When you see the S-Meter reading values above S0 when the antenna is disconnected or connected to a dummy load, this behavior is to be expected. The radio is not malfunctioning.
Let me explain why.
In traditional radios, the S-meter data is derived from the AGC voltage. It is not a directly measured value of actual signal strength.
The FLEX-6000 Series radios digitize a wide bandwidth of RF spectrum (it is done ~245 million times a second for the FLEX-6700), using the digitized data to accurately and directly measure the RF power in each FFT bin (the FFT bin is a small segment of bandwidth where we perform a Fast Fourier Transform that separates RF data into its frequency and magnitude components). Depending on the zoom level, the FFT bin size can be as small as a few Hz.
RF energy (radio frequency electromagnetic radiation or RF EMR) that the radio (actually, everything because the RF EMR field is pervasive) is exposed to and the ADC digitizes, comes from multiple sources and are cumulative. The sources of this energy are natural (the Sun, lightning, the earth itself, and the ionosphere), man-made (transmitters, electronic devices, etc..), and thermal noise (always present in electronic circuits). The radio does not discriminate between the different types of RF EMR. Also, there is a small electrical connection between the antenna port and the input to the ADC that can also act as an antenna because that circuit does not have infinite RF isolation. The takeaway here is that there is always RF energy that the ADC is digitizing, even with the antenna disconnected or a dummy load attached to the radio. It is just about impossible (it's physics) for there to be no local RF EMR the radio is exposed to.
So what we have here is a system that is doing the following:
1.) Digitizing RF EMR in a broad spectrum of radio frequencies
2.) There is always some level of RF energy being digitized.
3.) The radio uses FFT to separate the frequency and the magnitude (strength) of the digitized RF in small segments.
Now, in those FFT bins, in addition to the frequency component, we also have the measured magnitude (RF power) of the signal. We use this measurement to generate the panadapter so that you can visually see the spectrum by plotting frequency on the X-axis and magnitude (RF power) in dBm on the Y-axis. It is the measurement of the RF power in each FFT bin used to calculate the S meter reading.
So how do we calculate the S-meter reading?
The standard technique that spectrum analyzers use and FlexRadio does the same is to define a unit bandwidth for the measurement. The unit bandwidth in SmartSDR is the size of the receive filter.
So for example, let's say we have an FFT bin size of 1 Hz and the slice receiver's RX filter is 500 Hz wide. This means we have a 500Hz filter which is comprised of 500 FFT bins.
Now we integrate (sum over a range) the RF power in each one of those FFT bins to get a value of total RF power contained in the 500 Hz filter. Since there is always some RF power in each bin, when you sum them up, you are going to have a value of total RF power for the entire RX filter bandwidth. This is how we directly calculate the RF power in the RX filter bandwidth.
This is easily demonstrated. In the example above, if you start with a small RX filter size, say 500 Hz, record the S -meter reading in dBm. Now Expand the RX filter to 3 kHz wide. The S-meter will show a higher level because there are more FFT bins used to calculate RF power inside the 3 kHz RX filter bandwidth than there are in the 500 Hz filter bandwidth.
More bins = More RF power = a Higher S-meter reading.
By using this method to calculate the S-meter value, our radios very accurately and directly measure RF power. Many commercial businesses use our products as spectrum analyzers due to their accuracy when measuring RF power in a defined receiver's filter bandwidth.
Now, to complete this explanation, I need to talk about the relationship between the S-meter reading for a slice receiver vs the signal level of the panadapter. As I mentioned previously, the panadapter signal level is based on the RF power in the FFT bins. In a simplified description, each pixel on the display represents the RF power in a single FFT bin. The RF power in each bin is always changing in magnitude. If we show this directly, then the panadapter signal level would look very chaotic; it would be a wide fuzzy line that would mask weak signals. So we perform averaging on the FFT bin over a very short time before we display it in the panadapter so it smooths out the data is more visually appealing and makes finding weak signals in the noise easier to see.
So you cannot equate the level of the panadapter as measured on the y-axis as the signal level because its absolute magnitude has been averaged over a very small bandwidth, whereas, for the S-meter, there is no averaging of the FFT bins, the total power of each bin is summed.
I hope this explanation helps you better understand how direct sampling radios operate, how RF power is measured and displayed for the user, and why you will always see some signal level represented on the panadapter display even when the antenna is disconnected.