Tweaks to NAIm CDS -Part 2- valve output stage

[Mr Tibbs]

Doesn't the RC filter interact with the variable source impedance caused by the volume pot and vary the frequency response according to pot position?

Not if the filter is buffered;

Input--RC filter--buffer--pot--gain/buffer--output.

Please could you name some - I would be interested to see how they get them to sound really good.

Linn preamps sound pretty good (esp if you are prepared to try them outside a Linn system). Mark Levinson preamps get excellent reviews (though I haven't heard them myself). Dynavector phono amps seem to enjoy almost universal approval for their highly detailed and involving quality. I think the CD12 uses OP627s (which I reckon come very close to wire-with-gain).

Mr Tibbs

 

[ron]

Doesn't the RC filter interact with the variable source impedance caused by the volume pot and vary the frequency response according to pot position?




Please could you name some - I would be interested to see how they get them to sound really good.

I have tried most of the raved about opamps + Jung's nested feedback scheme and although some are good, I have never got to "excellent" with opamps (or valves for that matter).

OPA627 sounds better IMO if you give it +/-18v, I know some will say thats too high but I used a pre for about two years with that voltage
The AD815 is a good op-amp although it comes in a different package to the 8pin DIL.
I built this one and was quite surprised with it http://www.diyhifi.org/forums/viewtopic.php?t=544&postdays=0&postorder=asc&start=0

 

[audiojoy]

Until someone tries a solid state version i am very happy with my valve version, I will tell you something more interesting I have discovered whilst hunting around for more info on the Naim Nac 102 (i just bought from E bay for peanuts). MR TIBBS was involved in a discussion on the official prison site of NAIM dated 2001 I think. DYNAVECTOR LV100 preamp BETTER THAN THE NAIM 52?????????????? what happened to that thread then????

and according to one of the members the dynavector was full of IC's not a discrete component circuit in site!!!


Luke

 

[ron]

Talking of discrete I've just built this to try out http://www.diyaudio.com/forums/showthread.php?postid=103883#post103883
I used BC184/214 transistors for now and will try the 2sc2240/2SA970 next, I've fitted Multicap RTX coupling caps

 

[S-Man]

Linn preamps sound pretty good (esp if you are prepared to try them outside a Linn system).
Mr Tibbs

I had an early Kairn which is a discrete design, the sound was OK but nothing special. The Kairn is rather better that a Kolector which uses some sort of car audio chip (and maybe opamps).

I guess you are talking about the newer stuff.

 

[Mr Tibbs]

S-Man, some time ago I made up a simple non-inverting buffer (battery powered), using OP627's. I placed it between the output of my CDX and the input of the 102 pre, so that I could hear what effect the opamps had on the sound. Sometime later, five people (two of whom are PF folk) were asked to listen to a few tracks of music of their choice, with and without the opamps in the signal path. Four of the five said they could hear absolutely no difference. One of the five thought the sound was better with the opamps in place.

Note; all five have a love of music, but little or no idea what an opamp is.

Mr Tibbs

 

[S-Man]

One of the five thought the sound was better with the opamps in place.

I find that I often go for the clearest/most detailed choice in this sort of listening environment. In psychobabble this is listening with the conscious mind. I find that high negative f/b circuits often sharpen up the sound very subtly and (in the short term) appeal to the conscious mind.

IMO listening with the subconscious mind is the only reliable way to choose - i.e. when you're relaxed and absorbed in the music and not actively thinking about it. Scientists call this "flow". I get "flow" big time on great single-track on my MTB, I find I get more "flow" with discrete circuits than with op-amp circuits (I'm talking preamp here 'cos my CDP is packed with NE5534s!!).

The active versus passive preamp debate seems to fit this as well:

Conscious/logical mind says; adding active circuitry can only make the sound worse
Subconscious mind says; I'm bored with listening to this passive, must get the line stage back in.

Maybe I have applied the OPA627 wrongly because it does not give me much flow.

S-Man
(Dont look so frightened,
This is just a passing phase,
One of my bad days).

 

[teddy_pardo]

Hi Teddy,

'BTW has anyone tried or considered using the naim TA729 boards at the CD output? By changing some capacitors/resistors this should provide same filtering capabilities but with discrete components'

Your idea of a discrete output stage somewhat akin to Marantz's HDAM modules, might be a neater and much simpler solution than mine, you might have a winner here, simpler to implement for all others to use, are you going to try it?

Luke

Look what I have found: the CDS555 is using descrete components:

"analogue stages, I-to-V conversion, filtering and output driver are built from discrete components"

Also, they say:"a separate low jitter clock circuit with its own multi-stage regulated power supply.", Wow, are they using a Flea?

Teddy

 

[martin clark]

analogue stages, I-to-V conversion, filtering and output driver are built from discrete components

This also applies to the CDS2 and CDS3.

Fleas in a 555, hoho. Naim have always known that clock supplies need to be very quiet; I recall this being mentioned in an interview with JV published some years ago. Interestingly, in my old CD2 the SAA7220 (which is the clock generator/oversampling filter chip) is fed by two cascaded regs from the raw digital supply - ie the 5v required is itself derived from a 9v regulated supply.

Moving one resistor and one cap to make this a tracking pre-regulator arrangement effected quite a noticable improvement too, pre-Flea

 

[rontoolsie]

very true...I have found that even the XPS2 has circa 10mV of ripple on the post regulated DC feed to the clock/digital filter/DAC chip, the negative impacts of which are very easily heard when the noise is removed (which I did by adding a second choke regulator inside the XPS2 to feed just the digital voltages, which now have less than 0.1 mV of noise. Digital noise=jitter=hashy sound with poorer timing.

This also applies to the CDS2 and CDS3.

Fleas in a 555, hoho. Naim have always known that clock supplies need to be very quiet; I recall this being mentioned in an interview with JV published some years ago. Interestingly, in my old CD2 the SAA7220 (which is the clock generator/oversampling filter chip) is fed by two cascaded regs from the raw digital supply - ie the 5v required is itself derived from a 9v regulated supply.

Moving one resistor and one cap to make this a tracking pre-regulator arrangement effected quite a noticable improvement too, pre-Flea

 

[Mr Tibbs]

"analogue stages, I-to-V conversion, filtering and output driver are built from discrete components"

Yes, but take care - the ccts here are probably nothing like the simple single rail pre-amp type we know and love (we love them because they are so amenable to tweaking). Split rail discrete (and IC) design can use techniques which have a huge effect on power rail noise rejection. A differential input stage (most often used in split rail design) can actually null the non-linearity inherent in the bipolar transistor - design is everything!

Mr Tibbs

 

[S-Man]

Yes, but take care - the ccts here are probably nothing like the simple single rail pre-amp type we know and love

Well the circuits I am talking about are designed by people who really know what they're doing! e.g. Jensen, Leach, Kaneda etc.