[Radio Wrangler]

Thanks, there's a lot of information there!

Noise concerns say that you want a low noise amplifier as early as you can get it in your structure, and you want to minimise losses before it.

Big signal handling concerns say you want good narrow selectivity as early as you can get it in your structure and you don't want much gain until you've got through your channel-defining filter.

Clearly these concerns are at loggerheads. And dynamic range concerns say you want as much as you can get of both.

So we wind up with a nasty compromise and no real optimum solution. Consequently people have played many, many games with these structures and parameters.

The devil really IS in the details. Details of how good an amplifier/mixer/filter you can make can cause dramatic shifts in the structural choices... and you're still left wondering if there was a better solution.

We can't be surprised that there are so many variants.

Al was asking about broadcast radios of the 40s and 50s. Did radios compete then on detailed performance? Did buyers and advertisers understand such things... I don't think the radio world ever has its 0-60 time equivalent. I'm sure most manufacturers of broadcast sets just did what seemed like a good idea at the time, working to get sensitivity for their more exotic/expensive sets, and just living with any consequences.

So we see a wide profusion of approaches.

It's a valid issue and a class of problem which we still have no general solution to. If we do too good a job in one respect, we can usually twiddle things to trade some of the benefit off into a more needy area.

It still gets frightfully empirical.

I designed receiver structures for HP on elaborate spreadsheets modelling thir performance. In many respects these were models, not exact provable calculations. The spreadsheets told me what I'd got, but not how to do better. To an extent a lot of design software does analysis, but you're paid to do synthesis. So you make a guess, analyse it, vary your guess to analyse it and see if you're onto an improvement and so on and so on. It seems unscientific. Yes there are sub-areas which are analytic and provable, but any real system outgrows them rapidly.

If there were analytic solutions, we'd need computers alone and no designers....

David