1176 - Sonic Relevance of less discussed Revision Differences

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

volki

Well-known member
Joined
Jun 9, 2004
Messages
178
Location
Berlin
(Sorry if this has been asked before - a search didn't reveal any significant info to me)

After having digested the schematics of all those revisions, I'm wondering about the sonic relevances of some smaller, less discussed changes which happened through the years.
To aviod confusion, I'll be referring to the revisions as indicated in this manual, where the original circuit has no revision marking, the first revision being A, then B (bearing in mind that the original circuit is alternatively called "A", with first actual revision being named "AB").

The big differences obviously are:
- original circuit and Rev. A ==> T600 attenuator + Xformer front end / 2-transistor preamp / 3-transistor Class A output amp + 5002 Xformer
- Rev. D/E ==> changed preamp to 3-transistor circuit incl. LN mod (revised coupling to the gain reduction FET)
- Rev. F ==> changed o/p amp to Class AB +11148 Xformer
- Rev. G ==> changed front-end to bridging active I/P stage


Now I'm wondering, which of the following, less discussed changes, would also affect sonic differences to a notable degree?

- different Attack time constants: R54 changed in Rev. A and B, C27 added in Rev. A and changed in B
- Changes in the associated network for Q-Bias Adjust (R35,59,60,85) in Rev. A and B (would probably only affect the line-up process, i guess)

- (if minor) resistor value changes / additions on the Ratio switches in Rev. A, B, C, E and F (=> Threshold sensitivity? Knee?)

- Different coupling of base of GR FET to the I/P signal: Before the preamp I/P through C1/R6, as per Original & Rev. A, or tapped off the preamp NFB loop through R6/C3/C9, as per Rev. B and onwards. (actually, I'm not sure why the GR FET is fed forward / fed back a portion of the I/P signal in the first place, anyone care to fill me in?)

- Original & Rev. A: both preamp and o/p amp have an FET in their 1st stage, from Rev. B onwards they're changed to all BJT circuits (probably this is negligable regarding differences of distortion... or is it?)

- Rev. A: additional Feedback path from Q6 emitter through o/p Xformer primary back to source of Q4
- Rev. B: Bootstrap from Q3 back to Q2 in the preamp (said to reduce 3rd harmonic distortion... by a noticeable amount?)
- Rev. E: Resistor value changes (and tolerance restricted to 1%) around Q9 in the S/C amp
- Rev. F: "Output" pot changed from 250K to 100K (different loading of Pream stage)
- Rev. G: switch from Zener to Voltage Regulator IC for the +30 rail (might result in different performance at high o/p levels?)
- Rev. H: change O/P xformer to 12614


I'm definitely going to build at least one version, maybe even incorporating alternative switching between different topologies within the circuit, but just out of curiousity I was also thinking about breadboarding versions of the individual building blocks and tweak around with the "minor" changes listed above... which I wouldn't bother to do if there are no significant sonic differences reported by anyone.

Any input is much appreciated!
Cheers

 
volki said:
...Different coupling of base of GR FET to the I/P signal: Before the preamp I/P through C1/R6, as per Original & Rev. A, or tapped off the preamp NFB loop through R6/C3/C9, as per Rev. B and onwards. (actually, I'm not sure why the GR FET is fed forward / fed back a portion of the I/P signal in the first place, anyone care to fill me in?)...

To reduce distortion.
 
David Kulka said:
volki said:
(actually, I'm not sure why the GR FET is fed forward / fed back a portion of the I/P signal in the first place, anyone care to fill me in?)...

To reduce distortion.

So is the trick that you superimpose a portion of the I/P signal to the rectified control voltage from the S/C? Which would mean that you're able to catch transients faster than with the S/C only? (Provided that C22 is the capacitor responsible for the release, the I/P transient would feed an extra charge to it, increasing the FET's base voltage and thus shunting more I/P signal to ground?)

But wait a minute  - if there's a negative input swing, the opposite would happen, driving the FET more towards pinch-off and thus reducing compression...? I'm confused :-#


BTW just to be clear - I'm not talking about the LN mod which adds another feedback path to the FET's source, there the linearizing effect is more obvious.
 
David Kulka said:
volki said:
...Different coupling of base of GR FET to the I/P signal: Before the preamp I/P through C1/R6, as per Original & Rev. A, or tapped off the preamp NFB loop through R6/C3/C9, as per Rev. B and onwards. (actually, I'm not sure why the GR FET is fed forward / fed back a portion of the I/P signal in the first place, anyone care to fill me in?)...

To reduce distortion.
That was the explanation when it was presented, but my analyses and the various simulations I've done tend to indicate that for the same amount of Gain Reduction, there is no difference between one and the other. I'm eager to be proved wrong...
However, the fact that the FET is submitted to half the voltage (one side sees the full LN amp input voltage, the other sees half of it) permitted a 6dB increase in operating level, or 6dB better S/N. this has been obtained at the cost of reducing the max GR.
 
You're talking about the LN mod here, right? A part of which is the FB path to the FET's source. Interestingly enough, a similar view has been expressed in this thread.

What I was actually referring to was the gate of the GR FET being fed a portion of the I/P signal through an R/C network (Rev. A: fed forward at the preamp's input, Rev. B: off the FB loop from the preamp's output).

 
No other opinions, anyone?

Zener stabilized vs. Voltage Regulator PSU? Would a Zener arrangement behave differently for large O/P swings when much current has to be sourced?

Different R/C arrangement for the Attack time in Rev. A as opposed to later ones? Would the Rev. A values mean an overall slower range of the attack time you can dial in, which would also contribute to the sonic difference of Rev. A?

 
Bumping this again, also interested in the original zener based psu. Are the parts available for a 2 channel unit? Which zeners and (I believe there are a couple inductors too) Where would one find these?
 
Back
Top