Measurements

Conditions	Channel	Signal dBm for '5' on the meter	Signal dBm for 0dB SNR	Comment
Before changes	Ch 2		-123.5	Coils had been previously tuned to channel 2
Moved mixer drain to centre-tap of first IF	Ch 2		-122.2	Gain loss not measured, 1.3 dB loss in SNR
As above, but the next day	Ch 2		-122.8	Shows relative accuracy achieved in these tests.
Restore mixer drain, but reduce RF gain by a 56K in parallel with R41	Ch 2		-123.2
Tuned L1 and L3 to Ch 2	Ch 2		-122.8	Again, shows comparative accuracy - shows the fractions of a dB are only estimates.
Tuned L2 and L4 to Ch 1	Ch 2		-123.2	Possible interaction of the middle section filter
	Ch 3		-123.2	The next three measurements ...
	Ch 4		-123.2	... show variation of SNR across the band, ...
	Ch 5		-122.9	... but tell us nothing about the gain.
Carefully retuned L1..L4 to Ch 1	Ch 1		-123.5
	Ch 2		-122.8
	Ch 3		-123.5
	Ch 4		-123.5
	Ch 5		-123
Final tweak of tuning, make all circuits resonate just below channel 1.	Ch 1	-107	-123.6	With old filter: -105/-121.0
	Ch 2	-106.2	-122.7	-104/-120.2
	Ch 3	-105.8	-123	-103.6/-120.5
	Ch 4	-104.5	-123	-102.5/-120.6
	Ch 5	-102	-122.5	-99/-119.5
Restore full RF gain to check	Ch 3	-106.5		with old filter -104.5, conclude that 56K had reduced gain by about 1dB.
Try 27K in parallel with R41	Ch 1	-106	-123.2	SNR loss is acceptable, but would like to get more reduction yet in RF gain to approach the reduction when the old filter was in use.
	Ch 2	-105	-122.5
	Ch 3	-104	-123
	Ch 4	-103.2	-122.8
	Ch 5	-100.8	-122
Try 22K in parallel with R41	Ch 1	-105	-123	SNR loss 0.5dB, gain loss ~2.5dB
	Ch 2	-104	-122.3
	Ch 3	-103.3	-122.9	SNR loss 0.6dB, gain loss ~3.2dB
	Ch 4	-102.7	-122.5
	Ch 5	-99.7	-121.8	SNR loss 1.2dB, gain loss ~6.8 dB

Test conditions

Generator: Marconi TF2015, uncalibrated but agreed with spectrum analyser
Signal: 5KHz deviation frequency modulated
Noise meter: specially written program, filtering generator modulation frequency
Channels: Ch1 = 137.30, Ch2 = 137.40, Ch3 = 137.50, Ch4 = 137.62, Ch5 = 137.85 MHz