Two channel class A amplifier

[Banjan73]

Hello!
Its been some years since i last visited this forum. In the meantime, I have worked with two amps on my own.
So I`m thinking, lets present the last one I`ve made first. This one is actually just finished, and it is not tested properly yet. It might need some tone tweaks, but that is minor changes in the circuit, so I guess it`s ok to present this one now.
The concept here is based on the Fender champ 5F1 amp, but with an EL84 tube instead of the 6V6. The transformers are from an old radio.
This amp is made as a collaboration with a colleague of mine, which share my interest of making tube amps. However, he has never made a guitar amp befor, only hifi-amps. So this one was a new experience for him.
Anyway, my colleague did all the basics on this amp, woodwork, chassis etc, and he constructed basically a straight forward "champ"-type amplifier based one the 5F1 schematics (with some adjustments off course due to different transformers and tube).
Then it was my turn to hotrod this thing.
This time I though the concept as follows: two channels ("clean" and high gain) with just a amp switch (no footswitch), effects loop, no tone pots /equalization.
In other words, quite straight forward... Ehh.. Not.
I was struggling a lot with hum, white noise and oscillation in this one. But after a while, I have now ended with a result that seems, after a few low volume tests, quite nice.
Well.. Enough talking. Lets see the construction:

[Banjan73]

Some pictures

[Banjan73]

and more

[Banjan73]

Now. A couple of comments.
First: on the schematics you will see R20 10ohms. That one is not going to be there probably. Its just a survivor from a unsuccessful low pass filter that I was trying out there. But since its only 10ohms, I have just left the resistor there.
You will also see that the build is quite messy. This amplifier has not been built to look nice inside, but to sound good. It also suffers from a lot of experimentation, so that also explains a little bit of the mess inside..

[Banjan73]

Oh. I forgot.
At the bottom of the schematics, you will see the recovery stage after the effects loop. This one is now not connected because it created a lot of white noise. So I removed it. I will test the amp today at a band rehearsal with my effects board. Hopefully it will handle all sorts of effects, but we will se.

[dogburn]

Interesting project! Is it loud enough to play with a band? And it looks like there is no negative feedback loop - I would have expected one for the "clean" channel.

[Banjan73]

Thanks Dogburn.
Well.. Loud enough is relative. Yesterday we played without drums, and quite laidback. Actually, I missed some more volume in the band to get a bit better distortion sound. I got the impression that it sounds a little bit thin at low volume. I might put on some anode decoupling capacitor or larger grid resistors to cut some treble in this one. Or maybe a better constructed low pass filter than the last one I made.
Sadly I realized that the effects loop didn`t answer to my expectations. When I use my full pedal board (8-10 pedals?), the volume drops significantly vs if I do not use effects at all. It also seems that the return side of the effect loop have some impedance issues. I get more distortion than I want with effects (and as I said low volume).
So I guess I got to try again (now is the third time) to get that recovery stage working properly. Any tips regarding this to avoid white noise? I already use shielded cables and DC for the filament.
Oh, btw. I considered negative feedback, but I dropped it to get as much sound as possible out of this tiny thing.
The loudspeaker is also extremly efficient. It is an old Norwegian SEAS symfoni 10" from the same radio I mentioned earlier. I think the efficiency might be at about 105 dB/W or something. It absolutely screams!

[Banjan73]

Under is my thoughts regarding the recovery stage. Are there any serious design flaws here that I don`t see?

[dogburn]

I usually put a switch on the negative feedback circuit, so it can be switched off to give more gain. As for the effects loop, that's beyond my level of useful understanding - if I were doing that, I'd look through the forum for successful implementations.

As for volume, I can say the AA764 Champ based amp I built ended up being surprisingly loud, and I expect it could do well with a quiet drummer or other subdued percussion.

[Banjan73]

Yupp.
There was originally a switch like that for feedback (my colleague built it that way). As I mentioned, I had a lot of noise issues in the beginning when I modified the amp (for two channels, effectloop and highgain). One of the major issues was interference from the primary of the output transformer. I solved this one by putting the cables outside of the chassis (see the attached picture). During this troubleshooting session I also removed the feedback circuit because of noise-paranoia. Anyway, when I tested the original build, I found out that the feedback wasn`t anything I liked. So in my world it was no loss removing it.
Well. An update. Yesterday I put on a 900pF capacitor over the 100k anode resistor of U3A on the schematics. The first tests sounds promising. I think that the ice-pick distortion now is a little better..
The next now is to really think thoroughly through that recovery stage and get that working properly..

[Banjan73]

Hi again (and sorry about my nagging here).
Attached is my new thoughs around the recovery stage.
Aiken amps has a nice article about resistors and e.g. Johnson noise:
https://www.aikenamps.com/index.php/resistor-types-does-it-matter
Now. Originally, I used a 1M resistor as a grid resistor. I use metal film 1W for this stage only. When reading that article, it seems like the thermal noise increases significally the higher the resistor value you have. So now I will try with 470k as grid resistor. It should be enough to maintain the input impedance.
And I will decouple the cathode to increase the gain in this stage since 12AU7 is a low µ tube (signal/noise ratio).
May I be into something here to reduce white noise?

[Merlin]

Originally, I used a 1M resistor as a grid resistor. When reading that article, it seems like the thermal noise increases significally the higher the resistor value you have. So now I will try with 470k as grid resistor.

You want to make series resistors smaller, not shunt resistors. 1Meg or 470k are both fine though. If you're having trouble with white noise, R2 is most likely the culprit! Make it 22k or something.

What bothers me more is your FX driver stage, which can dump a hundred volts of signal into your wimpy Boss Metalzone...

[Banjan73]

Well.
If you see my first schematics, the recovery stage (return stage) is not connected. That how the amp currently is connected and I experience no white noise. So I dont think that R2 creates any noise. R2 is also that values to obtain a little miller effect at the beginning of the amp. I therefore believe, based on no noise without the recovery stage, that the source of this is in that particular stage. It might off course be due to poor amplification in that stage (bad signal/noise ratio), but this is only my thoughts.
You mention that the cathode follower before the loop might dump a lots of voltage into the loop. The CF seems to me as a normal cathode follower, so if you please could explain this a little more, I would appreciate it.

[Merlin]

Well.
If you see my first schematics, the recovery stage (return stage) is not connected. That how the amp currently is connected and I experience no white noise.

When you added the recovery stage it adds more amplification, so the noise from earlier stages (e.g. R2) became louder in the speaker.

[Banjan73]

Yeah. I can see that one, although the recovery stage doesnt provide much ampiflication vs the other stages.
But I will bear that in mind. I eill first now try my theory and if thats not working I will try to lower R2.

[Banjan73]

Merlin:
First: I discovered yesterday that I actually have removed R2 sometime. The schematics is just not updated. So, there is currently no grid stopper on the first gain stage. 
Second: I replaced the 1M shunt resistor in the recovery stage with a 470k and decoupled the cathode with a 6,8µF.
Result: Still white noise. Maybe a little less, but still anoying.
So... I follow your theory with noise beeing amplified through all the stages. The thing is that I have done a lot to minimize noise in the first stages, and as by now, I am in a stillstand here. Completely out of ideas. I use shielded cables the whole signalpath (except between recovery stage and EL84 tube + between the jacks of the effect loop), metal film resistors, DC for filament.
So, do anyone have more ideas here?
Attached is the schematics as it looks like now.

[jbrrrrr]

---edited because my reading comprehension is questionable---

[Merlin]

Are you using a shorting input jack? You should assess noise with the input shorted to ground, otherwise you'll be listening to the noise of the input grid leak R1. Also, you might try removing the input cap C17, as it has a very high reactance at audio frequencies which stops R1 from being effectively shorted. You can test this quickly by shorting an aligator lead across R1; what happens to the noise?

[Banjan73]

Yes, I use a shorting input jack. I will try that test later today.
I will also try to shorten the grid on the different gain stages one by one, if this first one  doesnt help (starting from the first one) :-)
Thanks for the input! I suck in troubleshooting;-)
(BTW, that capacitor is there for bass filtration. But I guess you already know that:-))

[Banjan73]

Sh%& Merlin!
You truly are a wizard! Where is Excalibur these days! Hehe.
Update:
I did what you said, shorted C17. No change in noise. Then i shorted the grid on gainstage 2. Much less noise!
So, todays session was to replace R26 from 220k to 10k. Much better, but the high gain was still a little noisy.
So R7 was then lowered from 220k to 27k. And VOILÁ! Almost noisless! Fantastic! Hurray!
The amp sounds great now, both "clean" and in high gain mode. A little noise off course, but this is a high gain amp, and you will always have some noise. I haven't tested it with effects yet, but I recon that will work now as the recovery stage is in place also. Maybe a little "dark" distortion? But thats a quick fix since I used 3 capacitor a' 300pF in paralell for the anode decoupling. So I can just cut one out and get 600pF. We will see.
BTW..
This amp is a request from a friend of mine. He is the guitarplayer in Wig Wam (ref "do ya wanna taste it" Peacemaker series).
He tested my main amp in which is a thread here also (I'll find the thread and link it in into an answer here), and he wanted one.(This will probably be a practice amp. He uses now 50w EVHs or Mesa Boogie MK IV)
I decided that that build was too time consuming, so I ended up with this simpler version.
He will off course have to test it himself. And then I'll maybe do some tweaks..

[Banjan73]

To much pop ups on this site when using my phone. So its impossible to copy the link. Anyway, search "high gain two channel EL84 guitar amp" on this forum.

[Banjan73]

Intalled brave on my cell phone now. Much better!
I tested the amp with effects now. In the living room in our appartment, so I had to keep the volume low.
This amp really screams. "Clean channel" also sounds quite nice and breaks up pretty fast (off course when its 3W). I have attached a picture.
Now, my next thread will be something similar, but with another goal.

[Banjan73]

Too large file on the photo. Here it is:

[kagliostro]

To "post" big images upload It in a free host and share the image using the URL with the image tag of the forum (the Mona Lisa image)

Franco

p.s.: IMGUR is a good free host and don't delete your uploads after a certain time but recently UK people can't access to what is hosted there

[pullshocks]

Godt gjort.


Hilsen, Pullshocks