KGSS Mod(s)

[kevin gilmore]

Yep, and once you terminate the cables in their characteristic impedances, the

significant sound differences between even the most different cables become

so small that you will have serious trouble telling the cables apart.

Something the cable companies obviously do not want to happen as their

business will fall into the trash. Also makes it harder for outside interferrence

to effect the sound.

50 ohms (or really 46 to 110 ohms) is not a problem for properly designed

audio equipment. Even buf634's or the elantec equivalent are actually designed

to do video over 50 ohms. Even the toy opamps that drive many headphone amps

can drive grado's which are 32 ohms.

If the impedance really bothers you, you can always make up some ultra trendy

300 ohm twinline. Made from pure silver, with a large diameter silver shield.

Can't do that with OTL tube gear however. Requires a matching transformer.

Edited January 31, 2011 by kevin gilmore

[bobkatz]

I've never used the power-matched Krell systems. You're saying a 50 ohm source impedance and a termination resistor at the input to the amplifier? I had a Krell KSA-100 amp for years, class A, excellent sound, but standard high impedance input. "I'm from Missouri, you'll have to show me", so I'll continue with my own experientially-derived reactions to low impedance loads (less than 10K) on all kinds of circuits that I've tried. I didn't come about this conclusion by theory or measurement, but by listening. Once I was using a variable o-pad at the input to my power amplifier to adjust its sensitivity. The driver circuit is the excellent discrete opamp in the Cranesong Avocet, with excellent headroom (it clips at about +37 dBu balanced). This driver circuit is rated to drive most any load down to below 600 ohms. For various reasons, during a certain period of time, the attenuator turned out to be presenting a 1k ohm resistance to the Avocet. I redesigned the gain structure to produce a 10K or higher load and immediately noted a better sound quality to the system, a more open quality. So I tend to stay away from low-impedance loads on my line drivers. All levels in my system are calibrated, so the level to the loudspeaker was matched to 0.1 dB in all these comparisons. I realize there is a contradiction here in that RF circuits and loudspeaker driver circuits themselves, are able to drive very low impedance loads, but "once bitten, twice shy" so I've chosen to stay away from low impedance attenuators.

[bobkatz]

About 25pf for the fet is right. Add in wire and connector

and its 30 pf for sure. There is likely nothing left available

in solid state that is going to do any better and still have

an extended frequency response.

I run a 50 ohm system with 50 ohm correctly terminated cables.

The krell stuff is designed for this. So is the Ayre stuff.

Anything with significantly lower capacitance has to be tubes.

Would that be 60 pf balanced or 15 pf (two caps in series)? I just calculated a circuit using MacSpice: with a 20K source resistance, 10K load resistance and either 30 or 60 pf across the load resistor. That approximates what I have hooked directly to the input of my KGSS (without the equalization cap). With 30 pf load it's down less than .01 dB at 20K. Even if you double the input resistor to 40K you're not going to see an order of magnitude change and even that would be less than 0.1 dB. I expected that or else our typical solid state audio gear would be pretty bad news. And my KGSS measured frequency response is not affected significantly by the position of the 1 dB/step attenuator. You guys are going to have to wait till I have the inclination or a free weekend to take my input circuit apart and just measure the frequency response hooking a low impedance generator directly to the input terminals. Something's very strange here, there's no legitimate reason for the rolloff that I measured in my amp and all theory says it is not due to the input attenuator. My colleague in Belgium will also make a measurement when he gets his KGSS, made exactly like mine except using DACT attenuators. Stand by for the fireworks ).

Edited January 31, 2011 by bobkatz

[Craig Sawyers]

My colleague in Belgium will also make a measurement when he gets his KGSS, made exactly like mine except using DACT attenuators. Stand by for the fireworks ).

I've just sprung for the last pair of Justin's KGSS boards and a 50k 4-gang DACT. It'll be a while before I get it finished though, although I've already got most of the bits - but not transformer or casework.