Broadband loop amplifier (update Dec. 2018 of 2003 design)
The broadband loop amplifier design requirements
Of course the result will be a compromise, on all aspects good enough.
(see for the principle of the amplification: Broadband amplification)
Note: balancing and visibility both are very relevant on a residential location. Especially balancing is most relevant because of the high man-made noise levels on residential locations.
Practical implementation of the active broadband loop amplifier
This is an update of the 2003 design (2003 version): mainly improved the DC biasing.
The demo version of Simetrix is preferred for design and simulation instead of PSPICE. It is much more intuitive and aimed at the design process.
This trans-impedance amplifier is not very complex. Important is using a (V)HF construction method with short interconnects and VHF decoupling for the HF part of the circuit (Q1, Q2, Q22 and Q3). Especially the emitter connection of Q1 needs to be as short as possible.
The amplifier consists of three stages. The transistor Q1 is optimized for noise in a Common Emitter (CE) circuit. The output stage Q3 has to deliver the power to the load. The intermediate stage Q2 and Q22 provides extra amplification to get the desired large signal behavior.
The gain (antenna factor) is fixed with Rgain and Csgain. Csgain is also used to stabilize the amplifier.
The opamp X1 controls the DC bias levels. The response is fast enough to recover from an overload when doing QSK. Remember that your own transmit signal easily overloads the amplifier. Worst case the output of the opamp has to be 1Volt from rail capable.
The diodes D1, D2 and D3 protect the amplifiers input stage (also from your own transmit signal).
The VHF-notch/low pass filter attenuates strong local VHF transmitters.
The transformer together with the Faraday shield take care of the balancing and the common mode suppression. The coupling factor k of the transformer is about 0.8. Inductances are 8uH.
The design assumes a 1.3m*1.3m loop with 2-3mm diameter wire. The impedance of the loop is relevant for the large signal behavior via the amplifiers loop gain.
With a 1.3m*1.3m loop with 2mm diameter wire the antenna factor (E/Vo) is about 0.4.
The transformers coupling factor k=0.8 introduces an extra inductance in the loop. The loop is not perfectly shorted by the virtual ground. The inductance lowers the 60MHz resonance frequency of the loop behaving as a quad antenna to about 40MHz (see: Broadband amplification). It gives extra gain at 40MHz and at the same time a lower noise contribution of the amplifier. Any extra parasitic inductance or capacitance will further lower this resonance. Resonances affect the matching between loops and so parasitics have to be minimized.
In the next plot the transmission line circuit model of the loop is used.
Output voltage Vo in dB´s with k=0.8 for E =1V/m.
Latest measurement of the noise contribution December 8 2018 on a quiet remote location (rural location: 51.360271, 6.220209) and at my home location (loop at 1mtr above ground):
The plot shows the output noise increase when the loop is connected. Note: 3dB increase indicates that the noise of the amplifier equals the ambient noise!
The measurement at home is done at the quietest time of the day without any identifiable noise source like the plasma television set of the neighbors. In the evening noise level increases with up to 12dB.
The measurements confirm the lowered resonance frequency from the loop behaving as a quad antenna. The 40MHz resonance gain peak is >10dB higher. At this resonance the noise contribution of the amplifier is lowered, so the noise measured is ambient noise. EZNEC simulation and the transmission line circuit model both confirm this 40MHz resonance.
Note: It is a valid measurement method, because the noise contribution of the amplifier to the output noise with and without the loop is equal within 3dB. The output noise without loop is 3dB higher at the 40MHz resonance peak. So the 40MHz peak is actually 12-13dB higher.
Large signal behavior
The following plot shows the measured results in the (version 2003) simulation (PSPICE).
The large signal behavior measured in simulation (PSPICE).
Real measurements on the updated amplifier indicate a better large signal behavior. E.g. the second harmonic at 14MHz of a 7MHz S9+66dB signal is clearly lower than 39dB below S9. That is more than 105dB below the 7MHz amplitude. On 14MHz it can just be heard deep in the noise without the loop connected and so with only the noise of the amplifier.
The Faraday shielding (including air slit!)
This construction improves balancing with 6dB. See also the picture of the amplifier:
Bias-T power supply (needs low series resistance 100uH inductors!)
Two updated amplifiers
No PCB layout or SMD devices used, but the BFR93A:
Two orthogonal loops
Discussion on the main design compromises
Using the common emitter (CE) circuit for Q1 for best noise behavior with the emitter directly to ground for minimal inductance, increases the complexity of the DC biasing.
The chosen transformer coupling is the easiest and most controlled way of maximizing the balancing. However the coupling factor k is 0.8. This attenuates the signal amplitude 3-5dB and so increases the noise contribution of the amplifier with 3-5dB. It also affects the flatness of the antenna factor, but at the same time it enables a better noise behavior at around 40MHz. Employing the resonance however makes it more difficult to match multiple loops for measurement purposes or for using them in an array. The turn ratio 1:1 is also a compromise. Optimal noise impedance matching is not possible over the full bandwidth.
Only the BFR93A is SMT
Limiting the diameter of the loop wire to 2-3mm increases the noise contribution of the amplifier with <3dB.
Last update: March 15, 2019