A quick performance test of the low cost 10M4D filters obtained from Germany, note loss indicated on plots includes that of resistive matching & ultimate attenuation isn't displayed due to reaching the analyser noise floor.
The indication is that the filter easily meets its published specification, the plots haven't been "frigged" in any way and were made with the interface shown. IMD tests haven't been made.
Centre frequency 10.695MHz
|
MODEL |
NO |
Pass |
Stop |
Ripple
|
Insertion |
Guaranteed |
Terminating |
|
|
10M4D |
8 |
6 |
80 |
2.0 |
4.0 |
90 |
0.9//10 |
The method I have used to "match" the filter into Pic-A-Star IF is shown above.
L1 is 22 turns 0.25 mm wire wound onto a T50-2 (red) core, T1 is 36 turns of 0.25 mm wire wound onto a T37-2 (red) core tapped at 29 turns from the "cold" end. Put simply there are 29 turns across R5 with another 7 turns connected to the filter via a DC blocking capacitor. T1 is resonated with a 60pF trimmer, note that as I couldn't obtain 100pF trimmers VC3 is 60pF with 33pF in parallel.
//pf)
2.0