[SOLVED] SA602A oscillator problems

[vladm]

Hello all;

I am building a direct-conversion receiver for 3-4 MHz, which uses the SA602A chip for frequency conversion down to audio.
The SA602A is a double-balanced mixer IC with inbuilt oscillator circuitry.

I followed advice from the datasheet:http://www.alldatasheet.com/datasheet-pdf/pdf/107777/PHILIPS/SA602A.html
and this page: **broken link removed**

and designed a voltage-tuned oscillator (see attachments).

To test the circuit, I connected pin 1 to ground via a 10k resistor and pin 2 to ground via a capacitor and connected pin 4 to an oscilloscope. The result was a sinewave of about 50mV p-p, variable between 2.8 and 3.8 MHz according to the tuning voltage (see attachments). Connecting the probe to pins 7 and 6 also had the same result, with a smaller amplitude.

To preform the next test i set up the circuit as shown in the attached diagram and inputted a 1MHz signal from a signal generator (1MHz is the highest the sig gen could go, so with the oscillator set to 3MHz the output should show frequencies of 2 and 4 MHz).

Now here is the problem: disconnecting the 10k resistor from pin 1 stops the oscillator. The output pin just outputs the input signal, no heterodyning products seen whatsoever on the 'scope FFT display. Connecting the probe to pins 6&7 now shows no signal.
This is very odd, I thought that the oscillator is isolated from any other pins. I need it to mix input signals with the local oscillator, which it is not doing at the moment.

The circuit is currently built on breadboard (see attachments), I have been told that it's generally a bad idea to use breadboard for prototyping RF circuits. (It is fairly easy for me to make a pcb.) The mixer circuit in question on the image is center, bottom is an RF filter and far right is an AF amplifier.

I know of issues concerning the poor dynamic range of the SA602A, so I had set the voltage of the sig gen to be well within the dynamic range of the chip.

Any help or suggestions appreciated.

 

[mvs sarma]

Please see think, where 7MHz Rx is managed.
https://www.google.co.in/imgres?img...91FUfzSHpHSrQfl_IHwDw&ved=0CEAQ9QEwBA&dur=607

 

[vfone]

The transistor inside of SA602 used as an oscillator is at pins 6 and 7. The transistor is internally biased and don't need external bias.
Seems that you got the rest of the chip oscillating, and when you disconnected the resistor from pin 1 actually you stop this oscillation.
Verify first if the oscillator works, supply the chip, and don't connect any other components other than those on pin 6 and 7. Use a coupled inductor to L1 to find and measure the oscillation using the scope.

 

[vladm]

The transistor inside of SA602 used as an oscillator is at pins 6 and 7. The transistor is internally biased and don't need external bias.
Seems that you got the rest of the chip oscillating, and when you disconnected the resistor from pin 1 actually you stop this oscillation.
Verify first if the oscillator works, supply the chip, and don't connect any other components other than those on pin 6 and 7. Use a coupled inductor to L1 to find and measure the oscillation using the scope.

Thank you for the advice, I did exactly what you specified.

I found that without the resistor to pin 2 (I had meant pin 2 instead of pin 1 in my original post) the circuit still oscillates, but the signal induced into the probe coil is ~5db lower than with the resistor.
The range of frequencies is also lower: with pin 2 grounded, varying the potentiometer generated frequencies in the range 2.7 to 3.8 MHz and without pin 2 grounded above 3.35MHz the oscillations stop.

To make sure it isn't the breadboard that is causing the problem (I have had problems with it before), I am making a pcb which should be ready tomorrow.

On a slightly different note, I had planned to use a KANK 5KR (the follow on from the KANK 3334 specified in **broken link removed** but I can't figure out which pins connect to the main coil, so I am using a 4.7uH inductor (it looks like a big resistor) at the moment which is 0.8uH lower than the value of the KANK 3334 (5.5 uH). Could this also be a source of problems?

 

[vfone]

No, the value of the inductor doesn't have anything with this issue. But layout, yes could be the problem. Also place a decoupling cap (100nF) between Vcc (pin 8) and GND (pin 3) of the chip.

 

[mvs sarma]

unshielded coils can have induction issues from local high power transmitters.
Thus, it is better to have the inductors, either shielded like KANK, or toroid models to be free from such induction.

 

[vladm]

I found the problem: It was capacitive coupling between adjacent tracks on the breadboard. I built it on a pbc and it now works fine. Also, trying to directly measure oscillations from the oscillator circuitary loads the circuit and changes it's behaviour.