Specifications

using gold plated contacts. This wouldn't be too bad if
it were just gold. But the normal gold plating process
leaves pinholes that require a nickel underplate to avoid
corrosion of the copper base metal over time.
It is both expensive and dicult to perform "direct gold"
plating for connectors, and I have never seen it done
in a relay. But the normal technique of using a nickel
underplate leaves a magnetic layer underneath the
gold that the signal *must* pass through on its journey.
Magnetic conductors create signal distortion that is easily
measured with modern test equipment.
A relay also has no "self-cleaning" action such as is found
in a rotary switch. Each time the switch is operated, the
contact of a rotary switch wipes any surface oxides or
other contaminants away leaving a fresh contact surface
Finally the optimal conductor geometry is of no concern
in a relay. Sometimes the conductor is steel so that the
contact is operated directly by a magnet. Sometimes
the conductor has a complex shape that complex and
undesirable impedance variation versus frequency in the
switch. Almost all relays are sealed such that it is virtually
impossible to clean them.
Just listen
The net result from a sonic standpoint is dicult to
describe. I think the best description was given by our
Senior Research Assistant, Ryan Berry. At one point he
asked, "Why not just use a relay?" I told him to go listen
to the best one we had found, an exotic device costing
double the typical price of the a top-quality small-signal
relay. After a half-hour of listening tests, he came back
and I asked him what he thought. He replied, "They just
make the music sound, well, WRONG!"
After the Shallco rotary switches, the best sounding
switching elements we have found are FET switches.
These have many advantages, including relatively low
cost, silent operation, and no moving parts. However
there are many other disadvantages that must be
carefully managed if an acceptable level of performance
to be attained. They are also prone to damage from
static discharge, especially when used as input selector
switches that are connected to external devices, and
their "On" resistance varies with the applied voltage
However, when carefully managed, these are also
capable of truly excellent sound quality, and the latest
generation of devices are also quite robust.
We spent well over a month selecting the highest
performing FET switches for use in the AX-5 that would
provide trouble-free operation for decades in the harsh
environment encountered when exposed to potentially
damaging episodes ranging from nearby lightning
strikes to a high-voltage discharge when connecting
(or disconnecting) cables during the dry winter months
when static shocks are commonplace.
So far we have focused only on controls that are suitable
for use in stepped attenuators. This is because to obtain
the full performance available from a truly balanced
circuit, the level matching between phases of each
channel should be held to less that 0.1%. Devices such as
ordinary potentiometers (volume controls) fail to meet
this standard by a factor of between 200x and 10,000x,
rendering them totally unsuitable for use in a high
performance balance circuit.
A question of balance
As is normal for every Ayre product ever built, the AX-5
is fully-balanced from input to output, and uses zero-
feedback circuitry. The advantage of fully balanced
operation has nothing to do with interference pickup
" I am condent of this to the point that I consider any product that uses relays in the audio signal
path to be unusable as a reference-quality device. At Ayre we use them to short out the signal for he
mute function in some products. But that is as close as they ever come to the signal path of even the
lowest priced of Ayre equipment. "
t: 01727 865488 e: [e-mail] w: www.symmetry-systems.co.uk
Symmetry, Suite 5, 17 Holywell Hill, St Albans, Hertfordshire AL1 1DT
Page 4
The AX-5 Story
from Ayre Acoustics March 2014