PA0LQ audio filter and mains QRM noise blanker
The following is a translation of a description I made for Electron of June 2000; the magazine of the Dutch amateur radio society VERON (to be translated as Society for experimental radio research in the Netherlands). Audio filter and noise blanker were designed and used by Harry Grimbergen, PA0LQ.
The upper part of the circuit diagram shows a non-ringing audio filter with 1000 Hz centre frequency and 30 Hz bandwidth.
The opamps are of the low noise type.
The values of resistors and capacitors may deviate up to 10% from the indicated values, as long as they are stable. For the capacitors polystyrene or mica types are to be preferred.
R1...R4 determine the resonance frequency of the four filter sections. Their reactance is about 180 ohm. To find the exact values they are temporarily replaced by variable resistors. The filter is fed from an audio generator of which the frequency is measured by a counter (its dial cannot be read with sufficient accuracy). The output voltage is indicated by an oscilloscope or a VTVM. The resonance frequencies are now adjusted as follows:
By R1 to 966 Hz
By R2 to 982 Hz
By R3 to 1018 Hz
By R4 to 1034 Hz.
When tuning is completed the value of the resistors is measured and the variable resistors replaced by metal film ones, or a combination of such resistors. Even the best variable resistors are not stable enough for this application.
Harry, PA0LQ, suffered from pulse interference that was related to the mains frequency, as caused by light dimmers with triacs or thyristors. He therefore made a noise blanker that is shown in the lower part of the circuit diagram. The system does not help against interference from fluorescent tubes, energy saving lamps and electric motors with a commutator, as used in electric shavers and other appliances.
The blanker functions as follows.
By means of two resistors and diodes half-sinewaves with frequencies of 50 or 100 Hz are derived from the secondary winding of T1. The two 50 Hz voltages that can be selected differ in time by half a period of the mains signal. Opamp TL071 converts the voltages into square waves that drive two monostables in cascade. Then follows a third monostable that determines the width of the "hole" that is made in the signal. Variable resistors "Pulse width", "Delay 1" and "Delay 2" permit to position the blanking pulse exactly on the interference pulse. A LED indicates when the blanking pulse is present. The blanking pulse drives the gate of TS1, used as a switch.
The selectivity of the 30 Hz filter causes the holes to be refilled almost completely. Readability of CW is affected only when more than 20% is cut by the blanker.
Also for LORAN?
John Rabson, G3PAI, asked me whether the PA0LQ noise blanker could also be used on LORAN signals. PA0LQ thinks this is possible but the blanker must then be driven by LORAN pulses that come from a separate receiver, tuned to 100 kHz. I think that this could be a rather simple receiver, for instance of the TRF type. It should not be too selective in order to preserve the shape of the LORAN pulses.
Dick Rollema, PA0SE