Quote:
You might have been better with two 10dB attenuators ahead of your combiner because this gives an extra 20dB drop in the ,egel of one generatr getting to the other one. Any close-by signal getting in backwards to the levelling-loop detector guarantees intermod creation.
If the combiner is very sensitive to its output load then yoou maybe do need a pad after it. Perhaps split the difference and have 5 in each port?
In serious cases, I've used minicircuits power amps and then arge attenuation after each sig gen into a purely resistive combiner. I was developing mixers in the >+45dBm TOI area.
|
If I understand you correctly, that was what exactly I did. I had two 20db "isolators" meaning attenuators in front of the combiner. Please see my sketch of the set-up in post 81
Quote:
used test frequencies of 99.00MHz and 99.05MHz.
The GALI-51 has about 18.1dB gain at 100MHz so to get two output tones at +3dBm each I would have been driving the GALI-51 with two tones each at 3 - 18.1 = -15.1dBm at the input of the GALI-51 eval board.
My combiner has about 6dB loss and there's some more loss for the LPF. I also had a 10dB attenuator after the combiner. Normally I can get in the ballpark of -100dB IMD3 terms after the combiner with this setup at VHF.
My sig gens have a narrow ALC option so when I select this there is very little crosstalk between the sig gens at a 50kHz spacing. I'm not sure what the ALC bandwidth is in the narrow mode but I think it is much less than 1kHz.
|
Thanks. I read that sometimes they use ferrite beads to wrap around the signal paths internally to enhance isolation and avoid crosstalk between the two channels.
I am moving on to the IF amp using ceramic filters in cascade using Nelson Jones' WW 1971 design (see attached schematic). I have used two 280kHz ceramic filters using the "cute" CA3053 IC in cascode configuration. If you compare the response curve of a single ceramic filter in post 85, the roll-off is much steeper. The overall gain is only is 9.3db due to the slight impedance mismatch of LC matching pads. Measurement has been conducted with a 15db attenuator at the input and a 10db attenuator at the output of the 50 to 330 ohms LC matching pads. Previously I managed to get the LC pads perfectly spot on at the centre of the Smith Chart(see post 85). This time I tried to adjust L-pads with the trimmers but I could not make it any better. I am going to build the TAA6611B IF Amp and detector stage....