5w, SE version of the Marshall 18w

[blackcorvo]

I decided to leave the ECL84 PP project aside for now, and work on something more simple.

I really like the sound of the Marshall 18W, but that is way more power than what I would really be able to use.
I went and built a prototype of a 5W amp, using a Sovtek 12AX7 and a 6BQ5/EL84, Made in Brazil.
I still wanted that 18w-ish vibe, but I didn't know how to get it... untill I took a look at runofgroove's Eighteen and the Vox AC4TV's schematics.
In the Eighteen, the 2nd stage simulates the input of the long-tail PI from the original 18w, just tweaked to sound good with fets. On the AC4TV, they used a similar design... and after I noticed that, an idea came into my mind: what if I use the values from the 18w's PI input for the 2nd gain stage?

So, that's just what I did. And to be honest, it sounds wonderful. I had to add a bypass cap (2u2) in parallel with the "tail" resistor (47k) to get enough gain to push the EL84 into distortion, but I might just reduce it to 4k7, or something like that.
I'm still tweaking a couple things here and there, and I'll soon put the schematic here.

Might take me a while to post the schematic, because I can't seem to find any softwares that have tube symbols in their library. I'm tired of using MS Paint.

[John]

http://www.expresspcb.com/ExpressPCBHtm/Download.htm

Go there, download that. It's free and works great. Then,

http://www.el34world.com/Forum/index.php?topic=590.0

go there and download Doug's components he came up with. Also, making and saving your own components/symbols is dead simple. It took me a bit to get quicker with the program, but once you've got it schematics take no time at all, and it's a great tool for keeping track of all your tweaks.

Cheers!

[topbrent]

Seems like your goals could be met with a GA5, or a Tweed Champ/princeton buildup.
Pick up a valve junior and hot rod it. 

The 18w pre amp is a extremely simple circuit.  No need to overthink things, imho.

[blackcorvo]

Seems like your goals could be met with a GA5, or a Tweed Champ/princeton buildup.
Pick up a valve junior and hot rod it.  

The 18w pre amp is a extremely simple circuit.  No need to overthink things, imho.

I can't buy a Valve Junior. That stuff is REALLY expensive here in Brazil, and I believe they don't even sell those here. I'd have to import it (original price + shipping + taxes = my parents having a heart attack )... and the sound I got with my prototype is pretty much what I wanted.
I don't like the sound of the Champ nor the GA-5. The distortion sounds too muddy to me.

http://www.expresspcb.com/ExpressPCBHtm/Download.htm

Go there, download that. It's free and works great. Then,

http://www.el34world.com/Forum/index.php?topic=590.0

go there and download Doug's components he came up with. Also, making and saving your own components/symbols is dead simple. It took me a bit to get quicker with the program, but once you've got it schematics take no time at all, and it's a great tool for keeping track of all your tweaks.

Cheers!

I tried that already, but I can't seem to find these custom symbols in the list when I open the software...
Nevermind, problem solved.

[Willabe]

So, that's just what I did. And to be honest, it sounds wonderful.

Corvo that is great.     

 
                     Brad       

[loogie]

This is better for drawing schematics:

http://sourceforge.net/apps/mediawiki/tinycad/index.php?title=TinyCAD

[blackcorvo]

And here is the schematic!

I don't really know what values for the RC filters on the power supply would be ideal, because I don't really know the math behind them... but the "hum annihilator" resistor is something I've seen a lot in brazillian tube amp schematics, because here they used a resistor instead of an inductor in pretty much all amps. I must say that's an effective and smart way to get rid of both hum and expensive/heavy inductors.

I'll try to record a sound sample and post it tomorrow. Might be noisy though, since it's a prototype, my guitar isn't shielded properly and the pickup is a single-coil. But that's gonna be all from the guitar/poor ground connections in my house (in the 2 years I took guitar lessons, none of my effects was noisy when I used them in that school - even the one I built in a wooden box with no shielding at all! -, but they always sounded like a bunch of angry bees in my house... god knows why).

There's only one thing different from this schematic to the real deal: I'm using a 6v+6v/2A transformer, with half of the primary connected to the wall and the other half used for the HT supply (same idea I used for the ECL84 PP amp), using a voltage doubler to give me my 250v, but I'll probably just get two new 6v trasformers, and use them for the power supply, Hotbox style.

[Willabe]

If it were me, I'd raise the screen R to at least 500R/5w, even 1k/5w. To help protect the EL84.

          Brad     

[blackcorvo]

If it were me, I'd raise the screen R to at least 500R/5w, even 1k/5w. To help protect the EL84.

          Brad     

The PA is based on the original Vox AC4, and it had the screen tied directly to the B+. I thought that could probably damage the screen, so I added the 100R.
I have a 1k5/5w here. I'll try that. Thanks!

[PRR]

The screen is already dropped through 2.2K between B+ nodes B and C.

The C-R-C power filter is pretty common in good radios. If R is fairly big relative to C, the buzz harmonics roll-off twice as fast as a single capacitor. A good trial value for the R is 1X to 2X the size of your cathode resistor. Smaller is more power out but also more hum.

[Willabe]

The reason I said that is because, I think our friend Corvo is in South America and might not have an easy time getting a replacement tube if one burns out on him?

Corvo, PRR knows what he's talking about, we all learn from him.


         Brad     
   

[blackcorvo]

The screen is already dropped through 2.2K between B+ nodes B and C.

The C-R-C power filter is pretty common in good radios. If R is fairly big relative to C, the buzz harmonics roll-off twice as fast as a single capacitor. A good trial value for the R is 1X to 2X the size of your cathode resistor. Smaller is more power out but also more hum.

That C-R-C filter is working really well, I must say. Without it, I get what sounds like that background buzz from unregulated power supplies... but with it in there, that noise is completely gone. I might try 220R or even 390R (those are the 5w resistors I have here closer to 2x 150R).
I just wanted to know if the caps are too big or if it's ok to use those values.

About the 2k2 resistor, all I did was use the values from the 18w's power supply.

By the way, I wanted to ask here what should be the specifications for the PT? I thought of using two 6v transformers back-to-back, but that might make the amp a little too heavy... from my calculations, I'd need at least 6v@3A for the step-down transformer, and 6v@2A for the step-up transformer. For a real PT, the specifications would be: 130v+130v@60mA and 6v@1A. Are these values right, since I'm gonna use a bridge rectifier for the B+ supply?

I really want an option that lets me switch between 250v and 150v on the power supply. Less power = more hours of rock here in my house.

The reason I said that is because, I think our friend Corvo is in South America and might not have an easy time getting a replacement tube if one burns out on him?

Corvo, PRR knows what he's talking about, we all learn from him.


         Brad     
   

It's not really hard to find these most common tube types here. They're just kinda expensive (depending on the brand, if it's new/used, and shipping...). The EL84/6BQ5 for example goes here from 23 bucks to over 95 dollars (only the tube, shipping price not included).

I was lucky to find this one. It's an RCA EL84, by the way. I forgot to mention that.

[blackcorvo]

Okay, here's a sample:

http://www.4shared.com/mp3/uAsbsXEd/LoudAssBitch.html

start - 2:48 = "GetMeOuttaHere" mode (230v B+). I start with the volume at 1/4, then I turn it up to 1/2, and finally, full blast.

2:28 - end = "Loud" mode (150v B+). I start with full volume, then I move to 1/2, and finally, 1/4.

I wish I could actually play something right... but at least it gives you an idea of the sound.

By the way, all clips are with the bypass cap in parallel with the 150R resistor on the EL84. I'll probably add that switch in my schematic on the final project.
I actually like the sound without the cap much better than with it in there, but I didn't have much time today to sit and record more than that.

[Willabe]

I wish I could actually play something right... but at least it gives you an idea of the sound.

I just listened to your clip and I think you sound pretty good Corvo. Got the Marshall thang goin on. 


             Brad       

[jeff]

 In the 18 watter they do use 820 cathode resistors. But they're shared by two triode sections. Using double the resistance 1640(1500 closest common value or you could use two 820s in series if you want to be exact) would set the bias voltage the same as in a 18 watter.

Not sure about the plate resistor, if you'd double that two seeing as you're only using one triode and the
current won't be the same. I'd guess you'd double that too If V=IR then V=(I/1) X 2R.(original 100K for two triodes=200K for one). Maybe someone else has the answer.

BUT... if it sounds good, why rewire it!

[blackcorvo]

In the 18 watter they do use 820 cathode resistors. But they're shared by two triode sections. Using double the resistance 1640(1500 closest common value or you could use two 820s in series if you want to be exact) would set the bias voltage the same as in a 18 watter.

Not sure about the plate resistor, if you'd double that two seeing as you're only using one triode and the
current won't be the same. I'd guess you'd double that too If V=IR then V=(I/1) X 2R.(original 100K for two triodes=200K for one). Maybe someone else has the answer.

BUT... if it sounds good, why rewire it!

I'll try that. I have 1k5's and 220k's here.
It sounds a little too bright IMO, and reducing the current through that triode will probably fix that. It's not the bypass cap on the 47k, because I tried 10uF and it still sounded just as bright-y as using 2.2uF.

[jeff]

You don't necessarily need the 47K and cap.

 Eliminating that should give you more headroom because you're basicial raising the plate to cathode voltage without it there, but maybe that's not what you want.

 Leaving it will make the triode react differently than the original, But I don't think you can get that one section to react exactaly like it would if both sections were working together without some serious designing. And it probally doesn't matter if it's exactally the same as long as it sounds good.

Don't take this the wrong way, it's not meant to be an insult, but just think of it this way.
You can't just take two wheels off of your car and get a motorcycle. You're building a motorcycle, so you don't necessarily need four wheels.

Using the 47K/cap will probally give more preamp distortion.
 I'd remove it, but if there is some mojo in it, why not:
 play it,
 see how it sounds,
 use a jumper wire to short the 47K and cap,
 see how that sounds,
 go with what you like best.

[Baguette]

Hey Corvo,
Neat sounding beast.
Have you tried jumpering the 47k "tail" resistor? How do you like the sound? That mod must be a major increase in gain.
I do not quite get the point of this "half LTP" triode BTW. Looks to me like a regular triode gain stage, only super cold biased.

[jeff]

I don't know what to think of this either.

 I don't think it's super cold biased. Only because the grid resistor isn't grounded, rather it's connected to the cathode resistor, connected to the 47K resistor. So the grid VS. cathode voltage(bias) isn't the same if it were "super cold biased" with a 47K cathode resistor, but it seems like you're wasting a lot of plate to cathode voltage potential.

I'm not sure but it looks like a "normal" gain stage sitting on top of the 47K/cap. So the tube sees the B+ minus the voltage of the current flowing through the 47K resistor(V=IR). If the voltage is maintained by the cap, is this in essence just lowering the voltage this gain stage sees?
 I've always wondered if this would by the same/a good way of lowering the voltage to a tube?
...Anyone?
 

[spacelabstudio]

I don't know what to think of this either.

 I don't think it's super cold biased. Only because the grid resistor isn't grounded, rather it's connected to the cathode resistor, connected to the 47K resistor. So the grid VS. cathode voltage(bias) isn't the same if it were "super cold biased" with a 47K cathode resistor, but it seems like you're wasting a lot of plate to cathode voltage potential.

I'm not sure but it looks like a "normal" gain stage sitting on top of the 47K/cap. So the tube sees the B+ minus the voltage of the current flowing through the 47K resistor(V=IR). If the voltage is maintained by the cap, is this in essence just lowering the voltage this gain stage sees?
 I've always wondered if this would by the same/a good way of lowering the voltage to a tube?
...Anyone?
 

The 820R is the cathode bias resistor, the 470K is the grid leak.  Jumpering the 47K would just turn it into a standard triode stage.  As is, it's just a split load phase inverter without the non-inverting output hooked up to anything.  That never would have occurred to me in a million years, but now I might have to try it. 

Chris

[zambo]

I have built a few SE amps with a cold biased stage like that to act like a pi. I think it works pretty good myself. I use 3 gain stages then the cold biased one. It does make for a really bright amp. Sometimes I will use one at 10k instead. that sounds pretty good. I want to say i used 22k tail and 47k plate for mine though. Its been awhile since i built one like that. will research and report back. Great amp though Corvo!

[zambo]

Just looked. 22k tail and 47k plate. I did use a presence knob/nfb loop on this amp as well. 100k nfb resistor off the 8 ohm jack. I believe i copied it directly from the express schematic. same with the preamp up to the fake pi.

[blackcorvo]

I'm working on something here... I might add a NFB resistor + potentiometer to the amp. Maybe even use a pot with an on/off switch, so I can completely switch off the NFB if I want to.

Still playing with values... I don't know if I want to just make this a GA-5 clone or keep it like in my original project... I might also add a tone control to it...

So many ideas, so little time to try everything!

[Baguette]

How about that "tail" resistor jumpering thing Corvo (the 47k hanging from 2nd stage grid and cathode resistors to ground)?
That would turn the "half cathodyne" into a regular gain stage. I'm eager to know how the gain, tone and feel of the amp changes with this mod (even better, put it on a switch).
Please try it and keep us posted.

[blackcorvo]

How about that "tail" resistor jumpering thing Corvo (the 47k hanging from 2nd stage grid and cathode resistors to ground)?
That would turn the "half cathodyne" into a regular gain stage. I'm eager to know how the gain, tone and feel of the amp changes with this mod (even better, put it on a switch).
Please try it and keep us posted.

I'll test that. I had a busy day today and couldn't do any changes or tests on the circuit.
I'm looking for a proper PT for this project. A good friend of mine knows a guy that makes custom transformers, and he once got the guy to make him a tranny for a SE EL82 amp he built. The guy asked for like 33 bucks for the job. Good quality, too, so I'm looking forward for that!

[blackcorvo]

Okay, I've tested it with the 47k//10uf in circuit AND shorted, but didn't hear a thing different. I'll just leave it there, though.

Anyway, here's the latest schematic. I've added a "feedback" knob! It really changes the personality of the amp.
I don't have one with a built-in switch at the moment, but I'll get one, and probably use it to completely remove the feedback circuit, so I get full, raw madness whenever I want it.

[Baguette]

Okay, I've tested it with the 47k//10uf in circuit AND shorted, but didn't hear a thing different. I'll just leave it there, though.

<if the "half cathodyne" stage does not sound any different than a regular gain stage, why bother with it? Then the amp is "just" another GA5 / Valve Junior / Mini Z variation.
Still, I'm kinda surprised there's no difference in the tone.

[blackcorvo]

Okay, I've tested it with the 47k//10uf in circuit AND shorted, but didn't hear a thing different. I'll just leave it there, though.

<if the "half cathodyne" stage does not sound any different than a regular gain stage, why bother with it? Then the amp is "just" another GA5 / Valve Junior / Mini Z variation.
Still, I'm kinda surprised there's no difference in the tone.

Actually, I tested it again today, and I've noticed it doesn't cut as much of the highs when you lower the gain knob, compared to just the standard gain stage. I was only testing it at full gain before... doh!

And I still have to go and call the transfomer guy and check the price of my PT. College ain't making things easy for me this year...