I stand corrected; yes I neglected mismatch loss.
Re-calculating using datasheet max values gives Pdiss values that
decrease with increasing frequency, which is what I would have expected
(but didn't get the first time). I realize datasheet values are just
that, all I'm trying to do is to get a feel for what the relay can do.
I can see from the discussion that measurements can't predict
everything. Fortunately for me I don't use 'elbows' or RA-coax
connectors at all (too many bad experiences at work, even with
'professional' grade stuff).
Thanks all for playing!
On 25/11/2023 20:24, Richard (Rick) Karlquist wrote:
You misunderstand what the definition of "insertion loss" (IL) is, as
used by RF component vendors. IL = incident power (input) -
transferred power (output). The device under test (DUT) has both
"mismatch loss",
where some of the incident power is reflected back to the source and
therefore is not transferred to the output; and dissipative loss,
where power is lost by being absorbed by heating the DUT. So called
mismatch loss does not result in heating of the DUT. Your
calculations are not valid. Hope this makes sense.
73
Rick N6RK
On 11/25/2023 10:51 AM, Scott Townley wrote:
If that's the case, then wouldn't an insertion loss test reveal such
an issue?
For example, I was deconstructing the power handling of a Charter
Engineering B5-series coaxial relay (N-female connectors). If I map
"max insertion loss at x GHz" to the average power rating on their
published chart, at every point the dissipated power comes out to 11
watts. That tells me that "good" N-connectors can dissipate 5 watts,
and for that mechanical form factor I should be able to derive a
power limit based on measured insertion loss.
Or are the physics of power handling in "not-so-good" connectors
different from what would be revealed by an insertion loss
measurement? Assuming of course that we are not in the realm of high
voltage breakdown of the dielectric (or maybe that's exactly the
difference?).
--
Scott NX7U
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|