Pre amp therapy / therapy pre

Here is a project which started as a thought piece for a colored pre for a friend and eventually developed into a therapy project. I just wanted to build something completely different from my product line and have fun. Let me say first that I'm not particularly experienced in p2p wiring and this project is probably not going to be commercial and does not reflect my general building practices. That said, look at the (bad) pictures.

http://picasaweb.google.fi/jonte.knif/TherapyPre#5443655352963125618

The topology is very simple. First there is an amp based on ECC88 and pretty common adjustable feedback arrangement. Basic gain is set here.
Output stage is based on easy to get and high quality E81L pentode. Mode can be switched between pentode, UL and triode.

Grid feedback can be applied and different loads switched to the trannie secondary. So, there are plenty of different distortion characteristics available to be applied _in a controlled fashion_ . With feedback and triode mode it has pretty low distortion.

All power supplies are passive. Many part choices are based on what I happened to have in stock. Caps are Rifa and Mundorf elkos, and the rest are Soviet PIO.

The pre has EIN of -134dBu A-weighted and maximum gain is 93dB with 1:10 input trannies, in pentode mode without feedback.

Max output is high enough and can be padded to play with extreme distorions without clipping gear next in chain.

This is somewhat preliminary post to ask if folks are interested in the schematics (which I have not finished yet.) I will not be providing exhaustive help, so this project would not be for everyone but I'm very open to discussion and suggestions how to modify and/or improve any aspects of the topology.

There is nothing particularly revolutionary in the pre, but many knobs and levers Sound is good and distortions pleasing, ranging from easy and fat to soft clipping.
It has started a studio tour, so I'll have more info about the sound later.

-Jonte

You are kidding about the "messy layout" right? Looks very clean!

and yes, ofcourse I want to see the schematic  ;D

Interesting idea with switchable pentode, UL and triode.

Jonte Knif said:

The pre has EIN of -134dBu A-weighted and maximum gain is 93dB with 1:10 input trannies, in pentode mode without feedback.

Click to expand...

That's a lot of gain, and I see little to no shielded wires there. What kind of noise did you manage with max gain? Any trouble with 50hz coupling around? I'm especially curious because I have built a p2p preamps with very similar layout choices (but only about 70dB max gain), and always had issues there.

I see you have also discovered those green Russian PIO caps.  An optimal choice for a "distortion generator" such as this preamp.

That's a lot of gain, and I see little to no shielded wires there. What kind of noise did you manage with max gain?

Click to expand...

Well, EIN is -134dBU (-131dBU if pad or line is use) and output stage has about -90dBU A-weighted. You can calculate the max
output noise from these figures. At 93dB gain: -131+93= -38dBU. EIN clearly dominates at high gains. (above 40dB) And because input wiring is far from any noise sources there really isn't anything else than source impedance noise+input trannie winding noise+1st tube noise (which is pretty low)
In UL mode without FB there is some induced noise in right channel output trannie, because it loses some hum bucking. (they are not designed for UL, but work very well.)
The only thing I was not sure about was overall stability but the pre was stable from start.

I don't quite get what kind of noise you would expect to have in a carefully laid out pre. Mains trannie has built in magnetic shielding.
There are shielded cables in the most critical places but since they are black you wont see them very well. The white cable from input connectors is shielded Teflon stuff. All non shielded cabling is teflon too. 

Schematics to follow soon.

-J

this looks like really interesting!

Very nice!  ;D

I'm surprised at your tube choices. I don't think I've seen a mic pre that uses a ECC88 at the input.

I took ECC88 because I wanted to keep distortion, impedances and B+ low. You don't have too many common options available then. Unfortunately not all modern ECC88 are real and good, so I put some used Mullard in and it is perfect. This first amp should have ample swing available so that it never clips before the second amp, which is supposed to take care of the distortion.

Ideally there could be some dissimilar double triode instead to raise the impedance of the feedback network, but I don't know them very well.

Here is the schematics.

-J

Looks good , nice work
would be curious to hear sound samples

Rossi said:

Very nice!  ;D

I'm surprised at your tube choices. I don't think I've seen a mic pre that uses a ECC88 at the input.

Click to expand...

my mastering lab mic pre's use an ECC88 at the input

Very nice build! wher did you get these turret boards and case?
I like the originality of the schematic too.

I bought 3mm glass fiber boards, sawed, drilled and riveted according to my plans. Fairly easy but I hate handling glass fiber. They are _sturdy_. Also the output trannies have turret boards.

Case sides are 10mm machined aluminum with threaded holes for other parts and cut outs inside. Front, back, and the internal panel are from Schaeffer. Obviously this is expensive, but not that far from high end commercial rack cases. I like this case type very much. Easy and stiff as h*ll.

I made a 3D CAD drawing of the pre to check that everything fits nicely. How ever, when building I noticed that somewhat more space around the turret boards would really have facilitated assembly. Also I don't like the looks of the power supply side. Okay, perhaps I'm going too far in the quest for "beauty" but that's what a therapy project is all about, in topology and build. 

great.

Absolutely stunning! Looks like it could withstand an atomic blast.
I wish I could hear what it sounds like.

Thanks for clearing up the noise figures above.

Some questions about the input sections if you don't mind.

What is the purpose of the RC filter going to the center tap of input transformer primary, just smoothing the 48V? But then, 48V to center tap? I'm probably looking at the schematic wrong somehow.

The "floating" section between input transformer and first tube stage is interesting. This is the first time I've seen input transformer secondary not connected to ground. What does this arrangement achieve?

Looks like it could withstand an atomic blast.

Click to expand...

Well we do share over 1000km border with Russia

What is the purpose of the RC filter going to the center tap of input transformer primary, just smoothing the 48V? But then, 48V to center tap

Click to expand...

Smoothing and providing soft ramp. You can also see that there is an extra switch which cuts phantom in line-mode. This is part of mic/line 3*on/on switch. To center tap, sure, that's why there is a center tap in Sowter (and many other) trannies. Resistor has to be half of the normal 6,81k. What you gain is that phantom powering will not affect input impedance.

http://www.sowter.co.uk/specs/9610.htm

The "floating" section between input transformer and first tube stage is interesting. This is the first time I've seen input transformer secondary not connected to ground. What does this arrangement achieve?

Click to expand...

Look at V76.
This arrangement provides reasonably high impedance for feedback and causes local feedback at low gains, making the amp absolutely stable without any compensation. Off course you could ground the trannie and use a cap to grid potential. Matter of taste perhaps. The 3.3uF cap is as much in the signal path as a cap to the grid potential would be.  I like this option because it looks cool and confuses people

Note: Using the center tap phantom arrangement changes phantom feed resistance a bit when pad is on, so if this troubles use pad at secondary side or the common 2*6,81k arrangement. I doubt that this is a real problem.

Kingston said:

The "floating" section between input transformer and first tube stage is interesting. This is the first time I've seen input transformer secondary not connected to ground. What does this arrangement achieve?

Click to expand...

Common in pre-war stuff.   Also seldom seen, sec connected to grid and cathode versus grid and ground gives higher gain.   Also common to see output transformers connected to plate and cathode.  Look at early WE and RCA circuits, if you can find them.   NewYorkDave posted the RCA 76 console schematic, and it has input connected in similar manner.  

Nice work!

What is the circle in the bottom 13v PSU secondary with the "x" in it?  Does it say "Lamp" or "1 amp" there?  (The scan is a bit small, I can't quite read the text.)  I don't know the symbol for a lamp, that must be it?

Very interesting project!  Thanks for sharing.  Makes me want to try it.

EDIT: Adding: Also, what part number Hammond are you using for the 14H choke?

And what are the black cylinders between the tubes? Caps...?

thanks!

Kingston said:

The "floating" section between input transformer and first tube stage is interesting. This is the first time I've seen input transformer secondary not connected to ground. What does this arrangement achieve?

Click to expand...

The main reason for that is that, in order to make the NFB gain control effective, the cathode resistance of the 1st valve must be made much larger than needed for basic cathode bias. Let's say the 1st valve runs at 1mA, the cathode voltage is ca. 10.5V. So the grid must be biased about 10V positive in order to properly set the anode current. If it was a grounded signal, one would have to fit a DC blocking cap on the grid. But with a transformer, it is more elegant to just connect the hot side to the grid.

Somewhat better resolution schema:

http://www.knifaudio.com/Therapy/therapy_pre.jpg

There are four Mundorf dual section 2*50uF caps. 47u in the schema. You see two caps up side down near toroid. These are the first 4 caps in the schema in the power supply from left to right. The rest are those horizontal mounted ones. Pretty short anode circuits for ECC88.

You can see the choke number from pictures. Unimportant. Total supply current is 40-50mA.

If it was a grounded signal, one would have to fit a DC blocking cap on the grid. But with a transformer, it is more elegant to just connect the hot side to the grid.

Click to expand...

Yes, and contrary to what I wrote in my previous message it actually _is_ electrically more elegant, because no current will flow in that 3,3u cap and its quality doesn't matter, but if there was DC blocking cap to grid there would also be a grid resistor and therefore current. My mistake.

-Jonte