There are situations where prototyping is the most efficient design
procedure. This is a model of my antenna-coupled tunnel junction
devices that I cobbled up early in the project (about the end of 2001)
to see if it could ever really work.

This setup consists of two pieces of coax with bazooka baluns, driving a
Yagi-ish antenna at both ends. One of the directors has a cut in the
middle, with a bit of wire soldered so as to overhang the cut. Pushing
on it with a Q-tip shorts out the cut, like a telegraph key.

The first plot is the end-to end-insertion loss of 18 inches or so of
1x1 inch alumina waveguide, normalized by soldering the coax feeds
together and doing a trace subtraction. The second plot is the
open/shut ratio.

This took less than a day with diagonal cutters and a spectrum analyzer,
and I've been trying to come anywhere remotely close to this with
clusterized simulators and IC fabrication for six years now. What's
that, some depressing number like 1000:1.

The problem is that there isn't anything remotely as good in the
infrared as copper at low frequency--normal conductors have a dielectric
constant of about epsilon_0 + i sigma/omega, where sigma is the
conductivity and omega is 2*pi*f, as usual. Changing omega by a factor
of a million really reduces the effective conductivity!

So score one for copper tape and diagonal cutters.

Cheers,

Phil Hobbs

Phil Hobbs wrote:
There are situations where prototyping is the most efficient design
procedure. This is a model of my antenna-coupled tunnel junction
devices that I cobbled up early in the project (about the end of 2001)
to see if it could ever really work.

This setup consists of two pieces of coax with bazooka baluns, driving a
Yagi-ish antenna at both ends. One of the directors has a cut in the
middle, with a bit of wire soldered so as to overhang the cut. Pushing
on it with a Q-tip shorts out the cut, like a telegraph key.

The first plot is the end-to end-insertion loss of 18 inches or so of
1x1 inch alumina waveguide, normalized by soldering the coax feeds
together and doing a trace subtraction. The second plot is the
open/shut ratio.

This took less than a day with diagonal cutters and a spectrum analyzer,
and I've been trying to come anywhere remotely close to this with
clusterized simulators and IC fabrication for six years now. What's
that, some depressing number like 1000:1.

The problem is that there isn't anything remotely as good in the
infrared as copper at low frequency--normal conductors have a dielectric
constant of about epsilon_0 + i sigma/omega, where sigma is the
conductivity and omega is 2*pi*f, as usual. Changing omega by a factor
of a million really reduces the effective conductivity!

So score one for copper tape and diagonal cutters.

Cheers,

Phil Hobbs

------------------------------------------------------------------------


------------------------------------------------------------------------


------------------------------------------------------------------------





Nice work and a credit to prototypes everywhere! I too a a believer of
building..so many rely on simulation solely... They must have bad Mitts
and lazy PINS

Marc