Flexible Waveguide Project

From MesoscopicWiki

results in reverse chronological order, scroll down to see older stuff, including pdfs of papers

Download Jason's waveguide impedance calc. Bottom line:

v p = c

at 330 GHz

For these tests I carefully mounted the source and receiver on the bench so there would be no movement except for the linear travel of the receiver on the computer controlled stage. First up, here is a plot of a 1 mm scan (as from 9/17) but with no tubing at all. The distance between the ends of the adaptors (now acting as horns) is 70.8 mm. For the rest of today's measurements this will be the zero mm position.

Next, I did two frequency sweeps with no calibration and the adapters in precisely the zero position. One with no tubing present and one with tubing. They are plotted on the same figures below. Green is with tubing and blue is without. The end-to-end distance between the adapters should be identical for the two scans.

9/17/09 distance scans

For this test I put the receiver on a motorized stage. Both source and receiver are at the same level. NB, in this section there are no baseline or calibration corrections made. These are raw data which will include all the standing waves in the system.

First up, here is a fixed frequency scan (320.6 GHz) while translating the receiver by 1 mm. The coupler slides easily inside the waveguide. At this frequency lambda = .936 mm. There are 1285 points in the scan, so delta x is about .7 microns. The distance between the two deep minima in this spectrum is .47 mm. This is 1/2 lambda. The corresponding change in phase is approximately -180 degrees.

Next we'll look at frequency sweeps as we vary the adaptor-adaptor distance in discrete steps. The absolute distance is not critical here, but the steps between each spectrum is precisely 100 microns. The second figure is just a zoom of the first.

setup for fixed distance frequency scans

cutoff behavior

comparing coated and uncoated tubing of exactly the same length

In order to understand the effects of just the metalization, I used pieces of uncoated PC tubing of the same length as the coated (270 mm and 85 mm) to do a calibration sweep. So, what you're seeing in this figure are the amplitude and phase changes due to the presence of the metalization.

Here is the same experiment but done with a shorter piece of tubing. The tube itself was 80 mm. But what really matters is the coupler-coupler separation. I measure this from the "knee" of the coupler, where the taper changes to circular cross section. So, these are 85 mm knee to knee and the plot above is 270 mm knee to knee.

For the long piece there are about 16 peaks in 5 GHz, or 312 MHz spacing. This corresponds to a 3 ns period. But remember, these spectra are measured relative to the uncoated tubing. For the shorter piece of tubing it looks like the spacing is around 450 MHz. This corresponds to a shorter period, perhaps 2.2 ns.

first tests

RC to RC (Rectangular to Circular) no base. These are absolute powers in dBm.
taper to taper. used piece of PC tubing to align tapers. RC-RC used as base. These numbers are dB relative to that base.
propagation through a 255 mm piece of uncoated PC tubing using the taper to taper sweep as base.
propagation through the coated PC tubing using the taper to taper sweep as base.

if you couple the two rectangular-circular tapers directly together here is what you get in absolute power.