I threw together this RF preamp for 88–108 MHz from junkbox parts. I had intended to use a low-noise GaAsFET, but I discovered that the ones I had were bad. I settled for a 3N204. (Dual-gate MOSFETs are so obsolete that I had to create my own schematic symbol.) Thinking I might mount it at the antenna some day to overcome feedline loss, I designed the preamp to be remotely powered. I found the Scientific Atlanta DA-PI power inserter at a garage sale.
The coil self-capacitance plus the 3N204 input capacitance resonate the input circuit in the FM band. The trimmer capacitor, which adjusts the input match, also affects resonance. I adjusted the trimmer and the coil shape until the gain peaked near the center of the band. For biasing I started with pots in the 3N204 source and G2 circuits, but I found that the arrangement shown worked best. It certainly minimized the parts count. R1 sets the drain current via the G2 voltage. I used 1kΩ with a 17.5-V supply, which yielded 13.5 mA at 4 V. Gain increased little at higher current.
The upper trace is the preamp response from 88–108 MHz. The lower trace is with a barrel connector substituted for the preamp and power inserter. The vertical scale is 2 dB/div. I tweaked the input trimmer for equal gain at the band edges. Midband gain is 10 dB, which is just what I wanted. This is sufficient to swamp a tuner's input mismatch loss while minimizing RF overload.
The noise figure of a 3N204 isn't nearly as low as that for a modern semiconductor. Still, the preamp improved the 50-dB quieting sensitivity of a factory-aligned Sony XDR-F1HD by 2.5 dB. I measured the preamp's third-order input intercept as 125 dBf. This implies that 87-dBf input signals will generate 11-dBf spurs, and 90-dBf signals 20-dBf spurs. Since this range covers the 50-dB quieting level for the preamp with any tuner, the RF intermod spec of the combination will be 87–90 dBf, or 10 dB less than the tuner's spec alone, whichever is lower. The latter figure usually dominates. Most of my tuners measure in the mid- to high-80s. At 99 dBf, my Technics ST-9030 could handle the preamp with little intermod degradation.
This shows how the preamp increases modulated S/N in a Sony XDR-F1HD. The horizontal scale is 2 kHz/div and vertical is 10 dB/div. I used a 13-dBf monophonic signal (approximately the tuner's 50-dB quieting level) 100%-modulated with 1 kHz. The upper trace is without the preamp. To compensate for the tuner's soft muting, I adjusted the traces for equal tone levels. Modulation greatly increases broadband noise for a signal this weak. Preamp gain counters the effect.
I built the preamp in a box that I salvaged from a defective RF slide switch (another garage sale find). 82Ω was as close as I could come to the 75Ω drain load. I adjusted the input trimmer to compensate for the shift in the response curve when I attached the press-fit box cover. Although the preamp works fine, the source, bypass, and resistor leads are longer than they should be.
I wrote this to show how easy it is to roll your own preamp. But don't duplicate this design. Instead, use a modern, low-noise GaAsFET or pHEMT and follow the manufacturer's datasheet.
88–108 MHz