In my recent project, the mini QRP SSB transceiver for 20 Meters (14 MHz) I encountered another unexpected problem: accurate sideband switching with the NE 612 as SSB generator can be difficult. Very difficult.
First my circuit looked like this as pointed out in the schematic:
Two crystals are used, one slightly higher than the passband of your ssb filter, one about 1.5 kcs lower.
When switching I found that the oscillator was not operating correctly and the frequency after the switch was put to the other position was some kcs off the crystal frequency. That time the switch was mounted in the front panel and the leads to the circuit were made of flexible cable with a length of about 8 cm. I speculated that this might come from unwanted capacities or inductances caused by the leads.
To avoid these unwanted influence I experimented with diode and transistor switches. No improvement could be found. Then I mounted a small relay instead the semiconductor switches close to the NE612. Again, no improvement. When analyzing the circuit I found out, that in case I remove the second crystal the remaining one operated perfectly. Oha!
So, I deduced that the internal oscillator was oscillating with two crystals in parallel. A lot of resultless experiments to have both crystals in the circuit but minimizing the influences followed. No cure was in sight.
Later I finally made the decision to use only one crystal. Due to the fact that I wanted to be able to operate on both sidesbands (don’t ask for a reason why! ;-)) I made up my mind to use only one crystal and to pull it to the desired frequency. So it ended up simply with one 9-MHz-crystal, a variable capacitor, a coil and the sideband switch:
Needless to say that this worked from the scratch. Coil data is simple: 12 turns on a TOKO coil carrier adjustable with ferrite core. The switch is again put into the front panel, no unwanted influences are observable.
Bye for now and 73!
Peter Rachow (DK7IH)