Ron W8RJL wrote:
> I cut my teeth on relays so I understand SP, DT, NO, NC, lathing etc, etc. I
> don't think there is anything parasitic involved, by parasitic do you mean
> the capacitive coupling between the moving contacts, stationary contacts and
> the electromagnet? Some relays have more capacitive coupling between
> contacts and between contacts and electromagnet/armature than others and in
> RF work less coupling is normally better. When replacing relays in RF
> circuits it is always best to try and maintain the parameters of the
> original relay since just maybe the design engineer knew what he/she was
> doing. 73, Ron W8RJL
>
I'm recalling (vaguely) that W8JI was concerned about the parasitic C
from the wires connecting the armature wires to the coil (inducing RF
into the coil and control circuits, and degrading the isolation when the
switch is "open" as well) and the series L of those wires (which tend to
be fairly small diameter). With the "shorting bar between the moving
contacts" scheme (which is what those relays have), the RF current flows
through the fixed contacts (which have "strap like" connections to the
base pin) and the moving bar.
Whether that's needed is a whole 'nother issue. Certainly, there are a
lot of relay box vendors using fairly vanilla SPDT or SPST relays.
I suspect it all comes down to isolation and loss specs, and whether
they are advertised/guaranteed. The TopTen sixway, for instance, claims
40 dB isolation at 30MHz, but the accompanying data plot shows it
doesn't quite get there.. about 35dB isolation between some pairs at
30MHz. The DX Engineering RR8A-HP has substantially better isolation
specs (70dB below 30MHz), but I think that comes because they use two
relays in series.
If you really get serious, you can get your relays from DowKey or
Transco or Sector.. HP used to make high performance relays, but I don't
know if Agilent still does. Clearly, they have really good relays
inside their test port adapters for network analyzers...
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|