[Prev][Next][Index][Thread]

Tesla coil tuner (grid-dip meter)



I've found a grid dip meter is a very handy tool to have for checking
filters and other circuits. The grid dip meter (a simple transistor
oscillator with a sensitive current meter to measure RF absorption peaks) in
the ARRL Handbook also functions as a tuned RF field strength meter when
oscillator current is cut.

I've read regenerative recievers work on this principal, but suffer from
poor dynamic range, noise, and two-handed operation to balance the
regeneration which changes with tuning. But for cheap test equipment this
could serve very well.

Wouldn't it be nice to use a microcontroller (PIC) to control the tuning and
Q (damping) of a filter/oscillator? By pinging the circuit with a step, and
timing the ring period, the Q could possibly be adjusted for very high
(10,000?) values. It would be very nice for the following applications:

* Spectrum-analyzer type oscilloscope probe
* Clean RF sine oscillator, & sweep generator
* Antenna and return loss measurements (grid-dip meter)

The single-chip controller PIC might even count frequencies (under 50MHz
without extra prescaler) and send RS-232 data to a computer for detailed
analysis & display. I would hope the RF spectrum from 1MHz to 1GHz could be
covered with 7 transistor oscillator circuits, each covering a 3:1 frequency
band.

As a test I've tried tuning a 1KHz - 500KHz active-filter Biquad's (similar
to state-variable) integrator resistors with mosfets. Found I could get over
2 decades of tuning with the same cap, but the mosfet voltage-variable
resistors distorted (even with biasing resistors) and caused oscillations at
high frequencies. The gain of the loop affects the phase angle of the
integrators, so damping and tuning interact. 

This was from a test on a breadboard with a ground-plane. I don't know if
its the breadboard or the 4007 CMOS mosfets that are making this filter
oscillate when my spice model (which is using regular integrator resistors)
says it should be stable.

Anyone got a 4000-series CMOS library for Orcad (MicroSim) Pspice? I need
the 4066, 4069, 4049 & 4007 & CMOS mux device models bad.

I'm wondering if this is possible, if it has/is being done. Maybe an
impossibility because of complex and chaotic frequency-damping control loop,
if the relatively tame 2nd-order tuned circuit's response is demonicaly
posessed by evil amplifier and varactor reactance and non-linearity.

My guess is that I'll spiral in and converge on a desired frequency & Q,
unless I use a clever algorithm that tunes a 2-dimensional pole-zero space,
with compensation for varactor or Fet-VVR non-linearity.

Do you know of any radio's or test equipment commercial products that use
regenerative/super-regenerative filters? Suggestions?