Musings on mixers, tri-filar wire and crystal filters.

[CJay]

So I've built a couple of double balanced mixers using FT37-43 cores, home wound Tri-filar wire and various diodes (1N4148, 1N6263 and now I'm attempting to find some HSMS-2829 diodes) and they certainly seem to work but...

How much influence on performance does the wire have?

I can find all sorts of information about how long the wires should be to get a number of turns on the core, how many twists per centimetre, what percentage of the core should be covered etc. but so far I've not really found anything 'simple' about how wire diameter affects performance, maybe because it can't be made simple?

Of course the thicker the wire (within practical limits) the easier it is to handle, but am I doing my amateur DBMs a disservice and not giving them the best opportunity to perform for me if I choose a wire that's too thin or too thick, is there a sweet spot or a way to calculate the optimum by rule of thumb or rigid equation?

Maybe I'm thinking too much into it?

I've also recently got my hands on half a dozen NATO marked crystal filters, centred around 43MHz and using, I think*, 8 crystals. They seem like they may be ideal for use in a homebrew receiver (HF bands) but I'd find it quite useful to have some data on them, does anyone know of a way to find data from a NATO part number or even just what they were intended for use with?

I could (and will) attempt to characterise them for myself but my CMU200 doesn't want to play nicely with GPIB and FRERES as yet and I'm not confident how I would work out input/output impedances and match them to prevent skewing results anyway.

* I think 8 crystals because my over enthusiastic disassembly efforts may have scattered a crystal or two around the garden but in my defence it is a  massively overengineered device that required a *LOT* of heat to take apart.

[German_EE]

I can't remember the formulas for this but  the wire diameter, the insulation thickness and the number of twists per inch affect the impedance of the overall circuit. Just think of networking cable, the twist is designed to make a 100 ohm transmission line. I have one design for a mixer in my files where one transformer must be made using 100 ohm twisted pair and the other must be made with 50 ohm twisted, not easy.

The only source of information for your filters that I can think of is a guy called Ben Knock G4BXD. He's been collecting military radios for decades and can be reached at military1944ataol.com