I only remember seeing distortion from the Lime if the tx gains are turned up full. The distortion may be in the spectrum analyser.
I have seen that here, though the analyser did start to complain when it happened.
That's a good point Charles - I didn't spot it but the level is -10 dBm in a 120kHz resolution bandwidth on a 1 MHz wide signal so the power into the analyser must be 0 dBm, 10 dB below the reference level, which one would think would be OK but worth checking.
The Lime on it's own, through an attenuator, at a presumably lower level on the screen should be OK. I am sure mine was much better than this except when driven to its maximum output, which isn't recommended. Perhaps another high dynamic range driver amplifier is needed. Getting the gain right is important and we have to be careful not to over-amplify the noise and spurious signals
I can confirm it is not spectrum analyser overload as I've reduced the signal level and also go the same result on two different analysers plus repeating the same tests under the same conditions with a DATV express tx results in a much cleaner signal.
Thus I can only conclude my Lime is not as clean as it should be despite have the full 5V applied from a USB3 hub.
I'll try again but with my spare Lime to see if that's the same.
but while my Lime is not as bad as your's it is somewhat worse than the other two.
I tried reducing the Lime output and it didn't make it any better.
The spectrum analyser was a SignalHound BB60C
your measurements are practically the same as mine for all three devices.
Interesting that Lime doesn't change much and in fact gets worse at lower powers at least in my case.
Ideally for signal fidelity reasons I'd prefer to use my DATV Express as the exciter, but as I also want to use a Jetson for H265 coding I figure that ties me into the Lime ?
I did write an application for Express that runs on the Jetson but it doesn't work properly and I have no idea why.
I ran it on a standard Intel Linux machine and it worked but not on the Nano. (Go figure).