I'm in the process of designing a FCMW radar capable of doppler, ranging and SAR measurements. The IF signal will be recorded with a laptop and processed via MATLAB script.

The design is based on the MIT OCW course. The biggest concern I have is with the IF filter. I modeled the provided circuit in Multisim and, although it's listed as a low pass, it's really a band pass. The frequency response plots are below.

I realize that the purpose of this filter is to knock down the mixer sum and harmonic products but I'm curious as to why fc was chosen to be 15 kHz when the audio IC should be capable of processing signals of at least 20 kHz. I'm also not sure why a band-pass was implemented. It seems to me that by filtering out lower frequencies, the minimum range of the radar will be increased.

I'm thinking of using a true low-pass filter instead. For the doppler measurements the phase response of the filter should be as flat as possible across the pass-band, correct? Naturally, the desired amplitude response should be as flat as possible as well.

Also, in order to reduce side lobes, the transmitter should be as linear as possible. Amplitude nonlinearity in both the VCO and LNA are unavoidable. Also, nonlinearity in the tuning voltage is also unavoidable. The tuning nonlinearity will affect the frequency of the side lobes while the amplitude nonlinearity will affect the amplitude of the lobes. Would it be beneficial to implement a few microcontrollers (1 controlling the tuning voltage, the other controlling a voltage variable attenuator) to improve both tuning and amplitude linearity? With a tuning ramp period of 20 uS, is this a feasible solution?

 

how far is your frequency excursion? that and the ramp period will determine the band pass filter width.

 

So I know that IF(t) = cos[(2∏(oscillation freq)*(time delay) + (chirp rate)(time)(time delay)].

However, the time is unknown since it is based upon distance. I suppose the proper course of action is to determine the maximum distance from which a return is possible (based upon tx power and rx sensitivity), calculate the IF freq of that distance and design a filter that cuts off as close as possible to that frequency.

 

actually distance is determined by recieved frequency, the sweep is at a 20 micro sec rate, that limits the max range to about 2 miles. so the maximum difference between transmited frequency and recieved signal determines the frequency every 20 micro sec.and the filter size is determined by the frequency chirp rate of the transmitter.

 

actually distance is determined by recieved frequency, the sweep is at a 20 micro sec rate, that limits the max range to about 2 miles. so the maximum difference between transmited frequency and recieved signal determines the frequency every 20 micro sec.and the filter size is determined by the frequency chirp rate of the transmitter.

First: major oops for not noticing I titled it "FCMW" instead of "FMCW."


On to the discussion at hand, maybe I should rephrase. If the max bandwidth of the tx signal is 1.4 GHz and the max IF bandwidth (limited by the PC audio processor) is 20 kHz, wouldn't the lower IF bandwidth restrict the maximum measurable range of the radar since once the difference between the tx and rx signals exceeds 20 kHz it would be attenuated by the LPF?