1. Hardware Design
The block diagram was designed to help us get a better understanding of all the portions that needed to be designed for this project. Given that we had the benefit of seeing what other groups had previously worked on, figuring this out was relatively simple even though it does not fully match our final design.
The bandpass filter was partially based on the work done previously by Caleb Froelich and Konrad McClure. We used the same tool that they used, rf-tools, to calculate the necessary values for a 3rd-order Butterworth bandpass filter for a lower cutoff frequency of 10 MHz and an upper cutoff frequency of 16 MHz. The number of turns for the toroids was determined to be 3nH per turns^2 for the Ts7-6 and 4nH per turns^2 for the T50-6. The 2.15V voltage bias part on the end is used to center the incoming frequency between the maximum and minimum Tayloe Mixer input values, which adds noise, but we ultimately determined that having that part is more useful here.
Calculations:
The instrumentation design is very similar to the Caleb and Konrad design. They designed it to work in a way such that the 0° and 180° signals, and the 90° and 270° are are differentially summed by the INA821ID Instrumentation Amplifiers.
The Tayloe mixer design is once again very similar. We decided that we did not need to re-determine how to set up the connections to get 0° on 00, 90° on 01, and so on.
Once again, this is something the previous group used to reduce some of the noise that resulted from the constant switching in the computer. Since for this project, some additional noise reduction seemed particularly useful, we pulled also used it for our design.
As the previous group used a Johnson counter, the local oscillator needed to be updated for this project, since Dr. Frohne said it was unnecessary. A 3.3V connection was added to the SCL_3V and SDA_3V connections on the pins 4 and 5, a crystal was chosen because it was the frequency needed for the Si5351 chip, and the previous group did not use many surface mount components, so this schematic does. We also arbitrarily set our I/Q signals to CLK1 and CLK2. In a future redesign, it might be slightly easier to set up the software if the clocks are set to 0 & 1 instead.
A voltage converter was added because with the different voltages needed for the oscillator and Arduino, we wanted to be sure nothing was burned up because an incorrect voltage was used. This design comes from the Adafruit Si5351 design. This was one of the places where we originally made an error in the schematic that made its way to the PCB.
Just like the bandpass filter, the lowpass filter was designed with assistance from online tools. This is one of the parts of our schematic that changed the most from the original design to the end result, but we ultimately decided on using a Sallen-Key Lowpass Filter that was calculated for a cutoff frequency of 100 kHz using this online tool. We also decided on using the LM4565 op amps for this (and the voltage smoother) because they were cheap and according to the datasheet, had a lower noise density than some of the other op amps we looked at.
Calculations:
