Behemoth update incoming..

I built myself a PC in 2011, and it’s finally due a major update/refresh. It’ll still be a bit of a behemoth..

The 2011 build was a hex-core Intel i7-3930k watercooled with a Corsair H100, 16Gb DDR3, with originally a Radeon 5770, with an Intel i520 SSD and 4x 1Tb harddrives for a Raid 10, all sitting on an Asus P9X79-Pro motherboard. A pretty decent spec of a machine, and a very good performer especially when overclocked.

Over the past 9 years I updated the graphics first to a GTX660 in 2012 and recently to an RTX2080. The CPU was swapped out for an E5-1680v2 Xeon. Storage was updated to 500Gb SSD, and to two 6Gb drives in raid1.

It’s become a good idea to future-proof this pc again, for the next few years at least. Incoming are the following:

  • AMD Ryzen 9 3900X, 12 core, 3.8-4.6 GHz clocks
  • Asus ROG Strix X570-E Gaming motherboard
  • 16Gb DDR4-3600 Corsair Vengeance RGB memory
  • 960Gb Corsair MP510 NVMe SSD.

Paired with my Zotac Maxx RTX2080 and the stack of storage, this beast of a desktop will continue to earn the Behemoth moniker. It’ll be a few weeks until the bits are installed and configured, but it’ll be a really fun time I think!

Coronavirus? My thoughts..

So Covid-19 had reared it’s little coronated head on the Island of Ireland. And the country is ill-prepared for the probability of what’s going to happen.

Here’s a middle of the road scenario based on the current data available. Cases appear to be doubling every ~6 days. 15% or so need hospital care, and about 5% of those needs ICU-level care. Deaths are averaging at about 3.5% at the moment.

Given there’s about 22 cases in Ireland (excluding the North for this one), we can expect the following:

  • March 9th; 22 cases,
  • March 15th; 44 cases,
  • March 21st; 88 cases ,
  • March 27th; 176 cases,
  • April 2nd; 352 cases, with ~50 hospitalisations and 12 deaths
  • April 8th; 704 cases,
  • April 14th; 1408 cases,
  • April 20th; 2816 cases,
  • April 26th; 5632 cases,
  • May 2nd; 11264 cases, with ~1600 hospitalisations and 400 deaths.
  • May 8th; 22528 cases,
  • May 14th; 45056 cases,
  • May 20th; 90112 cases,
  • May 26th; 180224 cases,
  • June 1st; 360448 cases,
  • June 7th; 720896 cases
  • June 13th; 1441792 cases
  • June 19th; 2883584 cases

That’s assuming current exponential growth. In reality the case increases should start to have less new cases by about the middle of May due to a self-limiting transmission (less uninfected people available) and a continuation slowly up to maybe 100 thousand cases by mid summer and maybe 1500 total deaths by then.

That’s a middle-of-the-road back-of-envelope calculation, but based on valid current numbers and trends. It really makes for sobering reading.

The elderly are more heavily affected by this disease, with a 5% chance of death in the over 60s and a 15% chance of death in the over 80s. For healthy adults below 40 the rate is currently 0.1% chance of death which is x10 times the standard winter flu (even including the flu shot for the winter flu).

What we should do individually is to limit our social contacts, minimising contact with people, minimising our own exposure to infection by practicing simple good personal hygiene, minimising touching our faces with our hands without disinfection. We should also limit the exposure to the elderly from children, who appear to be able to transmit the virus without appearing symptomatic ourselves.

The authorities should start the minimising of travel, both internationally and nationally. Sporting fixtures should be without the crowds of fans watching – if the fixtures go ahead at all. Pubs and nightclubs should close for the duration of this epidemic. Parades and marches should not be let go ahead. Public gatherings of more than 500 people should not be organised. If this is done, the rate of transmission will be lower, and the peak numbers at any one time of infected and treatment-seeking people will be lessened, hopefully to levels that our health system can cope with.

If we don’t take strong action in the very near future, we will have a proper national-level emergency to deal with, and a lot of our parents and grandparents will not live to see the next Christmas.

If you think I’m scaremongering or panic-inducing, please look at the north of Italy and see what’s happening there, and realise that there’s very little difference between there and here in our societal habits. A little fear is a good thing, if it can get people to think and to plan. Have a stock of medications to hand, and be prepared to stay at home for up to a few weeks. It’ll hurt, it’ll be terrible, but we should be able to weather this upcoming storm if we do that.

If you think that this is not much worse than the flu, there’s a doctor or two in the hospitals in Bergamo that would like to have a chat with you. Link here: (Google machine translation here)

How to do speed cameras correctly.

We are often told that speed cameras are a tool to improve road safety. However, the method of implementation does not fit that requirement.

To do speed cameras correctly and use them as a tool to improve road safety, there are a few prerequisites.

  1. The locations that are high-risk need to be correctly identified, ideally to within a 50m stretch of road. This would allow camera monitoring to happen at very specific areas that are known to be accident blackspots. This does not refer to 20km lengths of road – as currently used for criteria of camera location selection.
  2. The locations of cameras must be publicly and freely available, and kept up to date. The location database must be unrestricted in use so that satnav companies can use that data in their products – allowing advance warning of known dangerous areas. Mobile camera locations must be advertised either in the database or on a specific website a number of days in advance. The camera locations must also have their local speed limit associated with the camera being visible in the database.
  3. Permanently installed cameras and mobile camera vehicles must be highly visible, with fluorescent and retroreflective paint, visible to drivers before entering the monitored area. Cameras must *not* be hidden or camouflaged, either deliberately or accidentally
  4. Camera locations must be clearly signposted at the roadside in advance of the installed location, and the area of camera visibility should be easily visible on the road.
  5. Camera location warning signs must have the local speed limit listed on the sign.

Following these would mean that the brief of improving road safety via speed-sensitive camera monitoring would be clearly met.

For example, let’s take a hypothetical place. Let’s say there’s a 1km straight road (100km/h limit) with a road junction in the middle (80km/h limit for the 50m either side of the junction), that has seen a number of fatal collisions over the previous decade. The correct way to put a camera on this is to put a signpost 300m either side of the camera location saying “warning Camera Ahead – 80 km/h limit. The camera should be a bright rescue orange colour with retroreflective strips on the camera housing.

This would ensure that the posted speed limit at the junction would be advertised and observed by drivers, and drivers that are over the posted limit would be caught.

There’s absolutely no value in stating that speed cameras are to protect, when they are used mostly with ambush tactics. The cameras currently in use do not look onto the areas with the accidents, they appear to concentrate on the areas that are actually accident-free. The zones used to determine the dangerous areas cover such a long distance that they are meaningless for specific area protection.

Let’s call the current situation in Ireland as it is. The current implementation of speed cameras, with the private operator using non-reflectively-marked vans, parking on the edges of straight stretches of safe roads, not advertising the locations – this can only be described as a money-making measure as it *cannot* fulfill the brief of improving road safety.

As an aside, it’s a great indication of the sneaky ethos behind the current operator’s operation of the cameras, where the vans have a hatched decal applied to the van surface. This decal, while appearing similar to the reflective hatching applied to truck trailers, is muted in colour and utterly non-reflective. It appears to provide only lip-service to safety, and acts to better camouflage the vans when parked up. This can only undermine any public confidence in this implementation of the system.

Daylight savings – some thoughts.

One of the definitions of timezone being correct for a location, is that the average Sun is due south at local noon on the clock. This means that there’s about the same length of time from sunrise to 12:00, as there is the length of time between 12:00 and sunset.

In historical times, the local church would ring the bells based on the local time, such that at 12:00 the Sun would due south. Of course, due to different places being at different longitudes, the Sun would be due south at different absolute times. There’s 4 minutes difference for each degree of longitude difference in the timing of the Sun being due south.

This caused some problems when the railways allowed relatively fast travel east and west as each location had been operating on its own time. There were some issues with timetabling and advertising the arrival and departure timings of trains as a result of those differences. It was then decided that the railways would operate on a time frame that was consistent for the train company such that when the train company clock said it was 9:00 in Paris, it was also 9am in Brest. Previously it would have been 8.32 in Brest as it’s 7 degrees further west than Paris. Each location with a railway station then started to use the railway clock as their local standard, and the countries started to operate on a consistent time zone.

The advent of telegraphy and radio was another pressure on places to have a consistent time that was the same set of numbers on the clocks in each location. It became easier to have time signals that would allow easier synchronisation of clocks.

Then, in the 20th century, it was agreed to standardise timezones across the world, with Greenwich in London to be the zero point. As Greenwich was defined to be the zero meridian of longitude, this meant that there were 24 time zones of an hour difference around the world, each separated from the next by 15 degrees of longitude. Once passing 7.5 degrees east or west from the center line of a timezone, it was at that point that the next timezone was due to start.

Of course there were some political and social considerations in play now, as it was not really useful to have a country split into multiple timezones. Portions of Ireland are far enough west that they should be in the GMT-1 zone, but it was decided to put all of Ireland into the GMT zone. France and Germany decided to work in the GMT+1 zone, even though Paris is close to the center of the GMT zone.

That’s a little bit on the history of timezones.

The current status quo is that Ireland is set in the GMT timezone, and we change to GMT+1 from approximately Spring equinox to the Autumn equinox. This means that in Dublin on July 1st, the sun is due south at 13:30. In Killarney on the same day the sun is due south at 13.43.

There’s a decision that has been made to stop the annual change of the clocks for DST, and that change is to be welcomed. The change has been seen to cause health issues due to the enforced body clock changes with changes in sleep patterns, and there’s also a well-proven uptick in accidents also mainly due to the sleep pattern changes forced onto people around the clock change.

One argument that continually gets put out there for DST is that it’s safer for the children at school, that they travel when it’s brighter or get more time in the light after school. It would be a much better change to change the start time of school to suit, instead of forcing a clock change on everyone.

I would suggest that Ireland go to GMT ans stay on GMT year-round Any businesses that operate with European groups can change their start and end times to suit the Eurpean office timings, e.g. starting at 7am and finishing at 4pm. It would be the same effect on our bodies as if we were GMT+1 and working between 8 and 5 in that timezone. The agricultural sector follows the solar day anyway, so the listed numbers on the clock are of no relevance to a cow’s desired milking time.

To recap: Ireland should go to GMT, and stay on GMT. Schools should be able to change their start times to suit the light conditions, if the lighting conditions are considered to be a timing requirement. Business should have the office openings as appropriate for their business. Doing all of this ensures that our clocks are correct as per the Sun.

Zurich is lovely.

I hear Zurich is lovely this time of the year..

Also it’s apparently a lovely drive from Ireland via ferry to Fishguard and using the Eurotunnel onwards to the continent. I also hear that parking is fun in that city..

More to follow depending on updates.

Boeing proxy fails the sniff test..

Read this link first – https://www.moonofalabama.org/2019/09/14000-words-of-blame-the-pilots-that-whitewash-boeing-of-737-max-failure.html

It’s a sane critique of a recent NY Times article that appears to be from Boeing PR by proxy: https://www.nytimes.com/2019/09/18/magazine/boeing-737-max-crashes.html . To be honest, I would have expected better from the likes of the New York Times.

And Boeing wonder why the European flight authorities will not be certifying the 737 MAX until the Europeans have tested and certified under European testing – any FAA certification understandably won’t be transferred given the debacle.

Flightradar rankings

A few weeks ago, I set up an ADS-B listener, started uploading to FlightRadar24 with the ID T-EINN18 , and I’ve watched my ranking go up and up..

As of this morning (August 20th I have the 43rd highest ranking of 130 in Ireland, 3rd highest Shannon uploader, and worldwide ranking of 6204 of 23,000, and going up by ~2000 positions a day at the moment.

I’m using a RaspberriPi 2b, an RTL-SDR dongle, and RTL-SDR 22dB LNA, and the antenna is a self-built colinear coaxial cable antenna with 8 116mm segments taped to the inside of an upstairs window. The software is dump1090 and the Flightradar24 uploader. I’m seeing ~850 aircraft a day, and an average daily max range of ~170 nautical miles (250km). Not bad for €50 or so of an outlay! I can see planes about 8000 feet above Dublin and 11,000 feet above Knock. The curvature of the earth combined with the lack of height of the antenna prevents me from seeing any lower.

The ADS-B signals are transmitted by every commercial aircraft and by most private aircraft, on a frequency of 1090 MHz. The information transmitted includes at least the ID of the plane and it’s current position and speed. When these signals are collected and processed, aggregator websites like Flightradar24 can show the current state of our aeronautical skies.

Improvements planned are the installation of an ADS-B specific filter/LNA from the RTL-SDR people that is en route. I did get one already but that failed within a few hours of installation so the warranty replacement was sought. That addition should mean that I have a bette chance of hearing planes through the house and increasing the range between my southeast and the west, as the window the antenna is in is facing to my northeast.

It is fairly interesting to see what is possible with some fairly basic items and a bit of DIY electronical knowledge!

KiwiSDR installed and running

So, I got myself a KiwiSDR. This is now connected to the random longwire I’ve strung across the rear lawn. It’s available on http://kiwisdr.swibble.com:8073

The KiwiSDR is a completely self-contained device, that consists of a Beaglebone Green micro computer, with an attached ‘cape’ that provides 30mhz of HF bandwidth to ~8 tuners. This means that (in one configuration) there are four tuners each with a waterfall available for use. Because the device is self-contained and accessible across networks without the use of a PC to host the tuning, there’s great flexibility in the device that can play and display the radio signals.

I’ve mine set in a configuration where there’s only two tuners with waterfalls, but there are another four allocatable tuners as a result. I’ve also got a number of those tuners operating as a reporter for WSPR signals on 60m, 40, and 20m.

Currently the antenna feeding this SDR is a ~20m random longwire strung across the rear lawn in a North/South direction, feeding 50m of coax cable and ending up at a Nooelec 9:1 unun. There’s a reasonably low noise level on this antenna for some reason, though the signal level is also a tad low. I have plans for either an Active Antenna amplified loop antenna, or something like a Wellbrook loop or a Bonito amplified dipole

My KiwiSDR is indexed on http://sdr.hu and available for public use.

Adventures in SDR in suburbia..


I’ve had a bit of time to try out a few different antenna ideas, and it’s been a very interesting learning curve so far.
Self-build antenna fun:
The first antenna attempt was a simple short randomwire of a few meters whiule sitting on the lawn with one of the laptops and the RTL-SDR dongle. I was able to get some SW stations from Saudi and Eastern Europe with that.
Second antenna build attempt was that I set up about 20m of insulated copper wire across the lawn, and fed a coax cable to the receivers directly from the centre point. This gave reasonable results, but definitely left a lot to be desired. I tried better grounding of the coax sheath, and then I realised that I should try a proper centre fed dipole with one side earthed instead of both sides going to the signal feed. This improved things a little bit, but was still a little underwhelming, most likely due to the horizontal orientation and the 2m of distance from the ground.
I rejigged the antenna in the third antenna attempt to become a proper ~20m long randomwire, end-fed to the signal feed in the coax, and earthing the coax cable at that point as well. This gave a good step up in sensitivity and signal strength. Then, after a long-awaited Ebay delivery of a 9:1 NooElec balun, things got quite interesting. I modified the balun to become a proper unun after breaking the centre tap to ground connection, and I installed the unun at the receiver. This is the current best state of the longwire, performance descriptions to follow.
I had ordered a 4:1 voltage balun as an experiment. With 10m of cable each side, I had tried to have a 20m dipole with a proper correctly matching balun in DX operation. However, the experiemnt was not a success as I found that there was no LW or MW, and the signal/noise levels as seen by the receivers was quite low – lower than the pre-balun randomwire. This antenna has since been disassembled, but I’ll hold on to that balun for a possible future transmission antenna.
After the purchase of ~200m of 0.5 square stranded cable, I decided to try a long loop around the perimeter of the property I’m currently living in. I ran ~80m of cable along the tops of the bushes and walls around the edge of the back lawn (maybe 2-3m average height), raising up to 6m height by using the soil pipe vent stack and a gutter downpipe as supports for the wire. I can’t yet put the wires muvh higher due to possible neighbour issues. Both ends of the wire loop were run into the radio room and attached to a 1:1 balun, and a length of coax then running to the receivers. The noise level is generally much higher than I had anticpated, and the signals are present at least. It’s a fairly good option at this stage if I had no other option. The balun in use has been shown to give some operation down to the LF area at least.
Miniwhips.
I started out with an AiExpress-sourced miniwhip, needing a 12v source, and 239 type connectors. I had a slew of 12v power supplies to hand, and found that the RFI generated by most of them was utterly abysmal and made attempting to use them for VLF through the lower end of SW almost impossible. . I did find the best of them, and it gave a clean enough power feed, and I did get to use that miniwhip to try to listen to things. Noise levels were still into the -60dBm with little in the way of useful SNR, and the signals were definitely not clear and listenable. I may yet disassemble this one and replace a few components to get a useful antenna.
Then, I ordered a Chirio miniwhip with options of either 9v PP3 power or 5v bias-t power. This was a thinner antenna, and had ~6m of coax already installed with an SMA connector on the end. This antenna gave much better signal levels and a lower noise floor than the Chinese miniwhip, but still had a lot of crackling. Adding a better earth connection to the SMA connector on the cheapy Chinese upconvertor dropped the noisefloor by some 10dBm or so. I’ve currently got the antenna powered from the USB socket on the Pi3 that is hosting the FC0012 dongle and the rtl_tcp server that is serving this across my network.
Antenna performances – as like-for-like as I can get them.
The miniwhip has a noise floor currently of some -103dBm. With the RTL-SDR at auto gain and using 2048kHz bandwidth, Shannon Volmet at 5505 has a signal level of -80dBm., giving an SNR of 23dBm. The sound of the signal is not great, but definitely understandable.
The horizontal loop with the 1:1 balun, with the Airspy, 2.5MHz bandwidth, linear gain at 15, zero visual gain, has the noise floor at -82dBm and the same Shannon Volmet signal at -71dBm, for an SNR of 11 dBm. The sound quality is poor, and not very understandable at all unfortunately.
The randomwire with the 9:1 unun, with the RSP1a using 2048kHz bandwidth, RF gain at 9, IF at auto, and visual at zero, has a noise floor at about -134dBm and a Shannon Volmet signal of -103dBm for an SNR of 31 dBm. This signal sound is definitely the best of the three antennas currently in use.
Receivers.
I have a handful of RTL2832U dongles that I used when I was getting started in the SDR world. I’ve some FC0013 tuners, as well as an E4000 tuner. I bought two RTL-SDR 820t2 dongles with the antenna kits, and those were an absolutely fantastic introduction into the SDR world. The direct sampling mode is reasonable, though the lack of gain control is a bit of a pain. I’ve got one of the dongles set up with an upconverter on the Chirio miniwhip, and serving to my network.
I later bought an SDRPlay RSP1a, and an Airspy/Spyverter combination. The RSP1a is pretty much the best bang for the buck SDR system that I’ve come across. VLF through 2Ghz without a break, no external upconversion required, usb-powered, and a lovely clean signal output. These are all very useful characteristics. The Airspy combo also has its place, it’s worth noting. Linux and RPi support is better, the hardware is physically smaller, and there are a few additional programs that are very useful such as the fast-sweep spectrum analyser and the ADSB client. Both of these receivers are currently plugged into the Dell Precision laptop, and being successfully served to my network via the SDR Console V3 server.
When I’m working locally on the Dell Core I5 laptop, I prefer to use the Airspy combo with the 10Mhz bandwidth. If I’m listening to one particular signal. I’ll pick it out with SDR# and using the decimation feature, to get the best possible SNR for that signal. The 10Mhz is very useful for the general browsing of signals and seeing which one I want to home in on.
Future plans:
I’ve spent some time looking at the KiwiSDR as a useful addition to my radio listening system. The 30Mhz bandwith visibility, the perfectly networked interface, and the extra plugins currently available, added to the four separate receiver channels possible, are all very enticing for my use case.
I also have my eyes on a better LW/MW antenna than I currentl have access to. I’m carefully eyeing up the likes of a Wellbrook ALA1530LN loop, or an Active-Antennas AAA-1 set as a quad loop, or even a Bonito Megadipole type of thing. I’m spending quite a bit of time looking round at the KiwiSDRs that are publicly available, to see how different setups can perform.