Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Low Voltage 6N6p-OTL by H. Ahammer
New Posts  All Forums:Forum Nav:

Low Voltage 6N6p-OTL by H. Ahammer - Page 4

post #46 of 89

Unless I am reading things wrong:

The cap you mention should go straight to ground with its own wire. 

The "non-driven" grid of the input tube should also go straight to ground with its own wire.

 

It is VERY important that you DO NOT! Share the ground wire between these connections.  

post #47 of 89
Thread Starter 

Hmm... this thought might explain some noise I was getting in the previous build. This may be a stupid question, but how important is "VERY" important that some of these "grounds" aren't connected before they get to the star ground? I've basically got a "star ground" sitting just beside these caps, and will have a wire going to the PSU ground.

 

Well, actually, I have the non-driven grid (pin 7) connected to internal shield (pin 9) and then connected to pins 3/8 (cathodes) through a resistor before reaching the top "star ground".

 

Is this a clear cut case of "star ground means completely separate wires, go back and redo it!" ? I've always been under the impression that it's fine to attach everything to shared wires leading towards a central "ground point".


Edited by FallenAngel - 1/7/14 at 11:00pm
post #48 of 89

Shared wires going back to a central point is called by several names (christmas tree ground, hybrid star ground, maybe others), and it is a valid cost & time saving measure in mass produced amps where you have time to test what can be tied together and what can't without affecting performance. It generally works nicer than amps with a ground bus but not as time or materials intensive (both=money!) as an amp with a true star ground to build. 

 

In the case of DIY amps you can cut some corners with experience but until you are comfy with it star ground is nice and reliable. It means everything comes back to ground on its own wire. There isn't really much to ground here. Just keep things tidy. 

 

In the interest of cutting corners - I would ground things that appear to have no current going through them together. No reason to spend extra wire there. The 1Mohm input resistor on the "driven grid" and the "non driven grid" of the input tube can pretty safely share one wire although each channel should have its own ground wire here. The 10Kohm resistors to ground at the output of the amp for both channels can share the same ground wire as the headphones. 

post #49 of 89
Thread Starter 

Thanks for the clarification, that definitely helps and shines some light on why I was getting tons of noise last time. It also means "back to the drawing board" for the layout of this amp as I'm certain there are things that are grounded together which shouldn't be.

post #50 of 89
Thread Starter 

Good news and bad news:

 

The good news is that after wiring the star ground and fixing a few solder joints that had caused a little trouble, the amp is "complete" and plays music with basically no background noise/hum when no music is playing.

 

The bad news is that it sounds pretty awful with terrible distortion in the mids/bass. Incredibly muddy and "buzzing bass" when it touches anything above the high-mids.

 

I'm only testing with a cheap $5 pair of generic brand headphones but they sound a lot better than this when plugged into something else.

 

I'll try swapping different 6N6P tubes as I have a few, but only 1 pair of 6922 (brand new ElectroHarmonix E88CC).

 

Aside from that, the only things I can think of is the signal caps.

 

The coupling caps are Russian Teflon TF-1 0.22uF/400V which should be pretty awesome since I've used them before, but the output caps are a new experiment with Rubycon 1A2304A 410uF/300SV "PhotoFlash" caps bypassed by 1uF/400V Solen.

 

EDIT: Having done some more research on the Rubycon Photoflash caps, they're used often in the Bijou with good results, so I'm having a hard time figuring out what could make my amp so "muddy"!


Edited by FallenAngel - 1/11/14 at 6:33pm
post #51 of 89
Thread Starter 

A side note on the design - I've never seen this kind of wiring on an input tube before.

 

It seems to me that you could get away with using a single 6922 tube for both channels here or maybe drive both triodes of each 6922 or put them in parallel? The input signal goes into Grid 1, out of Anode 1, then through a pair of 7K5 resistors where it hits B+ between them, then goes to the output tube, while there's a ground going to Grid 2, and meets that signal as well.  I'm confused to say the least...

post #52 of 89
Thread Starter 

Reading a bit more here and understanding a little more http://www.freewebs.com/valvewizard/acltp.html

post #53 of 89
Thread Starter 

A little knowledge is a dangerous thing! Sanity check this differential monstrosity please!

 

post #54 of 89

That's not going to work. You have two out-of-phase signals from the input LTP driving the grids of the second tube in opposition.

 

That's all she wrote...

 

w

post #55 of 89
Thread Starter 
Good to know. smily_headphones1.gif what I'm not understanding is that the input tube doesn't have a pair of resistors like the 470k that I generally see in ltp implementations.

Little update, the amp is mostly cased up and playing music through my hd650s, but the bass is still really distorted and I'm having noise issues when I move the amp or touch the tubes, so I'm thinking solder joins may be troublesome...
post #56 of 89
Quote:
Originally Posted by FallenAngel View Post
 

A side note on the design - I've never seen this kind of wiring on an input tube before.

 

It seems to me that you could get away with using a single 6922 tube for both channels here or maybe drive both triodes of each 6922 or put them in parallel? The input signal goes into Grid 1, out of Anode 1, then through a pair of 7K5 resistors where it hits B+ between them, then goes to the output tube, while there's a ground going to Grid 2, and meets that signal as well.  I'm confused to say the least...

 

Its the same as a Melos. More or less. 

 

There are so many reasons to use a long tail pair (LTP) input stage. 

Lower gain

Lower miller capacitance = wider bandwidth

*generally* improved THD

Improved PSRR

Option of balanced inputs

 

So much cool. 

post #57 of 89
Thread Starter 

That's nice to know that it's similar to the Melos, but still really confused on why the input tube looks the way it does...

 

Shouldn't the second grid be joining the first grid through a resistor (equal on both sides), similar to this LTP example:

post #58 of 89
Quote:
Originally Posted by FallenAngel View Post

That's nice to know that it's similar to the Melos, but still really confused on why the input tube looks the way it does...

Phase
post #59 of 89
Thread Starter 

Simple enough :)

 

Ok, then I'm absolutely clueless as to why it sounds very muddy in the low end. I'll go over the entire circuit with a fine toothed comb and fix any and all solder joints which look off to hopefully track down the noise issue. After that, I guess I'll replace the 0.22uF cap with something else that I have on hand, but aside from that I'm stumped. The same issue twice, even after rebuilding it... I'm thinking it's either the tubes I have or that cap.

post #60 of 89
The next step really should be to measure voltages at the various nodes. Can you post a schematic with the measured voltages written in? If you have not worked with high-ish voltages much, it is a good idea to pick up a pile of $5 meters from Harbor Freight, connect them all with clip leads, then turn the amp on rather than poking around a live circuit.
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Low Voltage 6N6p-OTL by H. Ahammer