Beechlog‎ > ‎

September 2000

The on-line magazine of the Burnham Beeches Radio Club.

World Wide Web Edition September 2000
Here we are again, late as usual. In fact I've missed out a whole issue this time. My excuses? Well, I've been pretty busy at work, slaving over hot equipment, and also making a few trips here and there. And none of you lot have been much help, no articles or anything.

I've kept to the same format as last time. Since I received no comments, I'll assume you could all read it OK. It's all plain html, and the only fancy bit is the embedded style sheet, which is pretty minimal anyway. So lets see if I can remember how to use this editor!
Roger G0HZK, Editor
Munich in April, Roger on his travels.
GPS and How it Works. Just in case you don't know
Science in London A trip to the city.

External Links
The BBRC website. You can find 'what's on' here.
The RSGB for all about Amateur Radio
IEEE special issue on GPS
Germany Calling...
Off again, this time I have some work to do in Germany. I have been jetting around this year! This time to another country I have never visited before. It's Tuesday morning, and I have checked in at Heathrow for the 7.45 Lufthansa flight to Munich.
There was a nice queue on the airport spur this morning, even though it was only 6.30 am. But I was lucky - by the time my wife got up, the queue had stretched along the M25 and had made the BBC TV news.

The plane took off about 10 minutes late. It was a smooth flight, with little talk from the cabin crew, unlike last weeks Paris flight, where the passengers were told that "we have put the cat out and buttoned down the hatches" and suchlike! But a couple of hours later I stepped outside again into the warm sunshine at Munich Airport.

Unlike John 'GCL, a seasoned traveller to Germany, I speak not a word of the language. I managed to find a taxi whose driver spoke mainly Urdu (I think). Whatever it was, he spent a good deal of time talking into his mobile - I gathered he was having difficulty finding my destination. However we did get there, although it meant driving at 140 km/h in the airport, and overtaking on all sides of the road, including the hard shoulder, slip roads, and so on. I paid 80 DM for this fairground ride! Anyway the office is next door to the hotel, so I shouldn't get lost or need a taxi until I depart. I decide it might be a good idea to use the train for the return trip to the airport. There is a station here, Unterföhring, which is on the line to the airport, so all I have to do is find it!

Arriving here, I discover my equipment is still en route from Paris, so I get some time to start writing this stuff. The sun is shining and it is quite warm, I guess about 75 degrees. Quite nice after a dull weekend, but alas I shall spend most of my time inside. Eventually my equipment turns up and I get down to work. I get out of the office at about 7.30 pm. I booked an hotel close by, almost too close, for part of the hotel is two floors above me in the same building! But my room is in the next building, about 100 feet away. So soon I am in my room, which overlooks the office. I spend the evening in the bar. A meal of Pike and asparagus,washed down with too much beer. Most of the folk in the bar speak German, except me, which is OK because the barman speaks good English. There are some Americans, who speak loud English, and some Chinese, who speak a little English.
I talk to a German guy who works in Maidenhead, small world! And finally ring the wife via my mobile, E-Plus it says on the LCD. Oh well, day 1 over.

The second day starts with breakfast, luckily a standard buffet so I don't have to speak German. The day soon passes, as I have plenty to do. As the afternoon draws on, it clouds over and we get some rain. But it stops by the time I've had enough of working, so I walk across the car park to the hotel. In the foyer I meet Per, an engineer from our Stockholm office. I have to give Per some training the next day. As I drunk too much beer last night, I decide to give the bar a miss, and Per and I go out to dinner. We got directions from the receptionist to some restaurants, and set off on a half mile walk to this huge pyramid-shaped building.

There are a number of restaurants here, and we decide on a Greek. The staff speak German and Greek, we speak English and Swedish. Oh well, it turns out that they have a menu in English! I don't think I've seen so much meat on my plate before! These Germans don't seem that keen on anything else. I order some "Unfiltered Ale", and actually get a drink that looks like British beer, only a bit cloudy. It doesn't taste like bitter, but it doesn't taste like lager either. There is the complimentary Ouzo too. After an hours struggle, we leave enough meat to feed our families for a week. The food was very cheap, it must be the weak Euro. We manage to navigate back to the hotel somehow.

Next morning I had agreed to meet Per for breakfast, but although I had set the alarm on my Psion, it was still set to UK time, so I was an hour late. It was a mixed day, some clouds but some sunshine. I have booked on a late flight back, so I have a little time to look round Munich. So I get lunch in the Bistro below the office. It's a sort of fried soft cheese (none of the meals have vegetables). The menu in the Bistro was written in incomprehensible German.

I finish off with some training, and then head for town. You miss out so much travelling by taxi, so I wait at the bus stop outside the hotel. The bus destination is written on the board - Arabellapark - which is a U-Bahn (underground) station. The driver actually sells the tickets, so I use the fare to work out the cost of travel all evening! Munich has a zone system, with fixed fares between zones. This is explained on bus stops and stations, in German. It wasn't at all clear to me, but the Munich transport system seems to have almost no staff to check you have paid your fare. On buying a ticket (bus, tube or train - all the same), you immediately cancel it in another machine. Everything uses machines, no humans. There are no ticket offices on the undergound. There are no barriers on the stations. The bus to Arabellapark takes about 15 minutes, an interesting tour around the suburbs of Munich. The bus ticket said 2 zones, and travel to the centre of Munich was also 2 zones, so no trouble calculating the fare.

The centre of Munich has a mixture of architectures. There is an Italian church, highly decorated and coloured. There is a Romanesque old town hall, gothic civic buildings, and a huge 17th century brick church. I wander around for a while, there are familiar shops like C&A, Woolworths. The traffic-free streets are crowded with people and lottery ticket booths. Not many street cafes like Paris (Although I was to find these on my next trip). Eventually I make my way to the S-Bahn station. This is identical to the underground, but runs further out of Munich. I find the airport on the maps - 12 zones. There isn't a fare for 12 zones, just 1, 2, 3, and "ab 4 zonen". I choose the latter - about £5 - rather cheaper than the taxi. There is no-one to sell tickets, only the machine, so I stick in a 20 Mark note and get my ticket and change.

Once again no barriers. Indeed some of the stations on the journey have no fences - the road runs alongside the platform. These Germans are very trusting. After about 20 minutes the train arrives at the airport. The escalator leads straight up to the central check-in area.

There is about a half mile walk to the BA check-in, which seems the busiest in the airport. This is surprising because there are absolutely no signs telling passengers how to get there. Here I discover the plane is still at Heathrow - a thunderstorm at Munich has flooded a runway and delayed many flights. So I prop myself up at the nearest bar where there is a steady supply of Weißbier. The flight is actually delayed about 90 minutes - it becomes the last flight of the day on the departures board. It takes only about 10 minutes for everyone to board the plane. Being the last flight means there are no delays and we travel across the airport at a fair lick. Soon in the air and back home.

Roger G0HZK
GPS - How does it work?
One of the most curious applications of radio technology has been the GPS system. This was developed for the US military to replace an earlier system, but was also designed to be used for civil applications, albeit at a reduced accuracy. As you probably have heard, the deliberate distortion of the data on the unencrypted civil channel has now been removed, greatly improving the results obtainable. Cheap receivers can now measure location, speed and time to a precision not easily obtained by other methods. So how does it all work?

This has always been a mystery to me, although it is clearly evident that they use centuries old techniques of triangulation. Traditional means require you to take sightings with a theodelite from two locations a known distance apart, indeed the whole of the UK was mapped this way from a single reference distance over two centuries ago. But with GPS, you sight from only one point. How can it work?

In the early days of satellites, scientists needed away to measure the orbits and other characteristics of the objects they launched into space. If the satellite had a radio beacon on board, they could use the doppler shift of the signal to assist in their calculations. Soon it was realised that if the orbit of the satellite was known precisely, the location of the observer could be calculated. So the US Navy launched six satellites in known polar low earth orbits. This system, known as Transit, had drawbacks, like you had to wait an hour and a half between passes, and needed heavy dut processing to produce results. However, the system served its purpose until four years ago, when it was decommissioned.

GPS however is much quicker. You do not have to wait for satellites to pass over your sky, and the maths is quite simple. Normally you need to see four of the twenty four satellites, however if you know some of the parameters, you can use less, even "Transit" mode and do with just one.

Firstly a mention of the satellites themselves.
There are a variety of different GPS satellites up there. The design has improved over the years, giving better results to the users. Recent satellites weight about a ton, and are cylindrical - about 6 feet across and 6 feet high, with a span of about 38 feet across the solar panels. The panels generate a little over a kilowatt. There are two transmitters provided for users, one on 1575.42 MHz, known as L1, and another on 1227.6 MHz, known as L2. The L1 frequency is used by civilian GPS receivers, and the transmitter provides about 25 watts to a 13dBi antenna. The orbits are about 21,000 km above the earths surface, which means a path loss of about 182 dB. Incidentally encrypted signals are provided on both L1 and L2 outputs for US government users, both frequencies may be used by receivers to provide better results.

The nature of the signal is complex. A spread spectrum system is used, the civilian signal spread over about 2 MHz. All satellites use the same centre frequency, but the spread spectrum code is arranged so that no satellites interfere with any others. This is just as well, for the received signal is very weak, maybe below the receiver noise level.

There are 24 of these satellites in a "full set", and there are usually a few spares in orbit. They take about 12 hours to orbit the earth. This means that any user with a good view of the sky will "see" at least 4 of them. In fact, at the moment of writing this, I can "see" 7 from the middle of my garden, which is rather obstructed (I can see the sky only above 45 degrees from level ground.

So what information does each satellite send?
There are two groups of data sent, known as "Almanac" and "Ephemeris". Anyone familiar with tracking the Oscars will immediately recognise much of this data. The Almanac data consists of the approximate orbit data for every active satellite. So if a receiver downloads the Almanac from the first satellite it locks onto, it will receive sufficient data to calculate the current positions of all the other satellites.

This data is the standard stuff used for calculating satellite positions, like Orbital Inclination, Right Ascension at Time of Aquisition, Mean Anomoly, Argument of perigree, and so on. This data is updated periodically, but its intention is only to help enable the receiver to decide which satellites to use for its position calculation.

The Ephemeris data is the real accurate stuff. This consists of orbital data, clock corrections and the like. This accurate information has to be downloaded from each of the satellites we are receiving.

So how is this information used?
The method used is almost exactly the same as used by surveyors centuries ago. The traditional method of trianglation works like this: (i) observe the unknown position from two known positions, (ii) with a theodelite, measure the bearings to the unknown from each point, and (iii) measure the distance between the two known points. Then do some simple GCSE maths. The whole of Britain was mapped like this 250 years ago, using only one distance measurement on Hounslow Heath - the error over the length of Britain was only a few metres.

To use GPS, we do the same thing, but of course the observer is located at the unknown position. This makes it slightly more difficult! We can calculate the accurate positions of the satellites easily, using the Ephemeris data, but how do we measure the distance between the observer and the satellites?

The answer is to measure the time the RF signal take to travel from the satellite to the observer. Knowing the speed of light, we simply multply the time by the speed of light. But how do we know when the signal left the satellite? We need a clock in out receiver that is locked to the satellite system clock. But all we have is a crystal oscillator. Crystal oscillators are quite good, however, so we can read the time from one satellite, and add to it the approximate time the signal takes to arrive (say 50 mS), and set or receiver clock to the resulting time.

But before I go any further, we should consider how many satellites we need to take measurements of. If the observer and the satellites were in a flat plane, we would need two satellites to for the triangle and measure the 2 dimensions we need. But since we are in a three dimensional world, we need three satellites. Simple. Still GCSE maths. However, there is a fourth dimension, time, so in simple terms we need one more satellite to compute this fourth dimension.

It works like this. With four satellites, we can use at least two different groups of three, and calculate two fixes using our rough distance guess. Most likely, these fixes will not be the same, because our distance guess is not correct. So all we need to do is to alter our clock a little, and recompute. We do this until the two (or more) fixes converge to the same point. So that's all there is to it! Simple really. Well I have omitted problems arising from the fact that the satellites are in motion. Modern GPS systems take this into account by determining doppler data, dx, dy and dz, and working this into the equations. I did say earlier that I can "see" 7 satellites at the moment, so my receiver can calculate more than two fixes, and use the extra information to correct other unknown variables, like errors in satellite geometry and ionospheric distortion. Where it cannot do this, it can display a figure which alerts the user to the likely error in the readings. Mine says the error is within 13 feet - quite remarkable!

So what information can we get from a GPS receiver?
I have already mentioned what a GPS receiver measures while obtaining a fix, and these results can be presented in a variety of ways, usually with errors! The main purpose is to display a fix of the receiver position, which is primarily displayed as Latitude, Longitude and Height. This is a real can of worms! GPS natively uses the WGS84 datum for mapping the whole world. WGS84 uses an ellipsoid specified in both conventional ellipsoid terms, and in Cartesian terms. The former includes parameters such as the major and minor axes of the reference ellipsoid, and the latter includes the earths gravitational constant, the speed of light, and defined actual points on the earth.

So the GPS receiver produces a fix in Cartesian terms (x, y and z). This is converted to Latitude and Longitude for display. Most GPS receivers can also display locations using systems other than WGS84, but here there are limitations and potential inaccuracies. For example Great Britain uses a datum called OSGB36 for its maps. The datum consists of a reference ellipsiod called Airey 1830, after the Astronomer Royal who calculated it. The term datum just means that this ellipsoid is aligned to the surface of the earth at defined points. This datum is a much better fit to the geometry of Britain than WGS84, which is slightly tilted.

The problem which arises is that you cannot accurately convert a point on WGS84 to a point on OSGB36 - there will usually be an error. You can make an approximate conversion - in the case of Britain you can get 5 metres accuracy, which may be good enough. An unconverted reading can be several hundred metres adrift.

GPS receivers can also display projected grids, such as the Ordnance Survey NGR system. Conversion from x, y, z or Latitude and Longitude to the OS transverse Mercator grid is not difficult, but in the case of Britain, the scale factor varies across the maps (this is in order to minimise the measurement errors inherent in projecting a curvilinear surface onto a flat plane). The OS use a "rubber" formula which you have to pay for. You are not going to find this in a £100 GPS receiver! So you get an approximation, which may be good enough.

The Height displayed on a GPS receiver is of less accuracy than the x/y position. There are many reasons, none of them obvious (to me!). And what is it the height above? The ellipsoid, the datum, sea level (where)? I'm not going into this. It leads to too many arguments. There are some interesting documents available from the OS web site which explain the impossibility of getting accurate figures.

The GPS receiver, when fully locked to 4 or more satellites, knows the time with respect to UTC within 100-200 nanoseconds. But the display is not that accurate. Some receivers may be a second or more behind. Giving a user the accurate time is not the primary purpose of a receiver. It doesn't generally matter if the display is updated every couple of seconds, since most users have not moved far in that time, and probably have stopped to read the GPS. So the time displayed is usually updated at the same time. Even the serial output of most GPS receivers is not timely.

But a GPS receiver can be used as a very accurate clock. It is just a matter of designing it for the purpose, so that the display or digital output signal keeps up with the internal clock. GPS systems are often used as accurate time sources since they are extemely convenient and work almost anywhere in the world where the sky can be "seen".

The accuracy of the time carried on the GPS satellites is specified as being with 1 microsecond modulo 1 second (see a few lines down the page). Each satellite broadcasts three versions of the time, which are (i) the time kept by the satellite clock, (ii) the version of UTC kept by the US Navy, and GPS time. The differences between GPS time and UTC are also broadcast, these corrections enable a receiver to provide a clock accurate to within 28 nS of UTC. One interesting point is that GPS time is not changed when UTC is corrected by the addition of "leap seconds". These means that as the years go by, the GPS clock gets further behind UTC. However, the system accomodates this by broadcasting the "leap second" correction every 12.5 minutes. If you perform a complete software reset on your GPS receiver, you may well see this difference until the receiver downloads the correction.

The GPS receiver can also measure speed very accurately, provided it can keep "sight" of sufficient satellites. In practice, it takes surprisingly little time for a receiver to determine the speed at which the user is travelling. The original design objective was to provide an accuracy of 0.1 metres/second. It can also determine the direction in which a user is travelling, which is quite obvious really, considering what it is really measuring.

What of the future?
There is a rival system, GLONASS, built by the Soviet Union. However, since the breakup of the latter, there are fewer satellites in orbit than was envisaged. Whether there will be a steady decline, or a revitalising of this service, who knows?

The EU are also considering launching a network of positioning satellites for civil use. Since the removal of the GPS Selective Availability, GPS has become considerably more precise, so the EU system will have to offer an advantage over GPS. However the EU system is being designed for civil use from the start.

GPS itself is not standing still. New satellites are launched from time to time. These replace those at the end of their 10 year lifespan, but offer better precision, due to improved technology. There are plans to improve the service, with additional transmissions on additional frequencies, in order to minimise some of the current inaccuracies caused by ionospheric effects.

Roger G0HZK
Science in London
I've had some time off from work this week, and as it was a nice day on Tuesday, I decided to have a look round London. It's some time since I've wandered around on my own.

Of course, things don't work out quite as planned, but after doing a bit of early morning shopping I drove the car to Langley station, hoping to park and catch a train. Ha! The first thing I discovered was that it cost 50% of the train fare to park and ride. The second thing was that the station car park was full up, so I didn't have to pay anyway.

I did a bit of thinking, and now have an inkling of why so many people drive to work in London every day. Ever since I started travelling by train in the 50's I was convinced that the railways were run by imberciles, nothing seems to have changed. Anyway I parked some distance away and walked to the station.

It is a long while since I've travelled on this line. I normally catch my London trains at Datchet, which take me to Waterloo. So it was interesting to travel on modern trains (the last time I went this way, the trains were the old rattlers, whose exhaust seemed to circulate round the carriages). It was interesting to see that the line was now electrified from Hayes to Paddington. I suppose it's only a matter of time before the overhead cables are extended to Slough and beyond. This electrification has been to accomodate the new Heathrow Express trains, but both Express and suburban lines have overhead cables now. Also the electrified lines must extend through West London to join other lines - I noticed several Eurostar trains parked at the depot at Old Oak Common.

While in London I decided to pay a visit to the Science Museum. It's been about 10 years since I was last there, and this year the museum has been reborn with the opening of refurbished galleries and new exhibits. However this has been at the expense of the removal of old favourites. In particular, the railway and shipping sections have been thinned out. Being a small boy during the days of steam trains, I have many fond memories, and was particularly disappointed that no 4073 Caerphilly Castle is no longer in the museum. This locomotive, built in 1923, was one of an extensive class of express passenger train locos that thundered up and down the GWR line through Slough.

Gone also are the working models - the museum used to fill with small boys pressing buttons and turning handles. But the big static steam engines of the late 18th century are still there, as is the collection of aircraft on the top floor.

I also had a brief visit to Tottenham Court Road. When I was younger, this was one of the centres for the electronics enthusiast, as was Edgeware Road and Lisle Street. Tottenham Court Road is still stuffed with computer and Hi Fi shops - might be a good place to visit when I run out of hard disk space? Edgeware Road used to have Henrys Radio, where you could buy pocket dosimeters and OC36 transistors, the wonderful ORP12, and all the TTL chips I built my Spectrum-powered RTTY outfit with. H. L. Smith sold aluminium chassis, 6BA nuts and bolts, and so on. I should think they have long gone. As for Lisle Street, the home of No 12 sets, PCR2's, command receivers. By the time you got this stuff to Paddington Station your arms were a foot longer. I still have some old catalogues and magazines somewhere...

Roger G0HZK