[hpsdr] ALEX - Call for Comments - III

[CT1IZU]

> General:
> Interesting reading suggested by other emails earlier today:
> "G3SBI H-Mode Mixer as tested by PA3AKE"
> Specifically - preselector comments/measurements on Page 9.
> Best I can tell, we are using the improved topology he refers to.

Not so sure, -using the Coil Craft inductors may impose a limitation give
what appears to be the high IIP3 as by the Phils comments..
Work by Horrabin G3SBI on a 3 section ladder with a loaded Q of 8 gave IP3
as follows

 Micro Metals  T50-6 (coil Q 70)  +50dBm
Loadstone Pacific L41-6 (coil Q 70) +55dBm
------------------ L45-6 (coil Q 125) +47dBm
Toko 10K  (coil Q 70) +37 dBm

The Star filters used the Toko 10K  coils but specified the 145ANS type for
the 1.8-14MHz filters. These have a large cup core which screws OVER the
coil winding giving a larger core diameter and air gap in the magnetic
circuit. Note also that the filters were designed for a much different rx
archeticure ie a 9Mhz IF. In Merc wider filters may not be at a disadvantage
and would offer lower insertion loss.

I`m concerend about PA3AKE`s composite filter results with the T50-6 which
had a IIP3 around 36dBm. Subtract the differential off the Toko/CoilCraft
above would make them marginal to say the least.

The L45/41 are a pain to use as they are formers only, a similar problem on
the Toroids and the need to use a variable C.

No easy answers but something to consider.

In respect of C any rf filter like this needs NP0 caps or if you are flush
some nice Silver Mica.YV5 is a definte no-no (put one on a meter and hold it
between thumb and finger and which it move!). NP0 is easily available up to
220pf  and Murata and Epcos have some units up to 1000pf. From a kitting
point of view it may be better to parallel some 220`s for the required
higher values.



Shel CT1IZU


>



 1173394488.0