78.5MHz AM receiver self-oscillation

Question

Circuit appears to be working but very easily breaks into oscillation - sometimes just getting up from my desk triggers it, or while tuning, and sometimes after a few seconds with no disturbance that i'm aware of.

circuit is built on a tin plate from a flattened 9V battery case... because i'm keeping leads as short as i can inductors are probably too near each other?

I'm not confident yet with s-paramenters so i left it out of my calculations...

EDIT: as requested, image of circuit and scope traces when transmitter is off or when not tuned in tuned in, 400 hz mod clearly visible :) but it's hard to get this, typically it is self oscillating.. and suddenly.... the frequency of the oscillation is totally dependent on the tuned circuit, it is fully tunable accross the full range of the trimmer 5 - 50 pf.

Answer 1

When designing circuits for the VHF region, careful design and construction practices must be utilized in order to obtain stable and repeatable gain.

Here are a few thoughts that may improve the reliability of your circuit:

1.) Increase the drain resistance. This will reduce the tendency for UHF oscillations (and reduce the gain).

2.) Use toroids for your inductors if possible. Toroid forms help to provide self shielding of the inductor. If you cannot do this, orient the axis of the inductor such that it points towards the amplification stage in order to minimize stray coupling.

3.) Keep all wire lengths as short as possible but in particular, minimize the gate lead length and its associated connections.

4.) You may get some improvement by installing a vertical shield over the FET to isolate the input from the output. Notch the shield to allow the shield to fit tightly over the FET.

5.) Consider using a common gate topology with sufficient drain resistance to combat UHF oscillations. This will have lower gain which can be made up with a second stage of RF amplification if needed.

Answer 2

I see two coils in close proximity in the "as built" photo. One coil is connected to the detector; the other to the antenna -- and their axes are colinear. Whether this was the intent or not, your receiver is acting as a regenerative type -- detected signal in one coil is coupling to the other. Without a control to limit current in the coil that's acting as a "tickler" or an adjustment to limit coupling (one classic method was to rotate the tickler -- one way it was maximum, rotated 180 degrees it was out of phase and minimum), the circuit can enter oscillation and there'll be no easy way to stop it.

Tiny disturbances, such as signal strength fluctuations, body capacitance, or stray inductance (from a wedding ring, for instance) can change the conditions enough to trigger this. Once it starts, it's generally self sustaining, in that the regenerated signal will be enough stronger than the un-augmented to keep oscillating.

Adding a variable resistance in series between the detector and its coil will allow you to reduce the coupling and control oscillation.

Answer 3

There is much about the topology of your first stage amplifier that looks like an oscillator. It's reasonable to try a design that is more likely to be stable: the common-gate amplifier. This topology sports low input impedance and high output impedance, which seem to be a good fit to the rest of your design:

Here is an LTSpice model of a possible circuit using the MPF102:

Note that the J310 or similar is often preferred for VHF applications because the reverse-transfer capacitance is about 10X smaller.

The 3-dB passband of the input filter is about 68MHz to 88MHz:

and the voltage gain is about 20dB:

Here's a plot of the frequency response:

Be sure to follow all of the recommendations in the earlier post, particularly shielding the input from the output as well as possible. Also, don't assume that your toroidal inductors are correct: the magnetic materials have large permeability variations, so measure them to ensure you get the expected results.