Recent Noise Mystery Solved

 

You may recall a recent blog regarding my installation of a new pair of inverted-V dipoles, fed from a common coaxial feedline at around 85'. Both antennas 'hear' very well and everything was peachy until I recently put the inverted-V's on my Perseus SDR.

I had been particularly interested in using the new high antennas for my general shortwave listening on 40m and below but what I observed was immediately troubling when listening on the 5 - 6MHz, 49m band. After much head-scratching about what I had noticed, I posted the following 'help' message to the Perseus reflector as well as to the Perseus SDR Facebook site:

Perhaps others can help explain something odd that I have just noticed with my Perseus while comparing antennas.

I was comparing signal levels and noise between a 40m half sloper and a very high dual 40/80m pair of inverted-V dipoles fed from a common feedline. Both the sloper and the dipoles are fed with 50 ohm cable and all three antennas are well matched at the low end of the band(s).

Listening to a 6 MHz signal from China National Radio around 10 am today, the signal was around S7-8 on the Perseus, using the 40m sloper. Listening to the same signal, at the same time, on the Yaesu FT-1000mp, with the high inverted-V, it was slightly better, maybe by 5-6 db and overall lower noise. I then put the inverted-V onto the Perseus and there was not even a trace of the signal! I made this check with several signals and always with the same result.

Now I suspect that the SWR of the 7MHz inverted-V when used at 6MHz, is very high and the load presented to the Perseus antenna input is likely highly reactive and far from 50 ohms but that doesn't seem to bother the FT-1000.

I then ran the inverted-V through my antenna tuner so that it effectively produced a 50 ohm non-reactive input load for the Perseus and did the tests again...with the same results. Swapping antennas for the 40m sloper once again produced the same signal levels in both the Perseus and the FT-1000.

So what is going on here and why does the Perseus balk at the 7 MHz inverted-V while listening on 6 MHz? Is there something in the Perseus front-end analog filtering system that is overly sensitive even though the reactance was tuned out via the tuner? Is it the 80m V on the same feedline as the 40m V that is causing some still unwanted reactance that is not tuned out with the tuner?

Any ideas what is happening here as it looks like I will not be able to use the inverted-V antennas on the Perseus for general SWL out-of-band listening for some, as yet unknown, reason.

As you can see, I was completely mystified by what I was hearing, or rather not hearing, and as it turned out, completely off the mark.

I received a few replies offering some possible reasons for what I was seeing but none of them proved helpful in solving my dilemma ... until Roelof Bakker (PAØRDT) weighed-in! Roelof suggested that I look at the antenna's performance while running Perseus's built-in 'HFSpan' function.

HFSpan is a stand-alone 0-40MHz spectrum analyzer, that comes with the Perseus software. Although I was aware of it, I have only used it sparingly. I next did some screen captures with all three antennas, one at a time, and sent them to Roelof.

His analysis did not take long as he immediately identified my problem ... a very high noise floor when using the mysteriously-performing inverted-V. Roelof suggested some common mode choking to eliminate the problem.

Already having an isolating transformer in hand from a previous experiment, I inserted it directly at the Perseus antenna input and looked at the noise floor again, around 40m. The transformer was wound on a small FT87-J core with a 3 turn primary directly opposite a 3 turn secondary. I was astounded to see the background noise floor drop from -85dbm to a very quiet -110dbm!

Evidently there was a lot of noise being picked-up on the inverted-V's feedline shield. Not hearing any of this noise on the FT-1000 indicated that the noise was probably associated with the Perseus power supply, the laptop or the laptop's power supply. This immediately explained why I wasn't hearing anything with this antenna when used with the Perseus.

Roelof then suggested that a 2 turn / 2 turn transformer, offering less inter-winding capacitive coupling, might provide even more isolation ... and he was right again. A further ~4dbm lowering of the ambient noise was measured.

It may not be pretty but it produced an astounding improvement!

Further comparisons between the FT-1000 and the Perseus revealed that the FT-1000 was still producing a slightly better SNR than the Perseus, when using the inverted-V so evidently there was still some noise affecting the signal. I had a few very large #43 ferrite toroids and decided to wrap a few turns (5) of the RG8-X feedline around the toroid to see if there would be any improvement.

Once again, using HFSpan, I compared the noise floor both with and without the #43 choke and saw a further 6-7 dbm improvement! Going back to comparing SNRs between the FT-1000 and the Perseus, I now saw no difference between how I was hearing on both receivers, when using the troublesome inverted-V ... eureka!

Here are the 'before' and 'after' screen shots using HFSpan, the Perseus built-in spectrum analyzer. Both screen shots were taken at the same time (mid-afternoon) and show the results of the noise mitigation work.

0-10MHz sweep - with noise problems

Same sweep, noise eliminated, signals now unmasked

As of yet, I have not determined the actual noise source. I suspected it may have been coming from the Perseus power supply or from the laptop supply but that was not the case. Perhaps it is coming from the laptop's processor via the USB cable which I will also choke and see if HFSpan reveals anything further

This problem was a great learning experience for me, in more ways than one and I am most grateful to Roelof for his detective work and experience with noise issues and for taking the time to respond to my initial inquiry. Hopefully you may find something here that can help you as well.