Image Rejection? Non-existent? Why?

[greenjacketgirl]

Background

I am now using a soundcard that has stereo (hurray). So this should not be the reason that image rejection is not apparent.

Description of Issue

The issue remains of image rejection not being present. Looking into this issue, I used the analog discovery II oscilloscope to look at the I and Q signals at each stage of the op-amps. What I found was that after the summing amps, they were 90 degrees out of phase and the same amplitude, and after the low pass filter, they were still 90 degrees out of phase. But as soon as the signal passes through the 10uF to remove any DC offset before entering the audio jack, they become in-phase --> very bad. essentially all our hard work at producing two signals 90 degrees out of phase is wasted in one issue.

[greenjacketgirl]

Solution Part 1

After analyzing the audio jack T,R and S sections for appropriate connections. I discovered (along with the help of my dad and Dr. Frohne) that the ring was not making any connection. Thus, in order to by pass this problem, I took some insulated wire, stripped the ends, and soldered them onto alligator clips on one end, and the audio jack pins on the other. This way, I no longer had in-phase signals, but rather the 90 degrees out of phase signals I desired.

[greenjacketgirl]

Solution Part 2

While the "solution part 1 outlines the first part of the problem, this part was also crucial to getting any sort of image rejection. Since the original soundcard only had 1 channel, I needed to find another soundcard with stereo input. My dad, being a packrat himself, happened to have a stereo soundcard: ONKYO USB Digital Audio Processor MSE-U33HB

This has two options for input: 1) mic input 2) Left and Right input. I started with using the mic input, assuming that it did two channels. It appeared that it did, according to audacity (as shown in the image below)

Evidently, by trial and error we discovered that the mic input was mono and therefore I needed to use the Left and Right inputs on the soundcard. This finally solved the issue! The result was a little of 30dB of image rejection. Shown in this image:

We can further isolate our signal of interest by using quisk's RX phase adjustment tool. Settings shown in the following screenshots, followed by the final output signal where we can hardly see the image at all.