High gain two channel EL84 guitar amp

[Banjan73]

Hi.
I am new to this forum, but I have been recently thinking about telling you my story of how I construced my first 2 channel, high gain, push/pull EL84 amp.
Some years ago I constructed a 1W class A high gain-ish amp (2x12AX7) without really knowing what I was doing. I didn`t really bother to calculate too much bias etc. But the result turned out pretty nice.
My EL84 project this time started when a collegue of mine came with an old Tandberg Huldra 4 radio that he was about to throw in the garbage. The components left on the radio was just mains tranformer, choke, output transformer and the chassis. But these are the most expensive components in a tube amp, aren`t they? So I thought I should rescue the thing.
I had a quick brainstorming of what I wanted: Minimum two channels->one high gain, one «clean», an effects loop, preferably buffered, solid state rectification (I don`t like Fender «sag»). And push pull configuration off course (just to learn this technique also).
The Huldra radio uses originally EL41 power tubes and EBC41 preamp tube. EL41 was the forefather of EL84, so I decided to use two of these instead.
So, the first and most important issue here was to construct a proper power supply. After a lot of experimentation and research, including the choke almost catching fire due to too much current, I ended up with the configuration seen in the attached schematics .
Next was to construct the power amp. As you can see on the schematics, I ended up with two 2.2k screen resistors. These were not there in the beginning, but where added during the final tests to lower the plate/screen current (screen glowed red when playing loud). I also added two 1ohm resistors from cathode to ground to ease the bias adjustments. Oh! And the original output transformer turned out to be crap, so I changed that one. Be aware that the voltage plotted in on the schematics are not completely accurate. They are more a guidance. But the anode voltage is higher than the spec of the EL84. This is perhaps the reason for getting that much screen current.
Next up was to construct the phase splitter. I thought about using a long tail due to the saying of that this configuration offers less «sag» (again I hate Fender sag), but ended up with a cathodyne due to that this one only uses one half of the 12AX7.
For the buffer/post effectloop stage, I ended up using the other half of 12AX7 + a 12AT7. Not much abrakadabra there, I guess.
The preampstage was copied from the book «Design and construction of tube amp guitar amplifiers» by Robert Megantz. I made some minor changes to the tone stack sections, since Idecided to just use a «tone» control, not eq as the original and I had to adjust bias etc for my working voltage. Else, it is pretty similiar to the original schematics.
There is a lot of more details to this project, but fire away your questions, and I will try to answer as best as I can.
Conclusion:
This amp sounds pretty awesome! It`s a beast indeed. I have one Tommy Folkesson modified (4channel) Marshall 100W and one EVH 5150mk III as benchmarks. Off course this one doesn`t give the amount of power as those, but the distortion, harmonics and the cleansound also is pretty nice. When micked up to a PA, it works well in our band and we play pretty loud!
Soundfiles will come up soon.

[Banjan73]

Here are the schematics.

[Banjan73]

Some pictures of the working process and the result. I am currently thinking about building the amp into a combo. I am thinking about using a 10" speaker. Any tips of what to use? Celestion gold / greenback?

[Banjan73]

More pictures

[Banjan73]

And even more pictures ;-)

[tubenit]

That's quite an innovative project!  Thanks for sharing your ingenuity.


With respect, Tubenit

[Banjan73]

No problem :-)
I was actually a quite fun project. As I maybe said, the point here was to construct an amp that I really could and would use when playing in a band (rock/hardrock). As I hate using stomp boxes for overdrive/distortion, the only option was to make an amp with enough gain. AND it also needed "clean"-sound (it`s limited clean headroom here off course).

After I have posted some sound examples, I would love some feedback regarding which speaker to use for a combo, though ;-)

[pdf64]

Just to note that it’s unusual for the same HT node to feed the OT CT, screen grids, phase splitter and fx loop. There’s usually decoupling between them. If it isn’t causing a problem I suppose it isn’t a problem 

Combos will tend to exacerbate valve microphony and rattle issues, which high gain EL84 amps may be somewhat prone to.

[Williamblake]

I always thought something terrible would happen if you ground the centertap in that rectifier, maybe a typo?

[pdf64]

I always thought something terrible would happen if you ground the centertap in that rectifier, maybe a typo?

Have a good read through http://www.valvewizard.co.uk/bridge.html

[Williamblake]

.........

[Willabe]

I always thought something terrible would happen if you ground the centertap in that rectifier, maybe a typo?

Yes, with a FWB, but that's not a FWB. It's just a FW rectifier.

The 2 other diodes are for -bias, it's a FW -bias.

[sluckey]

I always thought something terrible would happen if you ground the centertap in that rectifier, maybe a typo?

It's only a problem if you ground the PT CT ***AND*** ground the anodes of D3 and D4.

[Banjan73]

Just to note that it’s unusual for the same HT node to feed the OT CT, screen grids, phase splitter and fx loop. There’s usually decoupling between them. If it isn’t causing a problem I suppose it isn’t a problem 

Combos will tend to exacerbate valve microphony and rattle issues, which high gain EL84 amps may be somewhat prone to.

Well. I haven't experienced any problems yet. I didn't know that decouple caps was necessary there either, hehe. Well, I'm here to learn.
There are though a couple of things I would like to point out here.
First, I see now that I havent named the components properly. Would be nice if we are about to discuss things.
Second. I skipped the secondary fuse. I know that it may be a little risky if I have a shorted tube or something, but there is actually some space problem on the chassis. I think perhaps I will solder one direcly into the circuit at some point.
Third. As far as I remember, I think I copied that biascircuit from the valvewizard page. Had to adjust a little bit though, to get the proper voltage.
This mains transformer is a little bit too powerful for this amp, actually. So I had to tame it. Thats the reason for using that type of rectification, and not a bridge rectifier.
Note also that, even there are two transformers in the schematics, this is actually just one in real life. I had just a problem finding the right symbol in the computer program..
And btw, its very nice to be able to discuss this build with someone who knows these things. My friends and my wife thinks lunatic building this crazy mess, hehe.

[shooter]

the image didn't copy well, hopefully you get the jest.
I "added" decoupling stages to the B+ line on top

[Banjan73]

Yeah.
Seen those before.
But in my world, those are low pass filters, not actually decoupling caps?
Or have I missed something crucial here?
The B+ has fairly little ripple already.
So, do I really need those? Are they there for some other purpose than to filter (and perhaps lower the voltage) a little more?

[shooter]

it's been 40 years since I had to set sit in class on AC fundamentals n theory. 
look at all the schematics for guitar amps and that is a "classic design" in the vast majority of 'em

[sluckey]

But in my world, those are low pass filters, not actually decoupling caps?
Or have I missed something crucial here?
The B+ has fairly little ripple already.
So, do I really need those? Are they there for some other purpose than to filter (and perhaps lower the voltage) a little more?

Decoupling caps? Wrong terminology. The resistors between the caps provide the decoupling, isolating one filter cap from another.

The B+ at that first filter cap connected to the rectifier has huge ripple. It can easily be 8 to 10 volts. That's OK to power a push/pull power amp because the common mode rejection properties of a P/P circuit cancels the ripple. But the earlier stages in the amp usually operate in class A and have no common mode rejection properties. They rely on the power supply to provide a ripple free B+. This ripple reduction is performed by addition R/C filter nodes. The more R/C filters you add, the greater the ripple reduction. For example, there may be 5V ripple on the first node, less than 1V ripple on the second node, and only a few mV on the third node.

There are other compelling reasons to use multiple filter nodes but the progressive ripple reduction should be enough reason to convince anyone to use multiple filter nodes.

If you have a scope you can see just how much ripple your single filter node has. Or, with a decent DMM you can measure the ripple voltage.

[mresistor]

My thought is if you are going to use an additional filter node up front in the power supply then one might plan for the additional voltage drop.  Sometimes it isn't significant.

[Banjan73]

Sluckey:
I have measured the ripple at some point. I don`t remember the values right now, but I as far as I remember, I approved it to be ok.
I can absolutely see your point in less ripple on the first stages of the preamp. Hmm. think I need to measure the ripple again, someday.
But! Anyway, I dont experience extensive hum and noise in this amp. Off course there will be some "hiss" when turning up the gain a lot, but compared to my two other professional built amps, its not that bad.
I learned some stuff when making that 1W amp  regarding the power supply. It turned out that rectifying the filament voltage made the amp a lot quieter. Thats why I did this this time also. And using shielded cables in the signal path (in the first gain stages) makes a lot of difference.
But anyway, I will someday measure the ripple throughout the B+ line. It may quiet the amp even more adding some LP filters there.

[pdf64]

A 6k OT with such a high HT voltage might cause the EL84 anodes to overdissipate at medium / high output power levels. It’ll be determined by how things play out with sag at the HT, anode and screen grid.

[Banjan73]

A 6k OT with such a high HT voltage might cause the EL84 anodes to overdissipate at medium / high output power levels. It’ll be determined by how things play out with sag at the HT, anode and screen grid.

Could you please explain this a bit more? I know that the voltage is a bit high, and that the tubes are getting quite a beat. As I said earlier, I had to reduce the screen current a bitt (2,2k resistors), since the screen grid turned red at high volume. But I have yet to see any red plating. Maybe I haven`t pushed it enough?

[pdf64]

Redplating is a good indicator that the dissipation far above the absolute max.
The absolute max will be perhaps 20% higher than the design max.
The design max is about 17% higher than the design centre, ie the EL84 12W rating.
eg the EL84 in AC30s tend to idle at about 17 watts, hence near their absolute max, with no redplating.
The AC30 gets away with that because it’s cathode biased and hence things tend to self limit, the dissipation probably doesn’t increase much higher with signal, and users expect to have to replace the EL84 fairly often.
The point being that the absence of redplating shouldn’t be taken to mean that the valve is operating within its limits.
And did you check for replating over a range of power output levels (continuous signal)?
There’s a few online load line calculators, I like nickb’s IVDS http://bmamps.com/ivds.html
Such tools can calculate max valve dissipation and output power etc.

[Banjan73]

I used a load line calculator when I first started drawing the power amplifier. But I believe I didn't follow it that exactly, hehe.
I know, as I said that the valves/tubes are driven hard. But I have a feeling they will hold (dont ask me why).
An alternative is to use a sovtek EL84M tube.

[Banjan73]

Ok.
I have used sunday afternoon playing a little on this amp. Sound examples are coming up, but first some pictures of my rig.
I hav tested the amp on a 1x12 redback 150W 16ohm, and a 4x12 marshall cab with original '68 25w greenbacks.
The recordings were done with a Zoom h2n. Guitar was a 2008 Les Paul standard faded.

[Banjan73]

Ahh.
Soundfiles are just a little too large.
Have to compress them..

[Banjan73]

Ehh.
Mp3 files are not allowed to upload?

[Banjan73]

Hmm. Maybe some zip files will do it?
Please excuse my horrible playing.

[pdf64]

Uploading them to soundcloud or similar hosting site, and providing a link, may make it easier to go about hearing them?

[Banjan73]

Uploading them to soundcloud or similar hosting site, and providing a link, may make it easier to go about hearing them?

Yeah it might. But I`m too lazy to create a soundcloud user ;-)
Anyway, I have now done some adjustments.
First, I have added another LP filter for the ECC83 anodes. I measured yesterday with my scope and I can`t actually detect any (!) ripple at all now in the preamp section. I have also measured more exactly the voltages for the anodes (and screen on EL84).
Second, I have a question:
I brought this amp to my former guitar teacher (a highly professional guitarist here in Norway), and one of his requests was to make a preshape filter on the high gain section. I have added an adjustable filter in the schematics which is attached. The filter is designed to cut frequenzies from~60hz to ~650hz.
My question is: is this an ok way to do this?

[shooter]

the only caveat might be the relay PS.  might induce noise if it's "common" to signal.
maybe start with a 9V battery as an isolated PS for the relay, Don't mix grounds.  If the circuit works n you're clam happy, Then maybe come up with a PS "solution" for the temp battery.

[Banjan73]

Dontbthink I understand this correctly. You mean that the filter after the relay creates more hum than without it? Or what do you mean?
BTW the relay is powered by the filamen 6.3V. It's not THAT pure dc, and you have a point if you mean that the power to the relay might cause hum. I will bear that in mind.
Anyway, my question was about the pre shape filter. Is that a correct way to implement such a thing?

[shooter]

if you mean that the power to the relay might cause hum. I will bear that in mind.

as to "will it work?"
that's up to the builder - guitarist, temp it in, you don't need the relay, just wire it "on".
test til you're satisficed it sucks or it rocks, adjust accordingly.  (that's the fun stuff!)

[Banjan73]

Updated schematics. I wonder if I ever will finish this amp, hehe
This variable high pass filter works pretty ok. I didn`t experience any "sag" when testing it, at least.
Now I will try to remove the "popping"-sound when changing channels. It is not always there. If Iswitch frequently, then its almost gone. Any ideas?
I will start with the battery trick, but if that doesn`t work, I have no clue...

[shooter]

"popping"-sound

that's the sound of your audio signal suddenly going away, then coming back.  the contacts breaking then making.
the larger the signal being disrupted, the louder the pop.


should be able to surf up "switch popping" and find a weeks worth of reading

[Banjan73]

should be able to surf up "switch popping" and find a weeks worth of reading
[/quote]
I realize that