Intermodulation in a shared mast preamplifier
A contest station shares a mast with a strong FM broadcast transmitter 300 m away and a paging transmitter on 150 MHz. The station uses a wideband mast-head preamplifier with 20 dB gain and an input intercept point of -10 dBm, feeding 40 m of coax to a receiver with a 10 dB noise figure. During contests, operators hear a mix of broadcast programme material and paging data appearing on 2 m SSB frequencies. The problem disappears when the preamplifier is bypassed, but then weak-signal reception is poor.
1. Explain the mechanism producing the unwanted signals and why bypassing the preamp helps.
Read the worked answer
The preamplifier is being driven into its non-linear region by the strong out-of-band carriers. Two or more strong signals mix in the amplifier's non-linearity to produce intermodulation products at sum and difference frequencies, some of which fall inside 2 m. Because the broadcast and paging transmitters are very strong at the mast, even a preamp with a modest intercept point generates products that the receiver then amplifies. Broadcast programme material is audible because the modulation is transferred onto the product; paging data appears the same way.
Bypassing the preamp removes the non-linear stage closest to the antennas, so the products are no longer generated there. Reception of weak signals suffers because the receiver now sees the cable loss directly and its noise figure rises by that loss.
The real problem is not the preamp's gain but its linearity and the amount of out-of-band energy reaching it. The cure is to reduce that energy before amplification, not simply to remove the amplifier.
2. What practical changes would you make, and in what order?
Read the worked answer
- Add a band-pass filter ahead of the preamp, with low insertion loss in 2 m and high rejection at the FM broadcast band and at 150 MHz. This is the single most effective fix, because it removes the interfering carriers before the non-linear device.
- Replace the wideband preamp with a narrowband, high-intercept-point design, or reduce its gain. A lower-gain, more linear preamp often gives a better system noise figure than a high-gain one that is overloaded.
- Improve the receiver's own noise figure or move the preamp closer to the antenna to reduce cable loss after it, so less gain is needed.
- Consider a cavity filter or notch filters if the paging signal is very close to the band edge, accepting the extra loss.
- Verify with a spectrum analyser at the receiver input that the products have dropped.
The order matters: filtering first, then linearity, then gain redistribution. Buying a 'better' preamp without filtering usually just moves the problem.
3. How do you calculate whether the system noise figure is acceptable once the filter is fitted?
Read the worked answer
Use the Friis formula for cascaded stages: F_total = F1 + (F2-1)/G1 + (F3-1)/(G1G2) + ..., with all values as linear ratios. Convert dB to linear with F = 10^(NF/10) and G = 10^(gain/10).
Suppose the filter has 1 dB loss (F = 1.26, G = 0.79), the preamp has 15 dB gain and 1 dB noise figure (F = 1.26, G = 31.6), and the cable plus receiver has 3 dB loss and 10 dB noise figure (F = 10, G = 0.5). Then F_total = 1.26 + (1.26-1)/0.79 + (10-1)/(0.79 x 31.6) = 1.26 + 0.33 + 0.36 = 1.95, i.e. about 2.9 dB.
Without the preamp, the cable loss dominates and the noise figure would be roughly 3 + 10 = 13 dB. So the filtered preamp still gives a large improvement. The judgement is that a 1 dB filter loss is cheap compared with the intermodulation it prevents; if the filter loss were 3 dB, you would weigh it against a lower-gain preamp and possibly a mast-mounted filter with better selectivity.