VDSL filter for the Kiwi SNR measurement
I've prototyped a VDSL "filter" to be used with the Kiwi SNR measurement (I used kiwirecorder for the prototyping). I works well for the one Kiwi I know about that has bad VDSL noise, causing it to have inflated SNR values.
If you know of other Kiwi's with this problem please comment here or email support@kiwisdr.com so I can try them out -- Thanks.
Basically, once the VDSL segment is identified it is excluded from the SNR measurement. Same as we have now for masked frequency segments.
Comments
The most obvious one is
http://zl4md.proxy.kiwisdr.com:8073/
It is consistently near the top of the SNR leaderboard, but has a massive block of noise from 8.5 to 12MHz
In the UK, but already poor SNR
http://21246.proxy.kiwisdr.com:8073/
http://remoteradio.changeip.org:8086/
http://kiwisdr1.hoka.co.uk:8072/
That ZL4MD one is what I used to develop the scheme. Thanks, I'll try the others. I'm curious to see how well it does in more marginal cases. I'm sure some adjustment will be needed.
http://remoteradio.changeip.org:8086/ is one of mine, but it doesn't have a VDSL problem, in fact we don't have VDSL at this location. I do occasionally have an issue with a neighbour and his DIY solar panel invertors, but I am applying pressure to him to resolve it.
Ah OK, however it would be interesting to see if John's filter, also responds to broadband noise that looks like VDSL.
I wonder if there are several consecutive FFT bins, that contain the same level, could this be interpreted as wideband noise rather than individual signals, and treated as such in the SNR calculation ?
Regards,
Martin
Well, this entire thing quickly becomes yet another SNR-related rabbit hole. One that I can ill afford at the moment given all the other important stuff I'm working on. I'm only doing it because of a person who is very unhappy the top NZ SNR score is due to VDSL noise.
I should have never added any SNR capability. The project has actually suffered because of it.
To be honest with you I was perfectly happy with the reception at my location on my various radios (up to 9 online) and judging by the many that regularly use them because of the good and generally free interference reception they are too. I have many that carry out QSOs, TXing from their own location, but RX through one of my radios.
But the SNR figures have caused me to become some what obsessive with getting the figures higher. It has recently struck me that I am probably wasting my time. I can still hear what I want, when I want and the only thing that changes are those damn SNR figures.
I find the RF spectrum display more useful as a gauge to how well I'm receiving and whether any new noise sources have popped up.
The only noise that bothers me is that between 1.5 and 2.5MHz, but that is only because I can see it, if I couldn't see it I would be none the wiser.
I concur that a single metric such as obtained through an SNR algorithm is not a good tool to use for evaluation in real life situations. I further agree that human examination of spectra, particularly broad bandwidth ones with an eye on noise signatures and comparison against ITU measurements such as Quiet Rural are valuable.
I have found that joining these kinds of interpretation, difficult as it is, with longterm WSPR spot results over many bands, MUFs ,and geographic locations further enhances the usefulness to real uses.
While we all may prefer simple answers, I don't personally find the best ones to be simple or easy.
Yesterday evening my SNR figures were driven down to 8/8 by static from lightning some 200 miles away. I wouldn't say reception was affected on any band.
So do these figures really matter and do we need this "League table"?
If you are looking to receive a particular signal, a high SNR doesn't guarantee that you will hear it on the receiver. Picking the right location will improve your chances more than going for the receiver with the highest SNR.
I agree that the SNR measurement function is flawed, and appreciate that you, John, may have spent more time on it than you would like. However, even in its current form it is far better than having nothing at all.
This is especially true when using TDoA, as it allows you to see at a glance which KiWi's would definitely not be useable for such purposes, and this "pre sorting" saves a LOT of time and otherwise wasted effort.
Likewise, by providing SNR colour coded markers on the KiWi maps, it is easy to see which ones are likely to provide reasonable reception. As only about 10% of the 800+ KiWi's deployed worldwide produce relatively noise free performance. It would take a long time to find these among the vast majority of poorly performing ones, without some sort of SNR or other quality indicator being in place.
I'd also suggest that the SNR value, performs a useful function in encouraging KiWi owners to consider improving their antenna systems, and chasing down interference sources.
Overall, I think the SNR measurement has had a positive influence on the KiWi network, as a type of quality control, and that without it, there would be many more poorly performing KiWi's in operation.
Regards,
Martin
For what it's worth here are my thoughts on the SNR topic.
My primary interest is how Kiwis perform at MF so a generic SNR doesn't help. For example a receiver could have an outboard bandpass filter (or HPF) that attenuates signals in the range of 500kHz and 1600kHz and yet still have a great SNR.
I totally concur with Martin's comment about the difficulty of finding Kiwis with good performance amongst the total population. That is why recently I conducted tests on all the accessible Kiwis in the UK to compare performance on the MW band.
Despite the decline in the number of MW transmitters I was fortunate that here the UK we have one transmitter located in the centre of the country with no other co-channel stations. Therefore I used this to test all the Kiwis in terms of how well this signal was heard in relation to distance between tx and rx.
The results were published in Medium Wave News and give a very good real world assessment of Kiwis in a particular band of interest. Of course this is not a substitute for other measures.
A receiver that performs well on MW may perform poorly at 30MHz because of the antenna, and vice versa. Hence a single figure of merit is just that.
With regards to TDoA I find that SNR doesn't indicate much - one just has to check if the target signal is receivable. Often receivers with apparently good performance cannot hear a target signal - on MW especially due to directional properties of loop antennas.
Notwithstanding these thoughts, I do agree with Martin's final two comments, and anything that encourages Kiwi owners to better their performance should be encouraged
regards
Steve G8KDL
Switching to any of my loops gives me massive signals, no noise and much reduced noise floor. I can hear every signal that is there. But the SNR figure is rubbish compared to the horizontal doublet. The doublet gives a good SNR, pushing me up the listings, but what is the point if it doesn't give the best reception? The more I consider it, the more I think SNR ranking is irrelevant for most operators and users.