1B as per Enzo's advice) that I'm putting in this weekend, actually, so we'll see how that works. That's what I was referring to at the end of the post. Sorry if that was confusing.
In regard to question #4 above, I just use smaller grid leaks on the next stage to knock down gain. A voltage divider in the same position will also serve the purpose.
One thing I saw in the pentode section at the input, wouldn't you want the volume after C5 on the plate output rather than in the screen circuit? I looked at several schematics like Vox, and they all had it on the plate output after the cap of course.
One thing I saw in the pentode section at the input, wouldn't you want the volume after C5 on the plate output rather than in the screen circuit? I looked at several schematics like Vox, and they all had it on the plate output after the cap of course.
Okay, that's what intuitively made sense to me, but I just wasn't sure and I wasn't seeing it (at least not obviously) in some of the other amps I've looked at. Though I admit to being lazy and not calculating stuff out for other amps to see exactly what the gains add up to. Math isn't my forte, so doing it for my own design is exhausting enough. Unfortunately the texts I've read so far (and I've hit quite a few of the online ones) don't really go into multi-stage preamp design. Thanks for that. I'll go back and re-work my grid leak resistors.
And yes, I realized a couple of hours ago I was missing a volume after the first stages. Might come in handy 🙂
In the post I made regarding question #2, I got some answers and it also has me rethinking the way I've got the 2 channels. Because of the loss due to mixing, I'm thinking maybe it would make more sense to just completely shut down the channel that's not in use. I admit to not having thought that through entirely, as to how I'd do it. But I think it would make for a better sounding amp. Could I remove the current channel select switch and replace it with one that switches B+1 between U2a and U1a? Could I then remove mixing resistor and call it good?
And yes, I realized a couple of hours ago I was missing a volume after the first stages. Might come in handy 🙂
In the post I made regarding question #2, I got some answers and it also has me rethinking the way I've got the 2 channels. Because of the loss due to mixing, I'm thinking maybe it would make more sense to just completely shut down the channel that's not in use. I admit to not having thought that through entirely, as to how I'd do it. But I think it would make for a better sounding amp. Could I remove the current channel select switch and replace it with one that switches B+1 between U2a and U1a? Could I then remove mixing resistor and call it good?
I was thinking that you could use a SPDT switch. You feed the output of each channel into the two connections that are not the common. Then hook the common to the grid of V2b with maybe a 10-20k grid stopper after the common of the switch. You could get rid of the channel mixing resistors. This way, you are switching only signal voltage, so you could use a mini toggle and the channel would be ready to go right after switching. If you switch the B+, you will need a high voltage switch and you will have to wait for the tubes to warm up before using the channel. It depends how you want it to operate. Your way is fine if you don't need to have the other channel ready right away. I know I was the one who suggested the channel mixing resistors in the first place in your earlier design, but at that time you didn't have a switch to isolate the channels, so the channel mixing resistors were needed to isolate the channels from one another.
This is the latest revision. I've fixed up the channel selection the way you suggested (I definitely like your way better than the way I was thinking!) Just want to verify I did it the way you said.
Latest Schematic
I still need to go through and fix all the grid leak resistors to adjust the gains. Hopefully I can find some time tomorrow to do that.
Thanks so much for the input.
Latest Schematic
I still need to go through and fix all the grid leak resistors to adjust the gains. Hopefully I can find some time tomorrow to do that.
Thanks so much for the input.
Last edited:
Yep, that's it. Now about the "volume" being in the screen circuit....
Here's another idea to ponder. Remove VR1 and tie C13 to ground. Then replace R19 with a volume control. That way you have preamp vol. for both channels. It could be 1Meg for more gain or 500k for some attenuation.
One other thought, you calculated your pentode input stage as a gain of 54. This is typical of a triode 12AX7 stage. For more gain, raise the screen by 20-30 volts or so by changing the value of R18. Otherwise your gain is hardly any more then the triode input stage. Just food for thought.
Here's another idea to ponder. Remove VR1 and tie C13 to ground. Then replace R19 with a volume control. That way you have preamp vol. for both channels. It could be 1Meg for more gain or 500k for some attenuation.
One other thought, you calculated your pentode input stage as a gain of 54. This is typical of a triode 12AX7 stage. For more gain, raise the screen by 20-30 volts or so by changing the value of R18. Otherwise your gain is hardly any more then the triode input stage. Just food for thought.
I've already made some of those changes. I've replace R19 with a 500K log pot and followed that with a 220K resistor and a 220K grid leak. But I may change those values a bit. I've also changed the bias resistors in a couple of places as I'm seeing some issues with how I had things biased.
VR1 is kind of a tone/compression kind of control. It's discussed somewhat in Merlin Blencowe's "Designing Valve Preamps for Guitar and Bass" (start at Variable Screen Bypassing) and I also stole that from the blues preamp on AX84.com (the circuits are almost identical in the book and in the blues preamp). I can always remove it if it's a problem, but I'd like to hear what kind of effect it has on the tone. I'm under the impression it's pretty powerful.
With the additional gain from the other stages, do you think I ought to be pushing the pentode more? I was thinking it might be smart to have it and the 12AX7 a similar gains so that there won't be a huge volume difference between the two just, hopefully, a significant tone difference.
But if the tone is going to be better pushing the gain, then that's what I'll do.
VR1 is kind of a tone/compression kind of control. It's discussed somewhat in Merlin Blencowe's "Designing Valve Preamps for Guitar and Bass" (start at Variable Screen Bypassing) and I also stole that from the blues preamp on AX84.com (the circuits are almost identical in the book and in the blues preamp). I can always remove it if it's a problem, but I'd like to hear what kind of effect it has on the tone. I'm under the impression it's pretty powerful.
With the additional gain from the other stages, do you think I ought to be pushing the pentode more? I was thinking it might be smart to have it and the 12AX7 a similar gains so that there won't be a huge volume difference between the two just, hopefully, a significant tone difference.
But if the tone is going to be better pushing the gain, then that's what I'll do.
VR1 is an active treble boost using variable screen bypassing. I referenced the exact place you did to know it wasn't a volume! It's a great book by Merlin. That will be neat to have boost as well as cut,not just passive cut.
I wasn't sure what your design goals were, but if it is just tonal difference and not gain difference, you have done well. At first, I figured maybe the pentode was for more gain, a sort of hot channel. Similar gains will be great as well.
I wasn't sure what your design goals were, but if it is just tonal difference and not gain difference, you have done well. At first, I figured maybe the pentode was for more gain, a sort of hot channel. Similar gains will be great as well.
I'm a really big fan of the blues preamp from AX84 (based on the clips I've heard). It's just one of the most glorious sounding amps I've ever heard. But I also listen to and play blues and rock from light pop all the way through hard rock. I figured the pentode preamp will help to produce that thick, bluesy sound that I crave and the 12AX7 will give me more rock and roll kind of sounds. And I figured if I need any additional distortion that I don't get natively with the amp, I can always get that from any of the great pedals that are out there. (I also have a small stock of germanium transistors that are destined for some nice fuzz/distortion pedals).
I've got kind of a full day ahead of me, but I hope to find some time this evening to finish making the bias and grid leak changes that I want and get a new schematic up before I go to bed.
Designing this out and actually getting help on this has been really educational. It's really bringing into focus a lot of the things I've read about. When you read it (and I've been reading about this stuff on and off for about 18-20 months), there's so much that it gets a little overwhelming, hence the time it's taken before I've actually gotten to designing something. But now that I'm doing it, I'm actually starting to see the things I need to focus on and the things that aren't quite as important. There's still a ton to know and it's still a little overwhelming, but I feel like I'm learned ton over the last week or two.
I've got kind of a full day ahead of me, but I hope to find some time this evening to finish making the bias and grid leak changes that I want and get a new schematic up before I go to bed.
Designing this out and actually getting help on this has been really educational. It's really bringing into focus a lot of the things I've read about. When you read it (and I've been reading about this stuff on and off for about 18-20 months), there's so much that it gets a little overwhelming, hence the time it's taken before I've actually gotten to designing something. But now that I'm doing it, I'm actually starting to see the things I need to focus on and the things that aren't quite as important. There's still a ton to know and it's still a little overwhelming, but I feel like I'm learned ton over the last week or two.
Your depth control will now kill your signal to the 6L6. Suggest you wire trem out to the wiper and grid signal to hot on the depth pot.
Is the 6L6 meant to be triode wired?
Is the 6L6 meant to be triode wired?
Oh my. I didn't see that at all. Thanks! Fixed!
As for the 6L6. No, that was a mistake that I've somehow left in there for a while. I need to fix that as well. Thanks for pointing it out. I will probably add a triode mode switch, but I didn't intend for it to be in triode mode. I intended for it to be connected to a 315V tap on the power supply (which I have yet to design).
I'm currently working on fixing the pentode which is totally messed up (bad plate voltage, gain is wrong, currents are wrong, biasing wrong. Ugh...). :-(
As for the 6L6. No, that was a mistake that I've somehow left in there for a while. I need to fix that as well. Thanks for pointing it out. I will probably add a triode mode switch, but I didn't intend for it to be in triode mode. I intended for it to be connected to a 315V tap on the power supply (which I have yet to design).
I'm currently working on fixing the pentode which is totally messed up (bad plate voltage, gain is wrong, currents are wrong, biasing wrong. Ugh...). :-(
Hint for the pentode design:
Put a blue LED (or two reds) in the cathode circuit. That will hold the bias at a set value while you adjust Ra and Rs.
Further: Select a moderate value for Ra as well, say 100k, then you only have Rs to adjust to get the gain you want.
This is an empirical approach vs theoretical, and you don't get caught out by individual tube variation.
Put a blue LED (or two reds) in the cathode circuit. That will hold the bias at a set value while you adjust Ra and Rs.
Further: Select a moderate value for Ra as well, say 100k, then you only have Rs to adjust to get the gain you want.
This is an empirical approach vs theoretical, and you don't get caught out by individual tube variation.
I love triode pentodes, and they love to oscillate. The 6U8 (both sections) was designed for vhf use (up to 220 MHz). Don't be afraid to add anode bypass caps to your design if the sound isn't what you expect. I also find 10n coupling caps are usually big enough and reduce mushy overdrive.
So I'm going to work on getting my pentode preamp fixed tonight. I'm kind of rereading everything I have on pentode preamps and I came across this on Valve Wizard's site:
"In general, the screen bypass capacitor has a greater effect on the gain of the pentode than the cathode bypass capacitor. The screen bypass capacitor should always be connected to the cathode rather than to ground. Yes, I know a lot of circuits connect it to ground and get away with it, but it is bad practice."
But he doesn't go on to explain why (there are a few places where he annoyingly just states something and doesn't give an explanation. Otherwise, awesome site.) I was just wondering the reason for this. Is it that the screen shouldn't go below the cathode voltage? I can't seem to find anything that explains this.
"In general, the screen bypass capacitor has a greater effect on the gain of the pentode than the cathode bypass capacitor. The screen bypass capacitor should always be connected to the cathode rather than to ground. Yes, I know a lot of circuits connect it to ground and get away with it, but it is bad practice."
But he doesn't go on to explain why (there are a few places where he annoyingly just states something and doesn't give an explanation. Otherwise, awesome site.) I was just wondering the reason for this. Is it that the screen shouldn't go below the cathode voltage? I can't seem to find anything that explains this.
Here's the latest schematic.
I've completely redone the preamp. Most of the 6U8 preamps I've seen have 220K Ra 580K-1M Rs and Rk of about 2.2K. But according to Valve Wizard, voltage gain is roughly Ra * gm and gm is about 2.5mA/V, so with a 220K Ra that's over 500 for gain. Way more than I wanted.
I think these values are more reasonable.
djgibson51, I took your advice, partially, on the LED, though I went with Yellow or Orange instead of Blue because I wanted the bias voltage a bit lower.
Also, I forgot to run the pot on the screen to cathode instead of ground (as per my previous post).
I still need to rework the various Rg values and I haven't finalized the PA end yet, but I feel I'm getting much closer.
I've started collecting pieces for the build and finally sold the guitar that's going to finance it. Woohoo!
I've completely redone the preamp. Most of the 6U8 preamps I've seen have 220K Ra 580K-1M Rs and Rk of about 2.2K. But according to Valve Wizard, voltage gain is roughly Ra * gm and gm is about 2.5mA/V, so with a 220K Ra that's over 500 for gain. Way more than I wanted.
I think these values are more reasonable.
djgibson51, I took your advice, partially, on the LED, though I went with Yellow or Orange instead of Blue because I wanted the bias voltage a bit lower.
Also, I forgot to run the pot on the screen to cathode instead of ground (as per my previous post).
I still need to rework the various Rg values and I haven't finalized the PA end yet, but I feel I'm getting much closer.
I've started collecting pieces for the build and finally sold the guitar that's going to finance it. Woohoo!
I put my stuff into Tone Stack Calculator and it looks like I'm looking at roughly 9-18dB drop with the pots centered. 9 in the low (10-100Hz), 18 in the mids, centered around 500-600Hz and then ramps back up to about -10dB for the highs.
So my question is, is this meaning that it's cutting my signal to somewhere between roughly 45-65% or so?
So my question is, is this meaning that it's cutting my signal to somewhere between roughly 45-65% or so?
- Status
- Not open for further replies.