I am doing S21 and S11 measurements with a plus4 on a Noolectric LANA and am seeing strange results.  The bottom graph shows the results port1 connected to the input and port 2 connected to the output.  I used 12" cables and calibrated at the end of the cables.  This results in a ripple on the S21 with the average less than 10 dB.  The top graph has added a 20 dB attenuator between the port 1 cable and the input.  The ripple only at the lowest frequencies and the S21 value starts at about 27 dB and tapers off around 16 dB at 2 GHz, taking the 20 dB attenuator into account.  This is more in line with the part spec.  I am trying to understand what is causing the incorrect S21 values without the attenuator.  I tried adding extra cable length to the input and varied the input power from -30 to 0 dBm, these made no discernable difference in either case.  I thought it might re related to the S11 value, but neither the S21 value nor the ripple rate or magnitude seem to correlate in the second graph.  Also, I see similar results when I drive the input with a Pluto.  Any ideas?

Thanks,

Bruce
KO4PRE

This is an LNA. To properly evaluate it, you should not use more than
roughly -30 to -40 dBm on the input port. That's why it looks so much
better with the 20 dB attenuator on the input port. Without that
attenuator, you are severely over driving the LNA and what you measure is
not representative of its performance in its intended application - very
low level signals.

With -10 dBm input power and 26 dB gain (per the spec sheet), you are well
within the spec 1 dB compression point (-10 dBm + 26 dB gain = +16 dBm - if
it were linear in this region). The spec for 1 dB compression
(non-linearity) is minimum of +13 dBm. You're at +16 dBm, only 2 dB below
the nominal output 1 dB compression point.

Dave - WØLEV

On Sat, Sep 25, 2021 at 4:09 AM <bwschober@gmail.com> wrote:

> I am doing S21 and S11 measurements with a plus4 on a Noolectric LANA and
> am seeing strange results. The bottom graph shows the results port1
> connected to the input and port 2 connected to the output. I used 12"
> cables and calibrated at the end of the cables. This results in a ripple
> on the S21 with the average less than 10 dB. The top graph has added a 20
> dB attenuator between the port 1 cable and the input. The ripple only at
> the lowest frequencies and the S21 value starts at about 27 dB and tapers
> off around 16 dB at 2 GHz, taking the 20 dB attenuator into account. This
> is more in line with the part spec. I am trying to understand what is
> causing the incorrect S21 values without the attenuator. I tried adding
> extra cable length to the input and varied the input power from -30 to 0
> dBm, these made no discernable difference in either case. I thought it
> might re related to the S11 value, but neither the S21 value nor the ripple
> rate or magnitude seem to correlate in the second graph. Also, I see
> similar results when I drive the input with a Pluto. Any ideas?
>
> Thanks,
>
> Bruce
> KO4PRE
>
>
>
>

--
*Dave - WØLEV*
*Just Let Darwin Work*

That occurred to me too.  I tested both cases down to -30 dB and see the same results with and without the attenuator.  Also, with 0 dBm into the attenuator, -20 dB to the LANA, I did not see the ripple.
I have attached the Pin to Pout graph from the spec sheet. I would expect to be good up to -20 to -10 dBm depending on the frequency.

Thanks for the feedback.

Bruce

I append my S21 measurements of the Nooelec LaNA.

The attenuation for the measurements was 20 dB, so you have to add 20 dB to the vertical axis.

Questions:
Do you use Bias-T power or external power? Voltage (3,3..5V)?