The Ultimate DIY Part 2 ? The KGITSOJC

[kevin gilmore]

Its a operational transconducance amp, and the load is 105 ohms.

Increase to 500 ohms for 10db of gain. add pot after the 105 ohms

and then a buffer.

[kevin gilmore]

So this one kind of came out in the mix.

 

http://gilmore.chem.northwestern.edu/nfd.pdf

 

unbalanced in, unbalanced out, about 8 watts into he6

gain programable with a single resistor (use a really good one)

and Thd < .01%

 

seriously sweet sounding.

will start on the board soon.

[spritzer]

Crap, just what I need... more stuff to build...

[Kerry]

All in due time...

 

This is fun, but agree it's hard pick the best.  I'm very interested in an amp to drive planar headphones.  It seems like this is going to make the list.

[Horio]

So I'm a bit confused as it seems like this thread has wandered off to some other designs more recently. Where does the KGITSOJC stand these days?

Edited February 20, 2013 by Horio

[kevin gilmore]

The posted front end with a lower power output sage makes a kickass

Unbalanced preamp.

Still trying to match this performance in a fully balanced version

[c12mech]

I was going to ask the same thing as Horio. 

 

In the end will this be able to do balanced/balanced, balanced/unbalanced, unbalanced/balanced, and unbalanced/unbalanced all in one unit.  I know that is a lot to ask for but it would be my ultimate preamp solution.

[kevin gilmore]

That is what i want to acomplish.

[c12mech]

If that is the case then I will keep waiting on this to materialize.

[Pars]

This sounds really promising Kevin. Are you thinking of keeping the higher power output stage... for a headphone jack?

[kevin gilmore]

i sure wish some company (any company) could supply

current manufacture complementary jfets. Then things

would be so much easier.

[Pars]

Any news from our friends at linearsystems?

[kevin gilmore]

They are either liars, or completely incompetent or both.

4 years to make a p-channel jfet, yeah right.

Or there are huge quantities of 2sk389 and 2sj109 somewhere

that none of us have access to that are preventing lsj109

from being released. Ayre evidently has 10's of thousands

of parts in stock, yet stax had to resort to glueing singles

together recently. Even the singles are not available.

[spritzer]

I'm tempted to ask Fairchild how many parts we need to order to get them matched in a single envelope... 

[wink]

Tempt away.

[c12mech]

I'm tempted to ask Fairchild how many parts we need to order to get them matched in a single envelope... 

 

If it meant being able to have available parts to design and build new circuits I would be willing to buy a fair amount of them myself. 

 

Is there no one that makes a decent SMD part that is equivalent to what is needed.  It seems that everyone is going SMD now.  I would hope they could make some good parts in those packages.

[kevin gilmore]

what i want is front end complementary jfets. Don't care if they are surface mount.

[spritzer]

Yup, package doesn't matter.  Fairchild does make a nice P-channel but I'd guess we'd need a lot of units for them to make complimentary pairs for us. 

[kevin gilmore]

here is the unbalanced/balanced to balanced front end.

 

http://gilmore.chem.northwestern.edu/ubaltobald.pdf

 

really sweet sounding.  0 db of gain balanced, -6db of gain unbalanced.

 

gain is controlled by R5,R32 so i could put a relay in that switches

the gain to maintain 0db (or any other number) when you switch

from unbalanced to balanced.

 

still even better with the that340 replaced with jfets.

mpsw06,mpsw56 replace 2sc2240/2sa970 when we can't get them.

Edited February 23, 2013 by kevin gilmore

[kevin gilmore]

first cut at the board

http://gilmore.chem.northwestern.edu/ubaltobald.jpg

[Horio]

Dumb question, but is the intent to use this between the "multiamp" board and an input selector board? Or would the gain be taken care of on this se/bal board and the "multiamp board would be eliminated.

Edited February 24, 2013 by Horio

[kevin gilmore]

This is instead of the multiamp board.
Input switching
This board
Pot of some kind
Diamond buffer output

 

There is nothing wrong with the multiamp, in fact it sounds great too.

But the input impedance is lower due to supersymmetry.

 

The idea on all of these is to eventually replace with jfets once

they become available. As it is, this new board is without feedback

of any kind, and is faster and more quiet. plus gain is a single resistor

and compensation is unnecessary.

Edited February 24, 2013 by kevin gilmore

[Horio]

That makes sense. Not to get greedy, but adjustable gain from the front panel would be a great feature.

[spritzer]

Relays and a stepped switch would be an easy way to adjust the gain.   Now if you want it to be a volume control then you are looking at KX-R levels of complexity though...

[Pars]

Preamps such as the Pass Aleph P use a simple panel mounted pot for gain control, one per channel, separate from a volume control.