GPS Bottle

The GPS Bottle Story - yet more detail

this page updated 1600 Wednesday 2nd June 2004 UK time

for the "bigger picture", see the GPS Bottle Story

There was just toooo much on that first bottle story page - getting longer every day as the story unfolds. So here is some of the older stuff, with lots of nitty gritty detail for those very few of you who might be interested.

final GPS bottles in production

bottle building final bottle Robin's priority task over the coming days is converting our prototype design into five reliable GPS bottles that have a chance of survival for long periods at sea and bumping into coastlines. We hope the finder of a GPS bottle will return it into the sea to continue its journey: if not immediately, then soon after. It is the 45 smaller bottles that will carry a special message and things for the finder to do. We now have all the components including STEPP units, solar panels, timers, batteries and the important 2 litre Frascati bottles. The design is based on that proven by Snoopy in the sea trial, but changes include reducing the length of antenna wires, changing the shape of the battery pack, removing parts of the solar panel that were not required, and adding 1/2 lb of lead ballast at the back to bring the top of the bottle clear of the water. final bottle in water The metal cap will soon corrode in salt water and so this will need to be replaced or augmented with plastic stopper, screw cap and sealing tape. GPS bottle 2 on the left has a label on the antenna giving information to be agreed with Layla to supplement that carried by the main label along the lower half of the bottle which can be A3 size. The details of this big label will be provided by Layla and will be slightly different from the label on the smaller bottles carrying the paper messages for the finder. e.g. it will ask the finder NOT to open the GPS bottle and may suggest the finder look for one of those other smaller bottles which carry a special message.

bottle launch pattern - in time and space

bottle launch bottle forecast Our DRIFT model, described below, is also alerting us to the importance of launching the bottles so that they spread out rather than stay together and all do exactly the same thing. There are several good reasons for this including wanting to increase the chances that at least some have a fighting chance to make it towards the Chatham Islands, and to make the tracks more interesting.

Layla has explained that there will be 50 bottles in total: 45 small bottles holding the notes for the bottle finders, and 5 "tracked bottles" to help the software calculate where all the bottles are now and predict where they might go next.

This choice of bottle launch pattern will be determined by very practical issues such as over what period of time can we launch the bottles due to constraints such as how many boat trips can the project afford (only one planned), how long children will want to stay at sea before they get sick (2 hours ?), and how fast our little fishing boat can sail. These things are critically effected by the weather, of course. No doubt we will seek to make good use of the tide. e.g. in the DRIFT model run below, and picture on the left, the launch pattern was a diamond shape. The boat started south east (white bottles), then north east (blue), and back north west (green) then south west (yellow then purple), dropping bottles all the way. By the time it returned to where it dropped the first (white) bottle, the tide has carried these bottles far south. On the right you can see the dramatic effect delay of launch by a day or two can have. This is not surprising since the wind has the most long term effect and the wind can change a lot within a day or two, if not hours.

Final choice of launch pattern should be chosen to make best use of wind and tide.

modelling of tidal current and drift due to wind

bottle animation Very early in the project it seemed a good idea to model, however crudely, the drift of each bottle due to the tidal currents and the wind (and waves). The fact that we will be tracking several bottles very accurately, hour by hour, means that it should be possible to adjust the "fiddle factors" in our model, included with the GPS Software, to make the model accurate and track ALL the bottles - if we record the exact time and position of each bottle launching.

On the right is an animated GIF file produced using the excellent "GIF Construction set" software from and 190 BMP images from one of the latest experimental versions of our DRIFT model, which is intended to provide "artistic tools" for Laylas' exclusive use. You can look at the larger size version here. In this particular run, the bottles hit the French coast. They hung around on the beach until the wind shifted before resuming their journey more widely seperated.

bottle dog The drift/wind factor is one of the most important of these "fiddle factors". It is the speed and direction of bottle drift due to the wind and resulting waves. We are using 3% (0.03) and in the same direction of the wind - but this will depend on things like how low the bottle is in the water. This is where we have started to do some real tests, on large inland patches of water on a windy days, measuring exactly how far (and therefore how fast) bottles move in a given wind speed. "Bottle Dog" on the left, and his radio controlled boat, have been recruited for the deployment and recovery of GPS bottles at places like local reservoirs. Tests on Robins pond showed he can happily tow a bottle tied on with string, so we now have a suitable matching wire hoop and hook "docking system".

Strange as it may seem, the experts at places like the Met Office - to whom Robin has spoken, and is very grateful - say that the drift may be at an angle to the wind due to the Coreolis effect. The more tests we do now the better: a realistic model could be used now to make continguency plans for where the bottles are launched, according to the weather.

drift modelling For those interested in the DRIFT model detail, here on the right is the PC screen of version 1b of the model simulating five bottles. The big white arrow over the red trails is the wind direction, controlled by the values near it. We can put these wind directions and speeds in for the next 7 days after visiting Weather Online for Dover but right now we are simply trying all sorts of wind directions and speeds, to get a feel of what we might expect. Below the wind controls are those used to set up the times and locations where the bottles are launched.

The smaller blue arrows show the tidal current, which oscillates back and forth with the tide, on a 12 hour 24.5 minute cycle. We are using a SIN function, with parameters of maximum speed and direction, interpolated from several points taken from Tidal Stream Atlases. Robin could use your help: he needs the simple equation ("A" level maths, "linear interpolation" ???) which provides an "averaged" value (of direction or speed) based on a list of values in X,Y (lat/lon) space, which might perform better than the equation he is using now. If you can help Robin - please contact him via the contact page. The other area where Robin would like practical help is in tests to measure bottle drift due to wind.

The predicted tracks of five bottles launched from near Ramsgate can be seen. The tracks are green where we are reasonably confident we will receive messages from the bottle, because they are within 5 miles of those mobile 'phone masts (green squares) along the coast. After 5 miles the trails go yellow, and then red outside 10 miles, since we are pretty sure they will be out of range. But we can easily adjust these numbers if we have better information - but do not forget that we are not talking about a mobile 'phone being used stood on the deck of a ferry, tens of feet above the waves - our bottle system is only just an inch or two above the waterline ! Small circles can be added along the trail to show where the bottles might report position every hour.

We hope to make the model sufficiently accurate that it will be a useful tool in planning details of bottle launch details, and then, as we adjust it from incoming GPS bottle reports, it will tell us reliably where ALL the bottles are. Also, fed with the 7 days wind forecast, we might have a good idea where the bottles might be in the next few days, including where they might wash up on a beach.

But remember, it will only be a model, and no substitute for "real life" - which always holds surprises up its sleeve :-)

glass or plastic bottles ? clear 2 litre wine bottles ?

glass prototype Lorina bottle I understand that Layla has now chosen the type of bottle to be used for the majority (maybe 40?) that are not fitted with GPS tracking - the smaller glass 0.75 litre Lorina lemonade bottle seen on the left. We now have a simple, proven electronic design for the five tracked bottles, capable of reliable operation within a moderately sized clear bottle, and now we must now build the actual tracked bottles to be launched in mid May. These must be constructed to look good and survive over what may be many months or years at sea. Plastic was convenient to work with for the early prototypes, but glass might better survive the occasional brush with a rocky coastline, or a few weeks erosion by sand on a deserted beach. glass prototype front glass prototype back Glass also seems a more "artistic" choice of material. We've discovered how to break the bottle where we need to from a search on the net for "bottle cutting" and finding the page. in short: 1) scratch with a glass cutter. 2) heat with a candle flame. 3) pour on cold water 4) repeat 2 and 3 until it breaks. I've been looking for a suitable clear glass bottle (1.5, 1.7 or 2 litre ?) to see if this is our preferred solution. On April 3rd I found what I thought might be a suitable bottle: Budgeons "SOAVE DRY 2003", 1.5 litre clear glass bottle at only 4.99 - and you do not need to throw away the contents :-) This bottle has a metal screw top with soft plastic inner seal, and it seems probable this bottle is used for other drinks. On Monday much of the work was done of squeezing the electronics into the bottle - and it WAS a squeeze. It looks as if we need a bottle of similar shape but slightly larger. However, shaving the odd millimetre off the sides of the solar cell, and lower corners of the timer, just about gets it in. The final versions cannot afford to be a tight fit, or different expansion rates might cause the bottle to break. If this bottle is used, then the label inside needs to be no larger than 21 cm x 17 cm. The timer at the front can be reached by a suitably bent stick. Right now the 20 available on/off programs are set to every hour, other than the small hours of the morning. i.e. 2300, midnight, 0200, 0500, 0700 then every hour again. I now understand that the final bottles will not include flashing lights or mascots, but our new glass prototype has Bart Simpson sat in the back with his hand held GPS :-)

2 L Scrumpy bottle Note from Robin on 18 April: the 1.5 litre sized glass prototype seen above sits a little too low in the water (almost exactly 50% below - including the top) compared with our plastic one which sits a little higher (no weight of glass). 2 L bottles Latest view is that either we must find a larger suitable 1.7 or 2 litre clear glass bottle, or revert to our original plastic bottle design. This is no big problem since our electronic package is the same, fitting in any of these bottles. We are also doing drift/wind tests on a range of bottles carrying the same 950 gm of electronics worth of ballast. Here on the left is the only large clear bottle I've found: the 2 litre scrumpy bottle just purchased from Waitrose. It was not too difficult to cut off one side, above the waterline - as seen here, testing in the rain. Note that the handle has weight, so must hang down. I would prefer a more elegant wine bottle: any ideas ?

2 L bottle floating bottle temperature Note added on 25 April: maybe we've found it: the 2 litre Frascati bottle seen earlier and here on the upper right with a bunch of friends :-) Robin would like to say thankyou to the "Mezza Luna", his local Italian restaurant, who kindly donated three boxes of empty 2 litre bottles. They were green, and therefore not suitable to take the solar panel, but they have been invaluable for our tests on Bray Lake which need to continue. I've now tested our solar package in this 2 litre bottle - that's Snoopy sailing it around our pond - and I've now started some temperature testing. That "greenhouse effect" inside the bottle raises the temperature from 21c outside to over 40c inside the bottle - electronics can only stand so much ! Hopefully this will not be a problem: the water around the bottle should help keep it as cool as Snoopy :-)

choice of SIM card and GSM service provider

The sea trials gave excellent results using O2, Orange and T-Mobile, and we are now talking to the mobile 'phone network providers, hoping some will sponsor air time for one or more bottles. We must now decide what SIM cards to put in bottles since this will have a major impact on the running cost of this project - particularly if these bottles continue their journey for many months or years, reporting their position every hour. At 10p per report this ammounts to over 2 per day per bottle, perhaps more off some countries. We do not expect any cost when far from land, of course. We intend to use SIM cards that also have a data line, which we can use to "dial into" the STEPP unit and "pull out" the recorded GPS history for bottles that drift out of GSM coverage, such as into the North Sea or Atlantic Ocean: maybe even the Pacific - see the Press Release from Turner Contemporary below.

prototypes used in sea trials - Tweetypie and Snoopy ! :-)

Tweetypie Snoopy On the left you see Tweetypie, our mascot in the bottle prepared by Robin, holding 12v of two 6v lantern batteries and Laylas' STEPP #1. This bottle is a "stretched" lemonade bottle, 48 cm high instead of the usual 35cm, made from two plastic bottles, one cut high and the other low. The bottom holds 750 gm of ballast to keep the bottle upright and the combined GPS/GSM antenna about 5cm above the waterline. On the right you can see the bottle being tested in Robins' garden pond. The two bottle halves overlap about 3 cm, are smeared with Geocel flexible aquarium sealant, then bound by two types of outdoor sellotape - we hope it don't leak ! The ON/OFF switch can be accessed by a stiff wire from the top of the bottle. Layla may also have time to make a bottle, similar to our first prototype, holding a single 6v lantern battery - it will be interesting to see if it performs as well as the 12v version. We expect any sea trial this week to be based on the first prototype already tested over several days of continuous running: i.e. one or two 6v lantern batteries switched permanently on, and not using the electronic timers. This will ensure that we are in continuous GSM contact with the bottle, so that we can be sure that any problems are due to the location at sea rather than components, such as electronic timers, that are still being tested.

Tweetypie in water Snoopy closeup Our latest prototype is a 2 litre plastic bottle, floating horizontally, holding the STEPP unit, with its combined GPS/GSM antenna, a 12v 1400mAH rechargeable NiMH battery pack, a solar panel, and two electronic timer units. The timers arrange that 55 minutes of each hour are spent with the STEPP switched off, giving the solar panel the opportunity (during daylight hours) to re-charge the battery. Every hour the timer switches on the STEPP, and waits two minutes to give time for the GPS to start tracking and registration with the GSM network to be established. A text message is then sent to the mobile 'phone, providing identification of the bottle, and the GPS data including latitude-longitude position, speed, course, date and time of fix in Universal Time (GMT). A small bright flashing light is also switched on for these few minutes each hour, to assist eventual bottle recovery, give us confidence that the system is working, and permit investigation of options such as optical tracking of bottles from the shore - at dusk or hours of darkness. Switches accessable from the top of the bottle will support us switching the bottle system from "continuous mode" (for initial sea trials) to "timer mode" (for longer duration). Yes, that is Snoopy, navigating the way with his handheld GPS, and stood in the space reserved for a small paper letter.

Use of digital timers

bottle timers bottle timers The section below on timers was written some days ago, and we experienced difficulty in getting these modified timers to work reliably. Problems included mechanical relays welding closed or not working at all, and timing not being reliable after adding much larger capacitors. Recently we had what may be a small breakthough: use of digital timers from mains products such as these on the left. The first type we tried with success was the "7 day Digital Timer L28AC" from Maplins at 10 - seen on the left in both pictures here. The timer runs off its own 1.2 volt cell - which was replaced by a much bigger one. Most of the mechanism including mains plugs and sockets and higher voltage circuitry, was thrown away. The switched output (rising from 0v to 1.2v) is used via suitable transistors to switch our 6v, 9.6v or 12v supply through to the STEPP and a flasher. These timers seem to be very accurate and reliable - tests continue. The main limitation of the L28AC is the fact that it can only be programmed (via buttons on the front) to switch on and off up to 8 times a day. e.g. every 3 hours, or at whatever time is required (less often at night?). However, we found the "SURE time, 24 HR/7 DAY ELECTRONIC TIMESWITCH model ST77" was similarly priced, had a similar internal design, and permits 20 on/off events per day.

Note from Robin on Tuesday 31st March: I just spoke by 'phone with Mark Sullivan, Sales Manager of Timeguard Ltd in London on and they have kindly offered to sponsor our project with the free supply of the ST77 product.
Daves bottle timer on Friday 2nd April: a dozen ST77 timers arrived from Timeguard today. Measured the tiny current drawn from the 1400 mAH battery at less than 2 micro Amps - which means over 80 years ! But I expect we will still trickle charge it from the solar panel main supply - just to be sure :-) Many Thanks Timeguard - these will be put to good use in the coming days.

on Friday 9th April: Robin is very grateful to David De la Haye, who has designed and built a digital timer in the same few days that we discovered and exploited the ST77 product. The reliability of these timers are critical to the bottle system, and so we are testing Davids' prototype, seen here drawing just 0.1 mA, and sat on Daves' excellent documentation. We will do these tests in parallel with our other tests and use of the ST77. This will be our "plan B" if we run into problems, or wish to add some "bells and whistles" in the future.

Timer to increase life 100x ?

bottle this is old stuff - to show you where we started...

On the right you can see the components we are now experimenting with to increase the duration from a few days to what we hope may be months or maybe even more. The idea is to introduce an electronic timer that switches the STEPP on for the required few minutes, every hour or few hours, to make its position report. The timer only takes about 1mA from the battery, compared with between 50 and 100mA when the STEPP is on. The result is a much lower average current drain on whatever battery is used. This approach may even permit use of one or more solar panels - if we are lucky they might keep the battery topped up forever ! We are using the standard Maplin VT27E timer kit (5 ?), but with the addition of two 10,000 uF capacitors to increase the adjustable time delay to one or two hours, and the switch on period to a few minutes. We are also using a reed relay able to operate on 6v or 9v since that in the kit requires 12v. The flasher unit is used so we can see when the timer switches on. We have not yet settled on the optimum battery pack, but we are using the "silly little" 9v PP3 seen here so that we can see how well the approach works.

first prototype ...

bottle prototype bottle floating The first prototype, seen here floating on the right, is based on a 2 litre plastic bottle (see construction details below), holding a Falcom STEPP unit with its combined GPS/GSM antenna, and one 6v lantern battery. Ballast, presently consisting of over 500gm of copper coins, keeps the bottle upright and with about 30% above the waterline. We've had "the bits" for this prototype of what might go in each bottle for some days now, and have tested these connected together (see "STEPP Testing Details" below). Layla has ordered a further two STEPP units, keeping some of her tight budget back for other alternatives as we identify and test them. We are now cofident that her budget will permit at least six bottles to be tracked with GPS/GSM, and this may be increased significantly if we get any further sponsorship. Right now we expect this prototype to provide continuous tracking for 8 days, at a rate limited only by our spend on text messages. In the coming days we expect to add a few small but important items like an on/off switch and a flashing light (timed to come on for a minute every 1/4 hour) to aid recovery after the sea trial, and also to investigate if the light can be seen from suitable viewpoints on land. It is possible we might also squeeze a second battery into the bottle, increasing the 8 days to 16. We will also be experimenting with power saving options, to further increase lifetime. What you see on the left is simply the components, with coin ballast, in the bottle being tested for balance in water. These compenents will be mounted within an inner, waterproof container. Laminated paper artwork can be placed inside the outer bottle, hiding the components and providing suitable information (in English, French, Dutch, Danish and Norwegian ?) for any lucky finder. However, we do not plan to leave this first prototype more than a few days in the sea, and hope to recover it within a few days.

bottle bits First sea trials might start within days: we need to checkout some basics: such as how well the GPS and mobile 'phone work when inside a bottle and only just above the waterline - and bobbing about like crazy ! :-) Lots of changes are expected, including the bottles themselves and what bits we put inside them. On the right, with a plastic coke bottle, is a Falcom STEPP with its combined GPS/GSM antenna on the left, and a rechargable 2000mAH NiCad battery pack. You can also see the "SIM card card" I got from Tescos. This combination works, providing tracking from any mobile 'phone (see testing section below) or GPSS PC connected to a GSM modem - as planned for the Exhibition. Testing does not even require a PC - just a mobile 'phone. A handheld GPS such as the Garmin etrex can be used to guide you to within yards of the last position obtained over the 'phone - could be useful for our sea trials - we do not want to lose the bottle ! :-) We estimate that a 2000mAH pack could keep the system working for 40 hours, nearly 2 days, based on the STEPP drawing 50mA. However, we hope to improve on this by reducing the STEPP current to perhaps nearer 20mA and using a bigger 14000mAH battery. So it is possible we might get nearer 700 hours or over 4 weeks.

latest tests - horizontal bottles ?

solar bottle The first bottle prototype ran for the equivalent of over 4 days continuously, which matches the 4.5 days calculated from 100mA consumption from a 11,000 mAH 6v lantern battery. A timer such as that above should increase this to nearer 5 or 6 weeks.

3L and 2L solar bottles We are now testing power saving features such as switching the STEPP unit on for only a few minutes each hour, and the use of solar cells. On the left you see a second prototype: a 3 litre bottle holding a solar panel and the electronic switch to only power the STEPP and a flasher for 5 minutes in every hour; a small 9v 170mAH battery to confirm that solar panel can charge faster than the system discharges. All mounted on a "plant solar turntable" so that the solar panel inside the bottle gets a typical dosage of sunlight each day. If we are very lucky, this system will prove the feasibility of solar panels giving a bottle unlimited operational life. The small 170mA battery would be replaced by a much larger capacity battery to avoid temporary loss of operation late at night or during prelonged periods of particularly gloomy weather. On the right you see the 3 litre bottle balanced to float horizontally to get more sunlight onto the solar cell. We have now managed to squeeze all this into the 2 litre size bottle, with just a 7.2v 2000mAH NiCad pack providing power and sufficient ballast for the bottle to float, keeping the solar panel up - as here below.

2L solar bottle floating We are moving towards bottles floating horizontal and with solar power a "definate maybe". No extra ballast seems to be needed, other than ensuring the battery is correctly positioned. Note that use of a solar panel means our label can only occupy half the bottle. i.e. A3 not A4 in size. Use of the timer gives a massive increase in bottle life, but it does mean we may need to make use of the STEPP "alarm" facility, to initiate the position report after power is switched on, since we probably cannot rely simply on the GSM system delivering a request in those few minutes the STEPP is switched on. However, there may be many ways of getting around this problem. The words below were put up a few days earlier...

We have started "duration tests" such as leaving a complete bottle system running outside the house, including on frosty nights (started 1800 on 2nd March). At first we thought this showed that a single 6v lantern battery may not suitable. However, later in the day we found this particular STEPP had been put into a power saving mode, and this is probably the cause of the problem. It seemed the type of battery we tested gives substantially more than the nominal 6v in its first few hours of use (Robin just measured over 8v on a new one). After an overnight test (outside in low temperatures) the STEPP began to "misbehave" (e.g. failing to reply, or reply without an updated GPS fix). Closer examination showed this battery still delivering 6.3v into the STEPP (Note the official lowest operating supply for the STEPP is 10v, so we should not be too surprised). Bringing the unit inside into the warm seems to fix the problem. i.e. it starts working again. We have added a single 1.5v A sized cell to the 6v lantern battery, bringing supply to 8.1v (at 1030 on 3rd). Unit was put in the 'fridge for a few minutes before taking outside. Duration tests continue with a second STEPP on the same single 6v lantern battery (restarted 0830 Thursday) and the first STEPP will be used for power saving mode experiments. Tests continue to see if this is the low temperatures making the supply voltage more critical, and reviewing what batteries should be used.

The Bottle

bottle glue You may think this question of the bottle trivial - compared with the more obvious technical details of what we put inside the bottle. But don't be fooled into complacency - poor attention to this detail could result in the bottle sinking, with all its' expensive content, within a few days at sea. More subtle points, beyond ingress of water, will be it floating upright, with the GPS and GSM antenna above the waterline, and it not building up a coat of salt or moisture that may attenuate the signals - maybe we will spray the outside with something ? Other important points will be strength, including avoiding cracks if dropped onto the deck before launch. We may also have to worry about unforseen events such as being pecked by hungry sea gulls ? :-)

But now to details of construction for those "Blue Peter" types among you: I've used a 2 litre plastic lemonade (or Coke) bottle, and another identical one to make things like the seal explained on the right here. Use a sharp craft knife to cut the bottle above what will be the waterline, but below where it starts to taper. Cut a piece from a spare (2nd) bottle, just above where it starts to taper and an inch or two below - to give a good overlap between the two halves of bottle #1. Apply gentle heat (a hair dryer?), some inches way from the cut top of bottle #1 lower part, to make it shrink just a fraction. This should enable you to push the cylinder from bottle #2 OVER it. You should also be able to push the TOP of bottle #1 OVER this seal made from bottle #2. After checking you've not made mistakes, glue the seal to the lower part of bottle #1 using sandpaper to roughen mating surfaces then "LOCTITE soft plastic adhesive" glue. Note that we've had "mixed results" with this and other glues on various plastic bottles. Any seals will probable need a "belts and braces" approach, such as addition sealing inside and out with water proof adhesive tape. e.g. parcel tape.

The first prototype above has the components such as battery, STEPP unit, ballast, etc., mounted within an inner bottle, which is itself expected to float. This was constructed from a split cylinder from another bottle, removing top and bottom. The side and bottom was sealed roughly with parcel tape, before pressing in ballast and strengthening using "polymorph" low melting point plastic purchased from Maplins. The intention is to produce a complete inner bottle, holding all the bits, capable of being easily fitted within the outer standard 2 litre plastic bottle. Suitable artwork would be placed between the two bottle surfaces with things like instructions for the finder in several languages including English, French, Dutch and Danish (emails have been sent to a few overseas friends for this). This special artwork for the bottle labels is where any logos of sponsors, such as soft drink makers, would appear.

I'll put up more details here as work on the prototype bottle proceeds. These will include any extra ballast needed to keep the bottle upright, but not sink. This will obviously depend on what we put inside the bottle. We want it to ride as high as possible out of the water, but to remain upright in all but the roughest conditions. For a stable bottle it looks like we need it to displace about 3 lbs or 1.7 kg of water. However, later tests may reveal that it is better to have less weight and let the bottle ride further out of the water, at an angle.

Note from Robin on 11 March: construction of later 2 litre size bottles showed we need not use a seal from a second bottle as illustrated above. Heating the rim of the lower half carefully to make it shrink, provides a tight push fit of the top half over the lower half. However, note also that a horizontal bottle puts the joint below the waterline and makes the waterproofing much more critical - lots of waterproof stickytape ? :-)

weights and measures ...

Now some numbers to do with weight: the battery will be the most signficant choice in our getting weeks rather than hours of bottle tracking. This, other than ballast, is by far the biggest and heaviest item and will be limited not so much by cost, as the size of bottle we use. So here are some rough weights:

  • total weight carried by a plastic bottle 3/4 below waterline: 1700gm.
  • weight of Falcom STEPP with its antenna: 200gm.
  • weight of a 9.6v 2000mAH battery pack: 500gm.
  • or alternatively, weight of a 12v 7000mAH battery pack: 1300gm.
  • or alternatively, weight of a 6v 11000mAH lantern battery: 610gm.
  • or alternatively, weight of two 6v 11000mAH lantern batteries: 1220gm.

Note that our first prototype weighs 1700 gm, including one 6v lantern battery and over 500 gm of ballast to keep the bottle upright.

We have been testing the STEPP running at 6v and it seems OK - more tests needed. We had been thinking of rechargeable 12v 7000mAH packs at 70 from but now see we could use common lantern batteries to give more capacity at lower cost. A 6v 11000mAH lantern battery supplying a STEPP drawing 55mA should last about 11000/55 = 200 hours, or over 8 days. Two batteries connected in parallel (+ to +, - to -) would give 22000mAH or 16 days (yes I chose 55mAH to make the calculation easier :-). Maybe we can extend this further by some combination of solar cells (see below) and/or timer circuit to switch the system on an off (it only need be on for a few minutes each hour, when it reports its position).

lantern battery Batteries and bottle should chosen so that the majority of the weight is at the bottom, so the bottle floats upright - giving less problem with keeping seals water-tight and the GPS/GSM antenna above the waterline. Unfortunately two 6v lantern batteries, one above the other, make a plastic coke bottle top-heavy. One lantern battery (see right) can be used with suitable lead ballast at the bottom of the bottle.

So right now we are assuming a 2 litre plastic lemonade bottle carrying the Falcom STEPP, a 6v 11000mAH lantern battery, and ballast. We will need to review these weights if we choose a stronger glass bottle or change the bits used.

Component Costs

Now some rough costs of "the bits" - remember that these do not include the time, travel expenses or profit of anyone putting each package together. But they should help to indicate how many bottles we might afford to track accurately, and where the major costs lie:

  • Falcom STEPP with antenna from 250 ?
  • Tesco SIM card 10 + 70 of top-up (4 positions/hour for 2 weeks?): 80 ?
  • 6v 11000mAH lantern battery (Maplin ZB65V): 4 ?
  • other items such as magnetic on/off switch, lights, etc. 30 ?

i.e. maybe close to 300 per bottle + 80 air time. Air time could be a significant cost. Each text from the bottle includes lat/lon position, speed, course, altitude (tells us when the bottle is thrown off the boat into the sea ?), time (GMT) and date. It is in the same standard NMEA format as the data from a GPS. That's the good news: the bad news is that each report may cost 10p, or maybe nearer 5p if the receiving phone is on the same network. Any phone asking for a report will also be charged a similar ammount for each request. However, the STEPP unit permits us to send a command telling the bottle to automatically report every so many seconds. e.g. 900 is every 15 minutes. If we use this figure, of 4 reports per hour, and we hope for 2 weeks of operation, then the air time cost might be 4x350x5p=70 - this is where we need and expect to reduce cost. But remember: we will need to send quite a few messages while we are testing, both at home/office and during those sea trials. It is just possible that a solar panel might give useful power in daylight hours, even if within the bottle and not getting direct sunlight.

Flashing Lights ?

cycle lights This pair of "Super bright cycle LED" cost me 15 from Halfords in Bracknell. One is red, the other green, and each holds two pen cell batteries, giving 3v. A button switches the light on, through a choice of flashing patterns, to off again. This might be a low weight and cost option for all the bottles, or just for those carrying the STEPP phone or similar device. One option is for the light to be switched on with a 'phone call to the STEPP when we want it to flash. e.g. for a few minutes after the hour in the evening, when observors on land can be looking out to sea for it. It may also be of value when recovering a lost bottle being tested in first sea trials. We have yet to see from how far we can see these lights, but I would be surprised if it is not several miles at dusk or night-time. Each light seems to draw an average of between 20 and 50mA, depending on what flash pattern is selected.

Another "flasher" for those who can use a soldering iron is a little kit from Maplin called "DUAL WHITE LED STROBOSCOPE" ( MK147 ) at about 7 and taking perhaps 1/2 hour to put together. It runs for a few hours off a small 9v PP3, but could last much longer running off a bigger battery, or only switched on occasionally. Right now we are experimenting with a Seiko electronic clock mechanism switching on the flasher(s) for a minute every 1/4 hour. Each bottle might have its clock a few minutes slow - so we might identify it by flash pattern and when it appears. The most practical aspect of the clock is being able to predict when a particular bottle with begin flashing, so we might locate it ourselves, without attracting too much attention from others at sea.

Solar Power ?

solar panel Products like this "Variable voltage charger panel", code RR53H, at 13 from might be used (if Layla doesn't mind a wierd looking solar panel on a stick over the bottle) to allow tracking of the bottles (or those we fit this option to) for an unlimited time - maybe months or even years. It is possible we might fit one or two INSIDE the bottle, and they might provide enough power from the indirect sunlight received. Described on the box as "2,6,9 and 12 volt solar panel with AA and 9 volt battery charger", I'm using it for some quick tests to see if such a product might be of use to extend battery life significantly. If we are lucky it might give unlimited life to the bottles, but it is also possible we would do better spending money on a bigger battery :-) However, a first, very simple test outside on a dull day, showed this solar panel charging 10mA into a 12v battery, so it is well worth further consideration.

what's all this mAH stuff ?

batteries For the less technical reader, here are some basic facts of importance when we are choosing batteries and what bits to use in each bottle. Battery capacity, in terms of what current it can supply for how long, is measured in ampere-hours (AH) or milliamp-hours (mAH). 1000mAH = 1AH. Current drawn by something connected to the battery is measured in Amps (A) or milliamps (mA). So a big 12 volt car battery might supply 40AH (or 40,000 mAH). A 6v lantern battery (rightmost of three on right) might give 11,000mAH and the small AA sized pen cell batteries (leftmost) might supply only 200mAH. The biggest "A" size cells like you might find in your torch (middle) might provide nearer 2000mAH or even more. Different battery types will vary a lot, so these are just rough figures.

So how can we use this information ? Well, if that cycle flasher above is pulling an average of 20mA, and we connect it to a 2000mAH battery, it should run for 2000/20 = 100 hours, or nearly 5 days. If we switch it on for a few minutes when we need it, it will last for years. Unfortunately the sizes of bottle we are considering will not support a 12v car battery, and so we will have to make do with smaller, less powerful batteries :-)

GPS 'phones such as those from Benefon and Thuraya

HNS-7100 How about the idea of a mobile 'phone inside the bottle, regulary reporting its' position, and available for immediate voice use by whoever finds and breaks open the bottle ? Something similar was done by Coca Cola in Australia last year - but they simply hid the 'phone in a specially constructed bottle - and it was switched off. For this they used a Benefon GPS 'phone - one of the first GPS 'phones put on the market a few years back. If you visit the Thuraya page you will see me testing the Thuraya GPS 'phone made by Ascom and Hughes Network Systems. This is more than just a GPS 'phone using the GSM network since it also includes a Thuraya satellite tranceiver and antenna. So if you are somewhere like the middle of the north sea, outside GSM coverage, you can still make 'phone calls or send and receive SMS messages. Iridium, operated by Boeing Aerospace is another example of a satellite 'phone system. Right now we are interested in any suitable GPS 'phone products which can continue to operate for weeks rather than days (e.g. by only switching on for a few minutes each day as is possible with the Benefon Track One). I expect to be contacting all the obvious mobile 'phone makers including Nokia and Ericsson.

STEPP Testing Details

phone + GPS Here are a few details for those like Layla who may want to start some real tests with our prototype now. They will need to be given the 'phone number of the SIM card inside the STEPP unit, and Robin will need to ensure the unit is powered on when the tests are done.

You need not bother about these details if the tracking is being done automatically on a PC running GPSS, the GPS Software. However, the fact that we only need any mobile 'phone to do it means that real trials, including at sea, can be done very simply.

Asking the STEPP to give you a GPS report: send it the text "&REQ POS"
(if you sent "&REQ POS 15" it would then send you a report every 15 minutes - but this is NOT reccomended until we are using the computer for this. Sending "&REQ POS 0" cancels any request, by anyone, for a regular report)

What reply does the STEPP return ? You should soon get a text message back (if the STEPP is switched on, and in a suitable position for its GPS and 'phone to work) holding two NMEA messages, $GPGGA then $GPRMC.

You may recognize things like lat/lon above, but this may help: this was the reply I got at 09:46 and 41 seconds, on 18 Freb 2004, from the STEPP here. The Lat/Lon was 51 degrees 23.8117 minutes North, 0 degrees 39.6179 minutes West. Lat/Lon is in WGS84, not the OSGB you will see printed around the edge of Ordnance Survey maps.

So any mobile phone, and knowledge of the bottle 'phone number, enables you to find out the position of the bottle in lat/lon which you could then look up on a suitable map. However, if you also have a handheld GPS such as the Garmin etrex, you could type this lat/lon into the GPS then ask it to guide you to the position. The GPS will then display the updated distance and direction to the position, to within a few yards. This could be invaluable for testing on land or at sea. But for sea I also reccomend having the bottle tethered on a long piece of string ;-)

The lat/lon can be abbreviated for typing into GPSS running on a PC which does not have a GPS or GSM modem as:
- followed by the Enter key then V for Visit. See TIPS page for details.

For a PC equipped with a GSM modem (or 'phone that includes a modem), a program called SMSH is used to poll the bottles and take their incoming positions and put them on the displayed map. The picture below shows SMSH (which is normally hidden, and only displayed when testing) taking the incoming position report from the STEPP unit. The unit was on my front lawn, marked by the square icon, and the reported position, marked by the circular icon, was 15 metres in error - which rapidly changed to only 3 metres when I realised I'd put the STEPP antenna upside down on the grass - it was amazing that it had worked at all :-)


Don't worry about the details of using this software, since Robin will probably do it for you. If you do get into GPSS use, you will need to have a Laptop PC and GPS with cable (such as the Garmin etrex) and start with step 1 followed by step 2 on the DOWNLOAD page. It is always best to take it in simple steps.

The bottle label and sponsorship

Here is a typical bottle label, for illustration only. It probably has little similarity with the labels Layla will design for the actual project. However, it does serve as a prototype to see what we might show to any "bottle finder", with important information such as assuring them that the bottle is harmless. Full contact details of whoever is taking responsability may not be such a bad idea.

An A4 sheet of paper will wrap nicely around the inner bottle and can be inside the outer bottle, protected from the sea. Further protection is provided if the paper label is laminated. The lower part might include space to acknowledge sponsorship, such as those who decide to "sponsor a bottle" and pay for the cost of the components within it. Typical sponsors might include soft drink manufacturers who already distribute their drinks in these 2 litre (or 3 litre) bottles. These include Coca Cola, Pepsi Cola, 7-Up, Sprite, Fanta, Schwepps, Dr Pepper, Virgin, R Whites, Tizer, Tango, Deeside, Buxton, Sainsbury and Tesco. Other sponsors we expect to contact include the mobile 'phone product and network players.

bottle label