ARC Ref3 or McIntosh C2300/2500

Old thread. Good discussion but without a proper technical answer. Clearly ARC is the winner in popularity for resolution/pure performance.

However, the REF 3 requires a high input impedance power amp to sound best (ARC power amps are typically 200K). Matching to a 10K input impedance power amp will have a slightly rolled off treble and higher distortion. (ARC 3 rises to 1400 ohm output impedance at 20KHz and distortion rises dramatically into lower impedance amps)

The Mcintosh has much lower output impedance and will work with lower impedance power amplifiers (as low as 5 K).

The ARC design is indeed outstanding and close to the best SS in performance with the advantage of a hint of tube 2nd harmonic warmth. However, a high input impedance power amp will be more susceptible to ground loops/hum. The Mcintosh is not as bleeding edge in terms of performance and appears to lean more towards old school “tubey” sound (warm and rich). The Mcintosh can be expected to be more robust and workable with a wide variety of power amps.

Mcintosh is like Mercedes (built for comfort & ease of use). ARC is more like Ferrari (built for speed but not as easy to use). Frankly I would avoid 200K input impedance power amps like the plague - with this design you are almost guaranteed to have RF/hum/ground loop issues from very tiny stray current which may show up as hiss - nevertheless a Ferrari is a Ferrari and when everything is going well for Ferrari, Lewis Hamilton in his Mercedes is looking at the back of a Ferrari  - like at Bahrain.

@kalali

No material difference between 100K or 1000K input resistance. Both are high enough such that any stray micro-amp current could generate a voltage input to the amplifier.

You need to get to 10K or below input impedance to start making a difference. This means the preamp needs to have a stronger drive output - all good stuff as anything that increases the need for a stronger line level signal means less noise as the end result.

Very old gear used 600 Ohm as standard and was immune to the kinds of noise now affecting high impedance gear but that standard actually came from transmission line theory for telegraph wires so it had other issues.

Another solution to noise on input is to go balanced (this means each signal wire to ground should have the same impedance) - this ensures that RF/EM hum/ground loops are affecting both wires equally and cancel out. However even balanced works better 10K and below for noise suppression.

So as a general rule, ultra high input impedance is a bad design if you care about noise. If you wanted to deliberately design an amplifier to detect very low level signal noise then I can guarantee you it would have a very high impedance just like an oscilloscope input.....