LimeSDR Mini QPSK Phase Noise at 3.4 GHz

I have been using a LimeSDR Mini to generate DVB-S QPSK 333 KS/s digital TV signals at 3406 MHz. Initial results were not good, with a really bad received QPSK constellation showing imbalance and phase noise, but as the LimeSDR Mini warmed up, it improved a lot. I reproduced this by cooling the LimeSDR in the fridge and it happened again.

Today I have done more tests with 2 units. I can reproduce the problem on both hardware versions 1.1 and 1.2. The time from poor phase noise to very good is about 5 minutes on the bench - it took a lot longer on a cold windy hilltop where I first noticed the problem. There is no issue at 2405 MHz (or below), it just seems to occur at the very top of the frequency range.

I am not going to spend much more effort on investigating this issue, as I now know that I need to let the unit warm up before (very occasional) use at 3406 MHz. However, I wanted to let other users know that the issue has been observed and ask if anyone else has seen similar problems or solved them.

Dave, G8GKQ

It’s bee suggested that this is probably a calibration issue, as calibration will be to some extent temperature dependent. @Zack, do you have anything to suggest here?

Thanks Andrew

I have also had other suggestions that it might be some stored factory calibration (as opposed to the pre-transmit calibration routine) that does not adequately account for temperature at the extremes of the frequency range. @Zack could this be a possibility?

Either way, there is no big issue - just so you know that we are pushing this amazing chip to it’s limits! Once it had warmed up, the DVB-S TV pictures on 3.4 GHz were received 28 km away.



Nice! Was that with a PA? What sort of antenna?

Yes, I use a Stealth SM3437-43L-1Q power amplifier which has about 50dB of gain and will give 10 Watts of clean DATV at 3.4 GHz whilst draining a 12V leisure battery. The aerial was a 60 cm prime focus dish - seen in the photo below with a different (10 GHz) feed. Dave4%20Stealth%202