First amp design - on track or off?

[Joe P]

Hi, good folks!

So, finally got around to designing my own amp. Now, this is my first tube-based design, and I have no formal education whatsoever, so I'm humbly asking for your expertise to figure out if I've made some huge mistake.
Now, I do realize that this might be a big project for a first build, but I've built larger non-amp circuits previously, so that's not really a concern for me. I also have no illusions that this is going to be cheap.

My goals are a very clean preamp with a switchable gain stage to get some fuzz, going into a very dirty power amp, with some unusual bits and pieces just for fun.
Signal path from input is as follows: half a 12AT7 as a common gain stage, going into a James tone stack, going into the other half of the 12AT7.
After that comes a relay switched 12AX7 wired up as gain stage + cathode follower as a dirty boost. After the boost comes a 12AU7 cascode into a LTP 12AY7 PI. Power section is cathode biased pair of PP EL84s.

Tweaked the tone stack in Duncan TSC to get the mid scoop in the vicinity of a 5F6.

To get a PT with enough current, I had to go quite high in output voltage as well, hence the twin 10W dropping resistors.
5U4G rectifier for maximum sag.

Disregard the weird look of the PT in the schematic, that's what KiCad had to offer, and I couldn't be arsed to roll up my own component...

[PRR]

R1 R2 6.4k is an exceptionally heavy bleeder, near as much heat as your power section.

Heater DC bias is taken from the WORST part of the DC. Why?

B+ decoupling is more art than science; but conventional practice is to use use much more than hundreds-Ohms between stages. (And before cheap electrolytics, we didn't use a whole crate of 100uFd for mere preamps.)

Why is R33 bias bigger than R35 load?

Counting on thumbs: you have A LOT!! of gain from volume pot wiper to output. Have you computed that? If you need less than ~~~50mV at Vol pot wiper to make full output, it is likely to hiss bad even at zero.

Typo: one plan shows 310V and the other shows 320V.

Inventory bad-dream: every tube a different type?

[OrganicEffects]

Doesn’t look like the gain switch is working how you intended? In its current state, signal goes from pin 2 to pin 7 and stops. Unless I’m not understanding what’s being switched here?

[kagliostro]

To me there is something odd on the relay section

are you planning for a single relay or two relays ?



I think you want the relay put on or disconnect the signal path from V2 section, correct ?

Franco

[Tony Bones]

V3a grid has no DC reference.

[Joe P]

R1 R2 6.4k is an exceptionally heavy bleeder, near as much heat as your power section.

Heater DC bias is taken from the WORST part of the DC. Why?

Lack of experience, first and foremost, did not think about that.

B+ decoupling is more art than science; but conventional practice is to use use much more than hundreds-Ohms between stages. (And before cheap electrolytics, we didn't use a whole crate of 100uFd for mere preamps.)

Didn't want to drop too much voltage, but I will keep that in mind when modifying the pre.

Why is R33 bias bigger than R35 load?

Copy mistake from two (hand drawn) versions back, R33 should be 500R.

Counting on thumbs: you have A LOT!! of gain from volume pot wiper to output. Have you computed that? If you need less than ~~~50mV at Vol pot wiper to make full output, it is likely to hiss bad even at zero.

I was aiming for a lot of gain late in the preamp, but I guess I overshot...
I'll have to have a think about if I should rework the staging, or just cut down on the gain.

Typo: one plan shows 310V and the other shows 320V.

Oh yeah, it's going to change anyway.

Inventory bad-dream: every tube a different type?

Indeed! Not planning on making more than one, so that's OK.

Doesn’t look like the gain switch is working how you intended? In its current state, signal goes from pin 2 to pin 7 and stops. Unless I’m not understanding what’s being switched here?

Yeah, moved the wires around when copying from paper to make it clearer, I see I got it very wrong. Sloppy. Will separate the parts on the next version so it's easier to follow.

To me there is something odd on the relay section

are you planning for a single relay or two relays ?

I think you want the relay put on or disconnect the signal path from V2 section, correct ?

Franco

Correct, and a single DPDT.



Thanks very much for the time and input! There's some work to do on this, but it's a learning experience, and I knew there would be.

[jjasilli]

"finally got around to designing my own amp. Now, this is my first tube-based design . . . I'm humbly asking for your expertise to figure out if I've made some huge mistake.  Now, I do realize that this might be a big project for a first build"


IMHO you're making a huge mistake.  Don't do it.  Build a simple, proven tube amp design first; maybe from a kit.

[2deaf]

The OT is missing a chunk of coil from 5 to 6.  Can't use shorting jacks on the speaker outs like that.

[sluckey]

The OT is missing a chunk of coil from 5 to 6.

That's just a drawing error. The chunk is really there. Still can't use shorting jacks though.

[st]

Haven't had time to really go through the schematic (and the thread), but cathode resistor on the power tubes seems unusually high.

[Joe P]

V3a grid has no DC reference.

Oh, right, missed that.

"finally got around to designing my own amp. Now, this is my first tube-based design . . . I'm humbly asking for your expertise to figure out if I've made some huge mistake.  Now, I do realize that this might be a big project for a first build"


IMHO you're making a huge mistake.  Don't do it.  Build a simple, proven tube amp design first; maybe from a kit.

I get the reasoning, but that would miss the main points of this, to learn about the actual process of designing a tube amp specifically, and to get an amp out of it that I really want.

The OT is missing a chunk of coil from 5 to 6.

That's just a drawing error. The chunk is really there. Still can't use shorting jacks though.

Yeah, that's just the closest symbol KiCad had to offer. Did not think about the jacks.

Haven't had time to really go through the schematic (and the thread), but cathode resistor on the power tubes seems unusually high.

Will double check when I make the next pass.

Thanks to all of you for replies!

[2deaf]

V3a grid has no DC reference.

Oh, right, missed that.

C20 (10uF) is on the wrong side of the relay.  Move it to between the CF and the relay and everything is cool.  Also, 10uF seems a little bit large to me.

[shooter]

to get an amp out of it that I really want.

m2c
build in stages;  PS up and running bullet-proof.  Tested with DL's to simulate tubes, spec'd to see if she's a hummer etc.
PA (pi included) again bullet proof, verified both DC and AC operation.

by staging the build you WILL find design flaws, construction flaws WAY easier,faster.  You will also have a better understanding of operation.

[PRR]

C18 C20 10uFd grid coupling caps are far too large; sub-sub-sonic response, cost, leakage.

Grid reference for V3A is "fixed" if you trade C20 for a Jumper then put C20 in the path from V2Bk to Relay. (As it is now, you have massive DC on that line, and massive DC load on V2B when relay switches.)

> build in stages

+1

[Joe P]

C20 (10uF) is on the wrong side of the relay.  Move it to between the CF and the relay and everything is cool.  Also, 10uF seems a little bit large to me.

C18 C20 10uFd grid coupling caps are far too large; sub-sub-sonic response, cost, leakage.

Grid reference for V3A is "fixed" if you trade C20 for a Jumper then put C20 in the path from V2Bk to Relay. (As it is now, you have massive DC on that line, and massive DC load on V2B when relay switches.)

Wiring around relay is messed up. DC is supposed to be blocked on all the lines going to it - didn't think about the coupling cap that should have been there.
Edit: I mean, I just treated the caps that are there to prevent pops as coupling caps and didn't think of the response.

to get an amp out of it that I really want.

m2c
build in stages;  PS up and running bullet-proof.  Tested with DL's to simulate tubes, spec'd to see if she's a hummer etc.
PA (pi included) again bullet proof, verified both DC and AC operation.

by staging the build you WILL find design flaws, construction flaws WAY easier,faster.  You will also have a better understanding of operation.

That's great advice, will definitely do it that way.

[2deaf]

For your consideration

[kagliostro]

2deaf you preceded me 

--

Relay arrangement proposal



(give it a try if you decide to build a circuit with OD)

Franco

[Tony Bones]

For your consideration

I'd still put a grid resistor on V3a.

Joe P, I admire your desire to do your own thing, but recommend that you avoid the attitude that you can design it right on paper, then build it and be done. Instead you should expect to try a lot of different things before you get to something you like. The paper design is just a suggestion of what direction you might eventually go.

So, echoing Shooter's advice, build in stages: build the PS and verify. Then build the output stage with a functional PI. Then add a simple preamp. Maybe just the first couple of stages of your design. At that point you should have a working amp. Then you can start adding and changing features. Don't get too preoccupied with fancy features like channel switching. A simple toggle switch will work fine for prototyping.

Eventually, your chassis will look like a complete mess, and the circuit in it might not have much resemblance to your original design. But that's ok. You can start over if you want with a clean layout and channel switching. Pick any awesome amp ever made and I promise it was the result of a journey of trial and error. And that was with people that have a lot of experience designing amps. This is your first one for cryin out loud.

[2deaf]

I'd still put a grid resistor on V3a.

What for?

[Joe P]

2deaf you preceded me 

--

Relay arrangement proposal

(give it a try if you decide to build a circuit with OD)

Franco

For your consideration

Oh, you've corrected the way I had it fixed, my approach was:


                   
              /o-| boost stage |-\
tone stack >-o                    o-> cascode
              \o----------------o/



I'm just going to assume your way is better and go with that...

For your consideration

I'd still put a grid resistor on V3a.

Joe P, I admire your desire to do your own thing, but recommend that you avoid the attitude that you can design it right on paper, then build it and be done. Instead you should expect to try a lot of different things before you get to something you like. The paper design is just a suggestion of what direction you might eventually go.

So, echoing Shooter's advice, build in stages: build the PS and verify. Then build the output stage with a functional PI. Then add a simple preamp. Maybe just the first couple of stages of your design. At that point you should have a working amp. Then you can start adding and changing features. Don't get too preoccupied with fancy features like channel switching. A simple toggle switch will work fine for prototyping.

Eventually, your chassis will look like a complete mess, and the circuit in it might not have much resemblance to your original design. But that's ok. You can start over if you want with a clean layout and channel switching. Pick any awesome amp ever made and I promise it was the result of a journey of trial and error. And that was with people that have a lot of experience designing amps. This is your first one for cryin out loud.

Yeah, I realize that "done" on paper = version 1 of who knows how many... Looking at a bunch of curves and numbers is one thing, hearing (and it working at all) is another thing alltogether. It'll take take as much time as is needed, I'm in no hurry.

[Tony Bones]

I'd still put a grid resistor on V3a.

What for?

To prevent very loud pops when switching the relay.

[2deaf]

To prevent very loud pops when switching the relay.

Both of the coupling capacitors feeding the relay are pulled-down to ground by 100K resistors, so the relay is just switching the V3A grid between two ground references.  There is no charge/discharge event that would typically accompany a very loud pop.

I never use relays, but Mesa does.  They have plenty of examples where the grid only has a ground reference from one contact or the other of a relay.

Sluckey's Marshall Dual 50 uses a switch wired such that the grid of V3-A only gets its ground reference from one contact or the other.  I strongly suspect that it doesn't have a very loud pop.

I don't discount the possibility that a mild, tolerable pop might be reduced by giving the grid a hardwired ground reference in addition to the relay, however.  I don't presently own any relays to give it a try and I don't have any plans to purchase any in the foreseeable future.     

[Tony Bones]

You might be right that the odd 12AU7 cascode might tolerate its input grid being 'free' very briefly, so perhaps "very loud" might be overstating the problem.

But, I still consider it bad form to let go of any grid like that even briefly. You'd never get away with it with an output tube. This application it might be tolerable, but I'd still add a grid resistor, no matter what Mesa has done. 

[2deaf]

But, I still consider it bad form to let go of any grid like that even briefly. You'd never get away with it with an output tube. This application it might be tolerable, but I'd still add a grid resistor, no matter what Mesa has done.

I agree.

[ALBATROS1234]

just found this thread and 2 things stand out to me,one the cathode resistor for the power tubes seems high  for a pair of el84 tubes that looks more like a 6v6 value. usually el84s use between a 100 to 200 ohm cathode resistor. second it seems like you are going to have excessive gain. my guess is this will create an oscillation which will end up as a shrill squealing sound if you turn the volume up past bedroom volume. this is just speculation from my experiences designing my own circuits. i disagree with the gent that said dont do it build something tryed and true unless this is your very 1st foray into tubes. if it is your 1st build try a champ style something to get your feet wet ,once that is working then start to experiment with tweaking that circuit by adding and subtracting to see what it does. i tryed to start off designing and the first couple were fails. but.... i learned alot. i kept experimenting and now i am building decent sounding stuff. what i found was that most of the average values are that way because thats what works best. so study as many schematics of amps you like the sound of and you will start to see a commonality between stuff. this is your bedrock. dont give up, read as much as possible, have fun tweaking its very rewarding and you will end up with a unique sounding amps.

[jjasilli]

Being the gent just mentioned, I'll jump back in.  If I understand JoeP's 1st post above, this appears to be his first tube build, his first amp build, and his first tube amp design.  Accordingly, Albatross & I seem to be in agreement re a build of a simple, tried & true tube amp.  Given the breadth & depth of the numerous constructive criticisms of virtually every aspect of the proposed "design", I stand by my Reply #6.

[shooter]

seem to be in agreement

I'm gonna dissent - for now - he's still on paper, it's cheap, fun, learnin a CAD program, all +'s

and I stand by my "build in stages"
additionally;
study LOTS of Dougs turret boards, layout, reasons WHY, etc
but at the end of day, I learnt; simple can make a GREAT amp AND I had to prove it with really complicated bad builds 

[jjasilli]

Sure, try your first downhill run on the xxx black diamond ski slope.  Ok, so you've got the wrong skis, the wrong poles and the wrong boots.  But heck it's admirable to do your own thing.  Maybe you'll even invent a new way of skiing.


(This is me being honest & supportive.    )

[shooter]

Not knockin, we've both got "the best interest" at heart 

I was a rock-climber 20 years, my 1st climb was 800+ft in Yosemite, The Harding route, 8 ppl in our party 6 (including me) clueless, 2 were moderate at best. we made it, not sure how, but when I took it up for real 10yrs later, I started with a "2 tuber" 

[jjasilli]

 


I hear that rock climbers get belayed on their first date!

[ALBATROS1234]

i also have to agree with simple is better. the best amps are simple minimalist designs. the modern stuff with all the crap jammed on the humongus circuit board doesnt sound better than a fender 5e3 deluxe , the simpler the more pure tone it will make and also let the tone of your guitar shine thru. the more crap and more tubes even the less you will hear what your fingers are truly doing. just my opinon mind you, i just look at a schematic for say a blues deville amp and dont get why fender doesnt just copy their tryed and true tone monsters that defined music instead of trying to reinvent the wheel. 

[tubeswell]

Hi Joe, the channel switch routing looked funny in the first schematic you put up. (Not sure if it’s been mentioned yet). Trace the signal pathways through the switch to make sure you will actually get a signal going thru with each switch connection.


(And I echo what everyone else said about simpler is better. YMMV)

[shooter]

get belayed

nice!
teachin newbies, "that 1st time", they're so scared sheetless they forget their own name    but 20yrs later they know who was holding the rope!

[Colas LeGrippa]

Off.

[2deaf]

Hi Joe, the channel switch routing looked funny in the first schematic you put up. (Not sure if it’s been mentioned yet). Trace the signal pathways through the switch to make sure you will actually get a signal going thru with each switch connection.

I think this is what he intended for the switching. 

[Joe P]

just found this thread and 2 things stand out to me,one the cathode resistor for the power tubes seems high  for a pair of el84 tubes that looks more like a 6v6 value. usually el84s use between a 100 to 200 ohm cathode resistor.

Yup, figured out that I had it calculated as if wasn't shared by the pair, so it's twice what it should be.

second it seems like you are going to have excessive gain. my guess is this will create an oscillation which will end up as a shrill squealing sound if you turn the volume up past bedroom volume. this is just speculation from my experiences designing my own circuits.

Sloppy calculations, and my thinking was on the lines of whack it with gain, I'll control it with the volume...

i disagree with the gent that said dont do it build something tryed and true unless this is your very 1st foray into tubes. if it is your 1st build try a champ style something to get your feet wet ,once that is working then start to experiment with tweaking that circuit by adding and subtracting to see what it does. i tryed to start off designing and the first couple were fails. but.... i learned alot. i kept experimenting and now i am building decent sounding stuff. what i found was that most of the average values are that way because thats what works best. so study as many schematics of amps you like the sound of and you will start to see a commonality between stuff. this is your bedrock. dont give up, read as much as possible, have fun tweaking its very rewarding and you will end up with a unique sounding amps.

I've been redesigning this, and have decided to cut it back and go for more of a studio amp than a gigging amp, so just the cascode into an SE output stage. There's so much gain stage juggling that the original ideas need, so I agree it's definitely better to get some real world experience with just those building blocks before I add on to it.

[ALBATROS1234]

yeah so build a basic concept and then if its not the exact tone you want you can sub this or add that and see what it does. rob robinettes sight helped me to understand what components do what to tone and i like the way he presents it. sort of like different guitar teachers methods are effective in different ways to different people but i like his method.i liked his how to voice an amp chapter it kind of breaks it down with common values that help a guitar amp builder wanna be like me.

https://robrobinette.com/Amp_Stuff.htm