Effects of Up/Down/Transverters on receiver front end?

[rwgast_lowlevellogicdesin]

So ive been looking at building a down converter for SDR use, and I stumbled acoss one that never got produced but was open source and the mixer it was using is an LT5551. This device has some incredibly good specs especially when compared to a standard mini circuits ADE-1. I know that the first amp in the rf chain will determine overall noise floor etc.. The LTE5551 is great but pricey especially when you toss in an AD chip to generate the LO. What im asking is if you want to perverse your receivers dynamic range and ip3 specs should you choose a mixer like the LT5551? When using something like the ham it up which uses an ADE1 mixer does that mean in basically trashing my airspys numbers? I know the ade-1 and lt5551 work in totally different frequency ranges it was just easy to use as examples.

[David Hess]

I do not deal with this very often but I think you basically covered it.  RF gain to make up for mixer loss or noise figure subtracts from the IP3 so you want use just enough RF gain to make up for the mixer's noise figure or loss.  In the HF bands, this is irrelevant because the noise figure is so high anyway making an RF amplifier before the mixer a waste unless it is for blocking leakage but at VHF and higher, this is not the case.

The ADE1 is a pretty mediocre mixer though at least as far as IP3.  A high IP3 passive mixer would provide better performance than the LT5551 but also require 20dB more (100mW) local oscillator signal so the LT5551 definitely has its place and it is easier to use.  There are some alternative passive and active mixer designs which are even better than this but the only place I have seen them used is in the highest performance HF receivers and custom gear.

[PA0PBZ]

One more thing: Mostly down converters are using a fixed frequency, so don't use an AD chip to get the LO, an old fashioned crystal oscillator can do much better in terms of spurs and (sideband) noise.

[dmills]

State of the art in HF band mixers are probably the 'H mode' switching designs, google PA3AKE for more then you ever wanted to know about really high performance HF RX design.

As the mixers get up there in performance you will increasingly find that LO phase noise becomes a problem, especially if you are doing high side LO and trying to put the IF at VHF, there might be something to be said for a good crystal osc if you can make the VHF IF tunable and run with a fixed LO.

Osc PN is insidious as it just raises the noise floor and looks like band noise.

Whatever you do, a diplexer after the mixer and before the filtering is a good idea as mixers are generally termination sensitive.

Regards, Dan.

[rwgast_lowlevellogicdesin]

Well I really don't wanna mess with an ad rf synth chip or programming it, I would rather have a fixed LO of 2ghz. Im not to sure how to do this besides frequency multiplication of crystal/tcxo. Iv heard about reflection oscillators but I dont seem to be getting any good info on google.

As far as mixers go, I know all about an H-mode , I actually have the fast bus switches right here but that is only applicable for HF. What are some good passive mixers that would work well for microwave down conversion, preferably not qfn packages. You could boost the LO and take care of its terminination impedance by using a filtered gain block right?

[dmills]

Ah, I got the wrong end of the stick, I thought you were trying to make a VHF/UHF SDR do HF, not microwave...

What are your performance goals?

For a fixed 2GHz LO, a low phase noise crystal osc and multiply it is entirely conventional (And generally quieter then a YIG or such).

Look at something like a 'Driscoll' oscillator followed by the usual cascade of tuned multiplier stages, MMICs are your friends for getting the level back between stages. A PLL and VCO module looks attractive, but usually has horrible noise performance up there. 

A conventional diode ring mixer will work, providing you can produce sufficient LO power (at least 50mW or so), but there are also printed structures with carefully placed diodes that you see in the microwave literature.

73 M0HCN