I’ve noticed that my SDR setup, an Apache Labs Anan-7000DLE transceiver used with the PowerSDR software, seems to rest the S-Meter at a quite high level when there is no signal present. This is unnerving when I try to compare my local noise floor with others who claim noise levels many S units below mine. I recently reinstalled PowerSDR and recalibrated the receiver amplitude. While I was at it I thought I’d document the HPSDR power meter performance, and also that of an Icom IC-7300.

These are some screen shots showing before and after calibration, and at the end a comparison with an Icom 7300 radio hooked to the same antenna. All the PowerSDR readings werre taken with the same receiver settings. The mode was set to AM and the filter bandwidth manually set to 2.4 kHz to be consistent with most radios in upper sideband mode. (In PowerSDR the S-meter reading is based on the total power within the bandwidth, so a narrower bandwidth will show a lower noise floor than a wider one.)

First, the display before calibration, first with a 50 ohm dummy load and then with a -50 dBm signal applied:

You can see that the meter reads about 6 dBm lower than it should. I can correct that by using the PowerSDR calibration routine that uses an external signal generator to calibrate the receiver chain. These are the results after doing that:

Now, the S-Meter shows exactly the signal generator level to be exactly -50 dBm, and the noise floor has gone up by 6 dB as well. The analog meter shows between S2 and S3 which, for -112 dBm noise power, matches pretty well the “Collins Standard” as shown by K9LA’s S-meter calibration table.

Now, let’s connect an antenna and see what happens:

Now the meter shows -89 dBm and S6 to S7, and again that tracks the Collins standard. The other rigs in Carl’s table would show a bit under S5 to S6. The noise floor shown on the spectrum display has gone from about -130 dBm to about -115 dBm.

Now, here’s the Icom IC-7300:

It’s showing between S2 and S3. For fun, I did a calibration of the IC-7300 S-Meter and came up with this table:

It’s pretty clear that the only thing Icom knows about the Collins Standard is that S9 is -73dBm. Below that point the scale is more like 3 dB per unit than 6 dB, and anything below -98 dBm doesn’t exist. That explains the very low noise levels this radio tends to report.