The datasheet for the MGS-71018 is very concerned with the relative DC levels of inputs and control lines. Is there any DC on the RF pins 2, 5 & 8? If so, that might explain why the switch isn't working to spec.
I think this is what is happening - While there is no DC on the RF input pins (5 & 8 in this configuration) there is once you drive the control pins into +ve voltages. I think this must have been different on the 1GG7-4218 switch that the schematic originally refers to. I believe that this is impacting the mixer bias - You can see this in the Part 6 video I just posted.
TonyG
Tony:
I watched part 5 of the video, and looked at the available schematic and the MMIC switch specification (why does it not list the truth table for the configuration we're using?
). I am not really sure what's going on, but I figured the overall circuit (up to the point things go into the mixer) is probably simple enough to simulate in the famous in-browser simulator.
In any case, I
had a go at it (warning: the URL link is very long so make sure you use a browser which can handle this). I think this may be the first time I'm
more confused after simulating something than before (mostly I'm confused as to why in the hell they used this +10/-10 configuration to drive this thing).
For the part of the circuit that we have a schematic for and measurements for, I put them in exactly (the +8.93 and -9.36 volt rails, the resistors, the capacitors). Obviously while there is a simplified schematic for the RF absorptive switch, I don't know any of the values of anything inside there (not that this simulator is really good enough for that anyway). One thing I do know, is that whatever the values are, they should present as a 50 ohm termination to the unused input -- a feat that I didn't really attempt to accomplish (though I included switches to try and look at this).
I also included switches to ground out the control voltages.
Someone earlier asked about "specific" values of resistors. It looks like the resistors with values
21.5k and
6.2k are chosen to provide a a 2.5V CMOS crossover point. I really don't understand why 34.8 ohm resistors are used on the control lines, but it is worth nothing that 34.8 ohm resistors (along with 1n caps) are used on
all control lines for everything: every single attenuator switch control, the switch cal control, etc.
I assume you've already noticed everything I said, but since I made the simulation I figured I'd share it.