joe, steve and comrades, i have thrown together yet another mixer after reading some more literature from w7zoi, joe's experiments and the second kiss mixer as well. so, here is what i did come up with. refer to the double_kiss.png attached. it simply a transformer with a DPDT switch on both sides of the primary. now, alternatively, one side connects to the input while the other side is grounded. so, all in all, the transformer's polarity is rapidly reversed. the advantage this has over the regular KISS mixer is that (a) it uses FST3253, hence, matching, biasing etc are already taken care of, (b) unlike the original KISS mixers, there is no 'unused' section of the transformer, hence, the LO isolation is very good. how is this different from Joe's mixer? this one uses a single transformer. it uses the FST3253 which can be hand soldered without resorting to SMD PCB. (see the picture). implementation details : there are a couple of resistors thrown in the provide biasing. they make the the circuit look more complicated that it really is. the output transformer is configured as a current balun rather than a voltage balun. performance: the performance is exceptiona. the losses are so low that the LPF, mixer and the crystal filter together are together less than -10db loss. The local oscillator leakage at the RF port is less than 5mv (except at 7 MHz where it is 10 mv). i couldn't measure the IIP3 properly as my dual generator isn't powerful enough. I suspect it is more than 20 dbm even with the crystal filter (reactive termination) at the IF port. i will need a stronger dual signal generator to accurately measure it (the IMD products are going into the noise grass). - f
Attachment:
double_kiss.png
Description: PNG image
Attachment:
double_kiss_circuit.png
Description: PNG image
Attachment:
double_kiss_mixer.jpg
Description: JPEG image