Bob W0EG 2025/03/15 12:50
Agreed, however, a one-port measurement on a long cable run can still be
useful. The scaler form of S11 is often represented as Return Loss in dB.
To me that is a more intuitive representation of performance than VSWR.
Return Loss is the difference between forward power and reflected power
(power returning back to the source). If you place a good short on the far
end of the line, ideally all power would be reflected back and the return
loss would be zero dB. If the return loss is 6 dB, then half of the total
power is lost in the system, so the line loss would be 3 dB. It doesn’t
matter where in the line or how the losses occur. You just want to have an
idea of how much of your transmitter’s power will reach the antenna.
To keep your transmitter safe and happy, a return loss of 20dB is
considered good (1.2:1 VSWR). If the return loss is less than 10 dB
(1.92:1 VSWR) maybe time to do something to improve it.
On Fri, Mar 14, 2025 at 11:59 PM Tom W8JI via groups.io <w8ji=
w8ji.com@groups.io> wrote:
> S11 is really just derived from a measurement of the impedance deviation
> from the calibrated or normalized port impedance. It is NOT a direct
> measurement of loss, it is simply a measurement derived from operating
> impedance error at the port.
>
> It is a horrible way to determine loss, especially when cable loss is low
> or the cable system under test is imperfect or non-constant impedance.
>
> The most accurate way to measure loss is to directly measure loss. A
> calibrated S12 or S21 transmission loss would directly measure cable loss
> instead of relying on a measurement of impedance at one point in the system
> (the port). We know the power going into the line, we know the power out
> the other end of the line. The S12 or S21 is far more direct than
> extracting loss from S11.
>
> 73 Tom
>
>
>
>
>
>
> On 3/13/2025 1:28 PM, Jim Lux via groups.io wrote:
>
> How so "not remotely as good" for cables with not too much loss (say,
> <10-15 dB)
>
>
>
> If the S11 measurement has some uncertainty, then the two way loss
> measurement has that uncertainty. Dividing the loss by 2 doesn't change
> the percentage uncertainty. These VNAs don't have markedly different
> uncertainty between the reflected and thru paths - they're all using the
> same receiver, and in a reflection measurement from an unterminated cable,
> the SNR of the reflection measurement is pretty good. The same is true of
> the Thru measurement - high SNR.
>
>
>
> (Unless the cable is truly horrible - if it has 30 dB of loss, yeah, the
> reflection measurement will be -60 dB, and that is getting down in the
> noise, vs still good SNR for the thru measurement at -30)
>
>
>
> Most of the VNAs have reasonably good (few percent-ish) uncertainties even
> with a reflective load (after all, you're close to a cal value).
>
>
>
> I would think the real question would be whether the far end is really a
> perfect short or open. I would think that for an HF measurement, the error
> is "small", for microwave "not small". e.g. what's the S11 of an open SMA
> plug vs an actual calibration open? I'd expect a bit more C than a "cut
> the transmission line off square to the end" because you have a 1/2 cm long
> "air dielectric" transmission line between the center and the coupling nut
> (which moves).
>
>
>
> The other thing you can do with a swept S11 is do an inverse FFT and look
> at the timedomain response, which can reveal oddities in the cable.
>
>
>
>
>
>
>
>
