excetara2 said:
Here is the circuit. Any help would be awesome.
You need a nonlinear mixer. Summing the two oscillators through resistors R5 and R7 makes a linear mixer, like you would use for audio. There are various kinds of nonlinear mixers. I played around with your circuit (the oscillators) and an LM1496 as a mixer. It worked well, except... let me get to that in a minute.
You said you wanted 100kHz. I think the base caps should be 10nF instead of 100nF. This will yield 100kHz.
Back to the mixer... The problem I had was that if I left one of the inductors at 500uH, and changed the other one so that it was less than 4uH different from that value, the two oscillators would injection lock, i.e., they would pull each other to the same frequency, with zero difference. This is a known "feature" of oscillators, and I have had it happen in hardware. In fact, I was surprised to see it in a sim, because I had assumed that the culprits were power supply impedance, ground plane impedance, capacitive coupling, and electromagnetic coupling. None of these exist in a simulation, unless you add parts to make them happen. Apparently all the coupling was through the mixer. An active mixer like the LM1496 should have less crosstalk than, say, a diode mixer. My point is, you are going to have to be very careful in layout, parts placement, power supply decoupling, and choice of mixer, in order to get true zero beats. You might want to build in a frequency offset of a few hundred Hertz to minimize this problem.