Use of quartz tuning-fork resonators to provide a sensitivity peak well above audio frequencies on narrow-band opto-diode nanowave receivers.

Here's a mail to the UK nanowave group that usefully introduces some thought on the subject:

'Another direction that the light box has allowed easy evaluation of is the use of a quartz resonator to null out the opto-diode self capacitance. Whilst only appropriate for very narrow band modes (well, unkeyed carrier operation really) it could be useful for evaluating cloud reflection paths, for example, and allows you to move away from all that 50/60 Hz optical crud, not forgetting the many  harmonics of this type of QRM.

The problem so far had been that crystals considered were of the AT type cut and these have a much lower parallel resistance at resonance than the 40 M ohm or so dark resistance at the usual opto-diode/FET gate interconnection. A 32.768 kHz quartz tuning fork resonator data sheet, though, looked more promising. Even the equivalent series resonant resistance was quoted as being about 50k ohm, so on the basis that the parallel resonant resistance is potentially 50k times the square root of the quartz Q, that gives 5 M ohm, which still falls short, but is a lot closer...

Simply shunting the opto-diode/FET gate junction of a Finningley receiver with a 32.768 kHz crystal from the junk box down to ground produced a good peak with a bandwidth of about 0.5 Hz. In comparative tests with a 1 kHz tone, s/n was down by 10 to 15 dB, depending on the quartz sample. So this looked promising.
A 75 kHz resonator performed about the same, as did some 100 kHz units that Farnell currently have in stock. It seems easy at this location (Cambridge/Newmarket area) to ensure that there is no direct electrical signals at these frequencies leaking into the front-end, so the SpectraVue display that was being used to measure the sensitivity remained completely un-cluttered, with just the test signal visible (OK, there's still some high frequency optical QRM from SMPU fed lighting, but the wanted signal is still much more obvious than when it is down at baseband frequencies).

The other potential feature that a quartz resonator can provide is a degree of impedance matching. So putting the live end of the opto-diode to a series connected resonator and feeding the FET gate to the other side of the resonator, produces a step-down of the diode impedance. Shunting to ground this second side of the opto-diode with capacitance allows a degree of control of the step-down ratio. With the resonator samples that I have and a gate to ground shunt resistance of 220k, the best s/n was achieved with between 33 and 68pF. You could also easily see a change of loaded resonator Q as you did this (ie, noticeably lower Q with no shunt capacitor).

With this arrangement, there is one obvious advantage (the lower amplifier input impedance is more sociable) and one disadvantage (you can't apply opto-diode reverse bias without using a sensitivity reducing resistance connected to the high impedance point, as far as I can see).

After all this, does it get us anywhere?. I'm not sure - but it's not far off, and the question should really be 'is there another form of quartz resonator that has an even higher parallel equivalent resistance?". Again, I don't know the answer to that. Watching the SpectraVue display slowly decay when you remove the light source is fun. Likewise, the slow build-up as you re-apply the light source.'

There was a zero response to this mailing! - perhaps because very few people want a receiver that is only suitable for CW reception, and CW with no keying applied at that, given the extremely low sub-Hz bandwidth. I just thought that it might be fun to try......

Initial evaluation

This couldn't be simpler:         

simple fe

A measurement of tuning-fork parallel capacitance on one sample produced a value of 0.5 pF, which is about one twentieth the self capacitance of an un-biased SFH213 opto-diode, implying that the addition of a shunt tuning-fork resonator will not degrade the opto receivers sensitivity at other frequencies in any meaningful way..

Going through a batch of various brand resonators from the junk box and applying them in the circuit above, resulted in receiver sensitivities at 32.768 kHz that were within 10 dB of that at 1 kHz. Two samples were within 3 dB.