[linrad] OT: Doppler module frequency acquisition

["G0RUZ Conrad" <conrad@xxxxxxxxx>]

I have a design project at University which is basically to make a handheld
X band Doppler 'speed gun'. The basic requirement is to build something that
will acquire and display speed in mph or kph with a range of 10-150 mph, an
accuracy of +/- 1mph and an acquisition time of 0.5 seconds. It must be able
to detect a moving vehicle at a distance of > 20m.

I am to be supplied with an X band Doppler module which has a low noise
amplifier fitted. The module has a WG16 flange for the connection of a horn
which I also have to fabricate. I have calculated that the mixer diode
output will be between 300-5000 Hz for the speed range mentioned. I will be
using a PIC due to time and budget constraints (with more time/money I would
have used DSP) and I can think of a few ways of getting a signal into the
PIC.

Traditionally students have used a few stages of amplification, followed by
a Sallen Key filter and either a comparator or a diode clipper then used RB0
on the PIC to generate interrupts at zero crossing points which are counted
subjected to some maths functions and displayed on an LCD display.

That is ok if you have a nice clean signal with good amplitude but I have
seen a few previous attempts and they all suffer from excessive gain with no
agc, a huge horn and are easily confused if another target comes into the
antenna aperture.

I thought that a better approach may be to use two PLLs (or one PLL with
switched filtering), one with a wide lock and capture range to acquire the
target quickly and the second with a narrow capture range to retain lock and
ignore other signals. I can either measure the VCO frequency (by counting)
when it is locked or measure the control voltage if it is linear enough
across the range of input frequencies. I will probably use a PIC16F877A as
it has a UART to drive my display (asynchronous serial VFD), 10 bit ADC's
and a PWM output which could be used for the PLL control if 2 PLLs were
used. It is also nice to be able to indicate that a target has been acquired
and that lock is maintained.

I have had very little experience with PLLs so here are the questions :-)

1. Does this approach seem reasonable?
2. Can I get PLLs that give good sensitivity when AC coupled?
3. Is it possible to get adequate linearity across this range of input
frequencies to use the ADC to measure the VCO control loop voltage. Of
course I could use a look up table but I'd rather not!
4. Would something like an HEF4046 be good enough or is there a better
device?

I'd be really interested to find someone with experience with PLLs and PIC
based frequency acquisition to point me in the right direction. Don't worry
I want to do the work for myself but a few pointers wouldn't go amiss.

Regards

Conrad G0RUZ


LINRADDARNIL