Exotic component choice for a GC

[matt_uk]

Firstly, having just read all the pages of debate about exotic components and whether or not they are worth it I don't want this thread to end up the same!!

I'd like to hear from people who have used specific component choices in their GC and how they feel it has made improvements.

I am preparing to build my first GC and really don't want to be dissapointed, I will probably be selling my Arcam amp first so there will be no way back!

So far on my list is A GOOD POT, probably Alps Audio grade. Will use one for both channels, as it is surely better than 2 at half the price. And currently only have 1 toroid.

Other parts on my shopping list are MUR840 diodes. Possibly Black Gate caps (need to check price), and I already have five LM3875TF's from NS and an old amp chassis.

Interested in resistor choices, going for a NI based on this sketch by Peter. Vishay's are expensive, Caddock not too bad, and I can't find Riken anywhere in the UK.

With a Pot added is there any need for the 220ohm resistor? And what is a good ohm value for the pot please?

Regards,

Matt

[Ropie]

Hi Matt,

Firstly, if you can afford it go with a stepped attenuator as (so I am widely told) that will degrade the signal quality less than a simple pot.

I changed input resistor from a nice Welwyn metal film to a very cheap carbon one and din't notice any difference in sound quality, at least nothing that made me wince!

I don't think you will need the 220R with the pot/stepped attenuator in place.

I use black gates for the input caps (4.7uF) but I had to wait over 6 weeks for them from Audiocom.

BTW, where are you from in Northants? I lived there for 19 years (then off to college!)

[Nuuk]

I second the suggestion to use a stepped attenuator instead of a pot. IMHO the attenuator is clearly superior.

For the IGC, use the one shown on my Gainclone FAQ page on Decibel Dungeon.

I like a pair of Welwyn RC55 resistors for the feedback and matched carbon films elsewhere.

The Blackgate caps are unlikely to sound that much different to the Panasonic FC's given the huge different in price.

I use the MUR860's for the rectifier bridges and have no inclination to try anything else given the sound of any of my Gainclones.

Do use a transformer for each channel with either one or two bridges.

Use a heavy heatsink, not only for keeping the chip cool but for resonance reasons too.

For an input cap, a polypropylene is ideal.

But do remember that this is all personal taste - there is no ideal recipe other than the one you achieve by trial and error!

[carlosfm]

If you use a pot, you don't need the 22k resistor from NI to ground.
I agree that a stepped attenuator is better, but if you're not willing to go to that trouble, an Alps Blue is very good.
Or an Alps Black Beauty, if you can still find one.

[Peter Daniel]

Originally posted by carlosfm
If you use a pot, you don't need the 22k resistor from NI to ground.

If you are using a pot that is more than 20K (like 50k or 100k), you definitely need a 22k shunt resistor at NI input, otherwise the offset jump is pretty high at full volume (300mV or so). This is of course when no cap is used at inverting input.

I also noticed that in some systems, when using a pot, at full volume (no series resistance) the offset increases as well (plus some noise is detectable), that's why I added 220 ohm series resistor.

While pots are OK, using switching attenuator improves the sound substantially. I bought some time ago a Spectrol pot used in one of older ML preamps. This is whole metal body pot, linear with shunt resistor added. The matching between the channels is perfect. I thought this could be a nice sounding one, but to my disappointmet it sounded worse than my Nobles. Eventually I built my own switcher and the best results I obtained with a fixed series/switching shunt resistance. I'm using a parallel combination of 10K AN tantalum and 10K Vishay S102 for a total of 5k series resistance. The shunt switcher consists of Holcos and max resistance here should be 20K or so. I used here my older Sonic Frontiers attenuators valued at 50k with Electroswitch switcher and I'm using additioanly 22k Vishay from NI input to ground, so the shunt resistance never goes higher than 22k, otherwise offset increases.

Regarding the parallel tantalum and Vishay, none of those resistors sound satisfactory when used alone. Tantalum emphsizes the midrange region and provides authority, but alone it's too tubey without enough air. The Vishay, OTOH is too soft sounding, but has very good resolution at highs, together they work very well

As to the exotic components in the amp, it's hard to recommend anything really, as for best results they have to be chosen by trial and error in a given setup. While the ones in a drawing work for me, they also work well only with an amp they were tried in (the AMP-1) and in a different setup, they might provide less satisfactory results. The choice depends on attenuator, chassis, PS and even wiring. As always experimentation is recommended.

http://www.diyaudio.com/forums/showt...433#post304433

[carlosfm]

Originally posted by Peter Daniel

If you are using a pot that is more than 20K (like 50k or 100k), you definitely need a 22k shunt resistor at NI input, otherwise the offset jump is pretty high at full volume (300mV or so). This is of course when no cap is used at inverting input.

I have an LM4780 NIGC, with a 50k log pot.
Initially, I had the 22k resistor.
I removed it and it made a slight difference for the better in sound quality (transparency, detail).
No problems with DC with or without resistor, even at max volume (I always test this).


There are no two equal GCs.

Oh, BTW I remember a detail now:
I don't use an input cap.
But I use a 33uf/40v Philips electro bypassed with 150nf poly from Ri to ground.
The sound is amazing, specially after I bypassed the electrolythic.
And no DC.

[Peter Daniel]

That's the reason. If you use a cap from Ri to ground, you don't have problems with offset, but if you connect it directly, you shoudn't use higher resistance (from + input to ground) than your feedback resistance. At least this is what I observed.

[Peter Daniel]

Regarding the previous schematic I posted, in place of 22K Vishay VTA55 (in shunt position) 22K Caddock MK132 works much better. Much more 3 dimentional soundstage. I just learned it today

[neutron7]

peter daniel,

besides the feedback resistor, which one makes the most difference to sound? I dont want to break the bank! but I want to be convinced by CG when its done

its non inverting

[Rob M]

This guy has lots of cheap stepped attenuators.

http://cgi.ebay.com/ws/eBayISAPI.dll...category=39783