Extensive work was done optimizing and testing the spectrometer in a number of different configurations and with three test samples with known NQR transitions. A brief collection of NQR data was initially done on p-dichlorobenzene. Then a study was carried out on sodium chlorate, during which most of the improvements for the superhet for us at VHF were made. The sodium chlorate signal was heavily optimized through incremental improvements to the design of spectrometer. A detailed report the experiments was completed in early 2015. Since then, the signal has been further improved by fine tuning the methods originally developed. In its current state, a signal average of N=500 results in an SNR between 50-100, a figure obtained in previous studies with N=10^5 or greater.
http://i.imgur.com/CPVOcsj.png FID of NACLO3
The hybrid tee isolation scheme was replaced with a quarterwave bridge circuit with crossed diode switches. D1 and D2 form switch Sin and D3 and D4 form switch Sout. When the pulse is on, Sin and Sout are both shorts. Sin passes the pulse to the probe matching network, and Sout shorts the right side of the quarter wave line to ground, forcing the anti-node of the EM mode to be at the left side of the quarterwave line. This guarantees maximum voltage is delivered to the probe and also that the RF amplifiers leading to the receiver are electrically isolated. When the pulse is off, Sin and Sout act like open circuits. The NQR signal originating in Lsample therefore is not divided as in the case of the hybrid tee, as no signal proceeds back to the transmitter. Sout is open so the signal see only Zin of the RF amplifiers, which are matched to 50 Ohms.
The probe was completely enclosed in an aluminum chassis and the coil was fixed solidly to a ground plane. The box was then clamped tightly to another ground plane to which other filters, amplifiers and external parts were also fixed in place. That surface was then connected firmly to a single strand of copper braid of thickness 1” leading to earth ground. Avoiding ground loops and potential mechanical oscillations proved improved the SNR. Additionally, shielding was added whereever possible, reducing the amount of electromagnetic interference.
The addition of 108 diodes (27 sets in series, each of 4 crossed diodes) to spectrometer between the transmitter and the probe eliminated unwanted noise and/or harmonics in the transmit pulses, and were experimentally confirmed to significantly improve S/N as more sets of diodes were added to the arrangement, until the resulting NQR signal no longer improved with the addition of more diodes.( Results show the spectrometer performs optimally for its typical resonance frequency of 29.9 MHz.
To test the spectrometer for lower frequencies, the 14N NQR in hexamethylenetetramine was observed with a hybrid tee bridge configuration. High S/N was achieved with robust isolation techniques, high quality fast recovery amplifiers, and high Q resonant tank. All studies were conducted using the same apparatus. The superheterodyne was largely unchanged for the three studies except for the tank section and the pulse protection circuitry between the power amplifier and the receiver section. An exact calculation of probe parameters was made. Evidence of an non-zero assymetry parameter in HMT was observed, but more research is needed.