Early Marine Seismics in Cambridge – Review
Bullard |Archive copyright 2015
The use of seismic techniques to investigate subsurface geology on land goes back to the 1920s in America where it was used for oil exploration, but it was not until 1936 that the attempts were made to use seismic investigations at sea. In Cambridge the first land seismics was done by Mr Edward Bullard and Tom Gaskell, Kier Grant and W.B.Harland in East Anglia from 1936 to 1938. The first marine seismics was undertaken by Professor Maurice Ewing of Princeton University (or was it before he went to Princeton ?) in about 1935, to whom credit is due for the taking the first steps into the unknown – at this time the no-one knew if hydrophones in the water would pick up seismic waves, or whether explosions near the surface could be substituted for charges placed on or in the seabed, so the initial course of action had to be to mimic land experiments as closely as possible by putting geophones and explosions directly on the seabed. Cambridge got in on the act pretty quickly! In 1937 Dr E.C.Bullard (Teddy, later Sir Edward Bullard) was invited to America by Professor R.M. Field, chairman of the American Geophysical Union for the Geophysical Study of Ocean Basins and went to sea with Ewing to witness the preliminary tests of the technique.
Bullard was impressed by what he saw and decided that this was a field that he wanted to pursue in Britain, where the relation of the continent to the continental shelf is different from that seen on the Eastern seaboard of America. It was seen as fitting well with the gravity studies of Vening Meinesz which B.C. Browne was interested in extending with a gravity survey undertaken on the continental shelf in a submarine, with Vening Meinesz’s borrowed pendulum gravimeter which was thought to be more suitable for work at sea than the Cambridge Pendulum Gravity meters as it was self contained whereas the Cambridge system used a number of separate units. The National Committee for Geodesy and Geophysics recommended the joint project to the Royal Society who vouchsafed funding and the Royal Navy offered the use of a submarine for the gravity work and a ship for the seismic work, plus surplus explosives, all in less time than it would nowadays take to write a simple grant proposal, let alone get it approved, all supervised by a committee of Admirals.
The experiment started with what was known and understood to work on land, but extended into a marine environment;- modified land geophones in a waterproof housing lowered to the bottom, and recorded on board ship on the film recorder with a 100 Hz tuning fork drive that had been developed by Leslie Flavill for the East Anglian land work. The source was military TNT charges electrically fired on the sea bed, and the experiment was constrained to shallow water because of the equipment and method used. The essence of any seismics is knowing the travel times for the waves from source to receiver, and whilst on land this could be achieved simply by a long wire from source to recorder, at sea it is not practicable to do this, and a radio link had to be used.
All the new equipment was developed by Leslie Flavill in a very short time and tested in lake Windemere in 1937, followed in 1938 by an experiment on the Royal Navy vessel HMS Jason, during which shots were fired from Jason’s launch ‘The Stork‘, with the shot break signal sent by radio to Jason. Initially a rather unreliable shot break signal was generated by transmitting a continuous tone from The Stork and relying on the shaking of the antenna on The Stork to generate an observable signal, but this was quickly replaced by a radio signal switched from a wire wrapped round the charge. The navigational accuracy of the day was inadequate to measure out the required shot-receiver distance, and the experiment had to rely on sextant observations of Jason’s mast height from the launch to position the shots! Problems were also experienced during data processing in measuring the water wave travel time as the recording relied on the seabed geophones of limited bandwidth, and at the short ranges used the water wave arrived in the midst of earlier ground phases.
The following year saw a further experiment aimed at extending the 1938 line in the English Channel using a working Brixham trawler, The Renown, as the shooting ship and another trawler converted to a gentleman’s yacht, the Arthur Rogers, lent by the owner Mr Byng and crewed by him and his friends as the recording ship. These ships had scientists Ben Browne and Tom Gaskell on the Renown and Leslie Flavill and Teddy Bullard on the Arthur Rogers. New technology for this cruise included a hydrophone hung 20 ft deep off the Arthur Rogers to record the water wave, and a new design of geophone that worked whichever way up it landed. This geophone, either by design or more likely serendipity, followed principles of design for optimised coupling to the seabed not fully appreciated or applied again for the next 50 years. Additional attenuated geophone channels were added to give a greater dynamic range for larger shots – on this cruise blasting gelignite charges up to 160lbs.
These shallow water experiments using bottom charges and geophones on the bottom came to an abrupt end with the start of World War II – indeed the Jason experiment only just crept under the wire, and the (almost) simultaneous gravity survey on the submarine HMS Narwhal was cut short at 9 p.m. on 13th September when the sub was urgently recalled to port after the international situation deteriorated (the Munich crisis). They had, however, by this time proved that crustal layers and velocities could be identified on the continental shelf.
It was some time after the war before things settled down, and Teddy Bullard and Leslie Flavill returned from war work, and the department got back to considering how to get round the need for two ships, and the modifications necessary to conduct seismic experiments in deep water. Military investigations during the war had established that sound waves could be detected on hydrophones suspended in the water column as readily as on geophones sitting on the seabed and in 1946 Maurice Hill and Pat Wilmore compared both types of detector off Plymouth in a Marine Biolological Association of Plymouth ship, M.V.Sabella and confirmed these conclusions. These tests used detectors deployed from the ship, used together with various schemes for firing charges suspended in the water column from buoys at a distance from the ship, including clockwork timers and long wires with intermediate floats to avoid the need for a second ship for shot firing. The danger and impracticality of handling charges in this way led to the inevitable conclusion that the only safe course of action was to keep the charges under control of the single ship, and find a way to deploy the detectors at a distance if it was to be possible to work with only one ship. As part of the initial work with suspended hydrophones Hill and Wilmore worked out a way of decoupling the hydrophone from the motion of the sea surface or ship, thus vastly reducing spurious noise. This opened the possibility of suspending hydrophones beneath free floating buoys, which had the additional advantage that several well spaced detectors could be deployed, although of course it added to the complexity of data localisation and instrument recovery.
October 1947 saw the first use of new sono-radio buoys developed by Maurice Hill using the experience gathered the year before to extend the work in the English Channel south of Plymouth. By May 1948 Hill had redesigned the radio equipment and hydrophones and added two more stations, returning in August to shoot three more stations. Data resolved a basement velocity of 14500 ft per second (4420M/s) deepening rapidly to 3300 ft (1000m) 25 miles south of Eddystone.
In 1949 the department undertook its first refraction shooting in deep water – using the weather ship Weather Explorer on a regular trip, 4 buoys were deployed with hydrophones suspended with neutrally buoyant horizontal sections 150 ft (46m) . below the surface, in depths of 7776 ft (2370m), with refraction shots using depth charges set to fire at 900ft (275m)depth, with additional 1 ¼ lb (500gm) TNT charges for obtaining reflection data. This experiment revealed the sea in the area to be “a rather complex structure”.
On the 1st May 1950 HMS “Challenger” sailed from Plymouth on a 3 year world surveying cruise with a set of ‘Hill Buoys’ on board. Over the next three years 50 refraction lines were completed by John Swallow and Tom Gaskell with Maurice Hill and John Cleverly joining the team at various stages of the cruise.
Novel Ways of Deployment: In the very shallow waters over the Seychelles Bank and atolls of Funafuti and Nukufetau a different method of deployment was adopted. The recording was done on the ship moored at anchor as normal but the buoys were towed away using the ship’s motor boat and buoys moored in position. The motor boat was then used to electrically fire the charges and shot instant transmitted back to the ship for recording. By using the H.T. of the motor boat transmitter to fire the charges and heterodyning the received carrier to produce an audible note a clear firing mark was obtained. At one station, on a peak of the Mid-Atlantic ridge, three buoys were moored in 100 fathoms (183m) of water using piano-wire and sinker weights. The buoys remained in position but with each mooring taking 40 minutes and the short operating duration of the buoy this method was limited to fairly shallow waters.
In 1954 a new set of buoys were produced primarily for short range shallow water work and weighing a 1/5th of the weight of the old buoys at just 50lbs (22.7Kg) were much easier to handle. Using dry batteries giving a working deployment life of 24hours and a radiated power of 3watts a working range of 12 – 16 miles was achieved but at this range it is not possible to detect the deeper 8.1km/sec layer.
In 1962, in order to enable refraction shooting at ranges sufficient to see the deeper layers, Tim Francis developed a recording sono-radio buoy, that had an internal photographic recorder – the Woodhill galvanometer camera originally developed for the Atomic Weapons Research Establishment as well as the established radio transmitters for short range work. The buoys also had a radio receiver used to command the recorder to switch on and off. By this date the electronics had all been transisorised, with the exception of the amplifier for the barium titenate hydrophone, which required a higher impedance than could be obtained by transistors, and so relied on a miniature thermionic valve.
the Woodhill Galvanometer Recorder
In 1963 John Jones sailed on Chain with Lamont Geological Institute to look at their seismic profiling system, and by February 1965 had a 30 cu. in. free running air gun based on the Lamont design and shallow streamer working for the Discovery cruise to Madeira. At this stage the air gun and array were run at shallow depths, with mostly high frequencies forming the image. In 19XX it was realised that running the air gun at ¼ wavelength deep used the reflected wave at the surface to reinforce the downgoing energy and increased the low frequency content – Tim Owen wrote a program to calculate the towing cable catenary that led to the construction of a heavy mount for the airgun, with the two air hoses separate from the towing wire and left to stream in a bight. In 19XX Fred Gray went to visit the Institute Francaise du Petrole in Brest(?), which led to the department purchasing a single section of the Geomechanique ‘Flexotir’ array with much reduced noise due to very careful static buoyancy adjustment and an elastic tow section, and towing at ¼ of the predominant signal frequency. In 19XX, after unsuccesful attempts to get the air gun to run with a solenoid firing system, a Bolt electrically fired gun was purchased. Records from the array were recorded on a Thermionics 16 channel 1 inch analogue tape recorder using F.M. modulation that was the size of a small wardrobe.
Tony Mertzer and the analogue tape recorder
Data display, both on-line and for replay was initially via a modified Muirhead ‘MUFAX’ weather chart recorder that used wet chemical paper passing between a rotating helix and a straight edge that went brown on the passage of a current. The basic MUFAX was a synchronous device that was used in research ships for recording the output of deep water echo sounders but in order to use it to record the free running Lamont type air gun Leslie Flavill modified it to include a solenoid operated clutch that was triggered by the signal generated from a trigger hydrophone towed behind the ship. Two smaller Mufaxes were purchased in 19XX – named Muriel and Murgatroid – and similarly modified.
Seismic Jet Pen,
The problem with the Mufax recorders was that they relied on a chemical process (with iodine) to mark the paper, and the process depended on the paper being wet enough to conduct electricity so a lot of effort had to be put into keeping the paper damp in a wide range of environments. Even when working well each trace was only a thin line that went black when there was a signal and didn’t mark where there was no signal – although there was some variation in intensity with signal amplitude, the display gave almost no amplitude information. Before Leslie Flavill died in 1971, he and Tim had been discussing the use of the pens from the Siemens Jet Pen recorder as the basis for a scanning seismic data display and after his death Tim went on to develop the idea through a couple of prototypes and a final version. The Siemens Jet Pen recorders were an ingenious device for the display of data up to several Khz consisting of a galvanometer made of a fine glass tube fitted through a small ring magnet and bent at right angles and drawn out to a very fine tube. The main glass tube acted as a torsion suspension, the ring magnet sat in the magnetic field of the drive coil and the tube could be twisted by the signal applied to the drive coil. The right angled tip produced a very fine jet of ink that scanned from side to side as the field deflected the magnet. By mounting the Jet Pen in a moving carriage a scanned line could be produced with the jet deflection normal to the scan direction. By moving the paper between each scan a full profile could be built up with considerably greater dynamic range than was available from the wet paper recorders. A prototype was built using stepper motors to drive paper and pen, initially with a clutch and spring to give the ‘flyback’ at the end of each scan, but later using the stepper motor drive in both directions. The first ‘Seismic Jet Pen’ worked well but the rapid starting and stopping of the carriage meant that sometimes the carriage got out of synchronisation with the timebase. In XXXX we began a major project to build a final version that incorporated a digital acceleration and deceleration circuit designed by Mike McCormack using dozens of I/C logic chips, and a comprehensive control and analogue signal display architecture. The building of the ‘Seismic Jet Pen’ was a major lab project that involved Mike, Mel and Roger Theobald, and produced a versatile and extremely useful single channel display system for seismic reflection profiling and disposable sono-buoys. A copy of the final version was also made for the Institute of Oceanographic Sciences.
At about the same time that we thinking about development of the Seismic Jet pen Mel and Tim put on a display at the first Oceanology International exhibition at the Metropole Hotel in Brighton, on a stand kindly donated by the organisers.
The next development in 1966 Cambridge sonobuoys was the purchase of a new sonobuoy system from G&E Bradley of London that had been developed for Birmingham University. The prototype sonobuoy was originally delivered in a plastic barrel buoy, but we had significant problems with grounding on a test cruise on the French CGG ship Andromede in the English Channel that caused us to revert to our own design of metal cans. The Bradley system used crystal controlled 27 MHz transmitters on the buoys with the seismic signal f.m. modulated onto a 3.375 Hz sub-carrier that was amplitude modulated onto the R.F. Initially the Bradley gear was used in sonobuoys without recorders, but after the film recorders were lost in the Red Sea new buoys with modified UHER recorders (below right) were built using a similar f.m. modulation.
Bradley system in a Cambridge buoy with the ship-board equipment
Sea Bed instrumentation;-
By the end of the 1950s it had been realised by Maurice Hill and Raitt that imaging of the shallow layers of the crust was made difficult because of the complications introduced by the water layer delay which caused masking of early arrivals, and that it would be necessary to put the receiver and source on the seabed to overcome this problem. It fell to John Shorthouse, a new PhD student, to work out how to do this in Cambridge. At the time John Shorthouse started his PhD in 1961 there had been a number of seismic experiments in shallow water in America () and England (), using sensors cabled back to recording systems in the ship, and bottom charges, and a couple of experiments in America using a 4 channel photographic recorder lowered by cable to the seabed in 1937 and 1938. There had also been an American experimental true autonomous deep water pop-up seismic apparatus that used 30 gallons of petrol as the buoyancy, tested in 1939 and used in 1940 – it was not a great success, with substantial losses. John Shorthouse was wary of the difficulty of locating the popped-up apparatus on the surface – the pioneer pop-up had used a series of floats on the apparatus that were release at 15 minute intervals during its deployment on the bottom to help keep track of where the equipment might be expected to pop up – and so opted for a recording system tethered to the ship. Since Cambridge research normally had access to only one ship, using the ship to mind the recording system at the time of the actual experiment meant that the shots had to be fired on the sea bed independently of the ship, and this was done by putting the charge in a pressure proof case, with a separate pressure case containing a clockwork timer for fire the charge, and putting the charges in steel pressure vessels down with their timers running before deploying the recording package.
Shorthouse’s equipment consisted of a 10 channel Woodhill photographic recorder (originally designed for the Admiralty Research Establishment) in an aluminium (?) pressure vessel connected to a string of 4 geophones spaced at 100m intervals, with a geophone later installed within the pressure tube ( although more or less all it did was to pick up motor noise from the recorder!).
The timing of the experiment was complex;- once started by the clockwork timer, the recorder would run for just 19 minutes, allowing 6 shots at approximately 2 ½ minute intervals, each shot being timed independently by a similar clockwork timer. The record timing was done with a Smiths electric car clock with an accuracy of 1 part in 1000. Later modifications added 40 Hz low pass filters to the hydrophone channels, and a rectified wideband hydrophone channel that was able to receive the pings from the PES ( precision Echo Sounder) which acted as a crude check on the internal timing. Filtering of the internal geophone channel more or less eliminated the motor noise problem.
In 1964(?) Bob Whitmarsh started a PhD. With the intention of developing the deep water sea bottom seismic capability, and initially used Shorthouse’s equipment modified to have a 200 ft film magazine to give greater recording speed and duration, and a rectified channel on each hydrophone on a Discovery cruise in 1965. Although the experiment yielded better data than before, Whitmarsh decided to develop the sea bed seismic capability using a proper autonomous pop-up system.
The challenges facing Whitmarsh were formidable – a system relying on petrol for buoyancy as did the 1940 American experiment carries enormous potential risk on board ship, and metal housings could not be made buoyant at the depths envisaged for the equipment. The only available alternative was the newly available glass spheres, which claimed to be suitable for depths down to 6000m and had, at that, stage no history of successful use in field conditions. The spheres that were available were 17 inches diameter, and thus too small to take the Woodhill recorders,so an alternative recording medium was needed. Transistors were taking over from valves and allowed simple digital clock circuits, although integrated circuits were still strange devices, and it was still necessary to use miniature valves to connect hydrophones as high impedance semiconductor devices such as field effect transistors didn’t exist.
Bob’s efforts to build a pop-up seismometer were another feat of endurance in what was rapidly becoming a traditional Cambridge seabed struggle! He originally hoped that he would find a manufacturer of domestic tape recorders who would help to modify their recorder to be suitable for use in his Seabed Seismic Recorder (SSR)by extending the recording time and reducing the power consumption, but in the end had to settle for very high tec tape recorders developed for use in rocket testing at a cost of £2450 each – at today’s prices that amounts to about $50,000 each. Things didn’t go well on the glass sphere front either – the intention was, as in later OBSs, to use the sphere as both an instrument housing and the main buoyancy for the SSR, but despite efforts to re-grind the mating surfaces, it proved impossible to prevent leaks at any depth greater than 200m, so hopes of a deep water instrument had to be shelved.
The equipment was finally built and tested on the MAFF fishing boat Platessa from Lowestoft and deployed in the English Channel from the CGG ship Andromede, with limited success:
“Unfortunately the useful data recovered from the records of the Andromede station are fewer than they might have been” ( 141)
Bob ended his time at Cambridge on an upbeat note, however;-
“In conclusion, it appears that ocean bottom seismic recorders can become useful tools in future seismic exploration of the ocean crust and upper mantle. It is hoped that the progress made so far with these instruments will not end but only mark the beginning of the end….
While Bob had been working in Cambridge, Tim Francis had been working at Blacknest (an offshoot of the Atomic Weapons Research Establishment at Aldermaston) on the development of an autonomous ocean bottom recorder for use in monitoring underground nuclear explosions during the cold war of the 1960s and 70s. Bob then went to the Institute of Oceanographic Sciences and continued development of OBSs using the forged aluminium spheres developed for Tim Francis. Cambridge then concentrated on internally recording sonobuoys and reflection profiling until 1975.
Our next sortie into seabed instrumentation followed on from the LISPB experiment to determine the crustal structure of Great Britain by shooting several lines on land. It was clear that to make a similar profile of the structure of the North Sea would require the use of marine seismic equipment, and that there were distinct advantages to using sea bed recorders. Alan Smith had taken part in the LISPB experiment as part of his PhD project and set about using elements of the LISPB recordinag system as the basis for a new seabed instrument. Given the limited depths that would be encountered in the North Sea, it was though to be unnecessary to plan for a full pop-up system and that instruments could be moored from surface buoys. Learning from past problems previously encountered with glass spheres, Alan decided to build the new instruments – christened PUSS for Pull-Up Shallow Seismometer – in aluminium tubes. The limited diameter meant that the high quality UHER tape recorders used in the LISPB equipment could not be used, and small cassette recorders were substituted, with space in the tube to fit 4 decks. The impossibility of receiving radio signals on the seabed meant that a crystal controlled clock had to be provided. The main sensor for the PUSSES was a 3 component geophone in a gimballed mount that was designed to be ‘sticky’ by the addition of a controlled level of sticktion/friction so that the geophone aligned vertically and horizontally, but had sufficient sticktion to remain rigid at stress levels corresponding to seismic signals. A hydrophone was a later addition. The 3 geophone channels, the hydrophone channel and the clock and flutter correction channels were recorded on the 4 tracks of the casssette deck by combining the flutter and clock channels.
The Pusses were first used in May 1976 in the North Sea as part of Alan Smith’s PhD, and then taken over by Phil Christie to continue the North Sea work. Modifications included changing the recording windows to allow 4 dispersed charges with a further 4 back-up windows. The crystal clock was also upgraded to improve temperature stability, as was the power supply. One of the problems with the old PUSSES was the noise and rumble on the geophone from the tape drive motors, and new cassette decks were sourced that gave less noise, but in order to reduce all units to an acceptable level it was necessary to remove the flywheels and have them dynamically balanced.
The first PUSS experiment was characterised by periods of high frequency noise on all three geophone channels that appeared to have a physical origin – tugs, settling or something similar, and this was tackled by using rubber bushes and rubber sleeving on wires near the PUSS.
The North Sea experiments of Phil Christie in 1977 used 13 PUSSES, 5 upgraded as above from the previous generation and eight new ones, and involved RRS John Murray and RRS Shackleton. The experiment went well with the Firth of Forth profile and a Crustal Control profile being successfully completed, in that only one of the three dispersed charges failed to detonate, and only one of the 13 PUSSES was lost, and gave the first experimental value for the stretching of the North Sea basin.
Penny Barton was the next student to take over the PUSSES, for a long line in the North Sea across the Central Graben. The instruments were used as modified by Phil Christie. The first attempt to shoot the line in 1980 was limited by bad weather, so a continuation into 1881 was planned. One problem that had been observed on the 1980 line was interference from commercial seismic shooting at up to 200km range, which was then extensive in the North Sea, and Penny Barton tackled this by liaising with the exploration companies for short windows over the programmed shot windows. The planned line across the Central Graben was successfully completed in 1981.
The next seabed seismic experiments by Cambridge started in 1978 when Martin Sinha began a PhD on seismic refraction, and sought to use seabed receivers in deep water in order to increase the range window over which layer 2 arrivals could be observed, and to avoid the difficulties of defining receiver positions when using free floating sonobuoys, and to move towards being able to record both p and s waves. The F.M. recording sonobuoy system already developed in the lab was suitable for use in a seabed instrument, leaving Martin to design the pressure housing, release and recovery system and a suitable deep water hydrophone. As with the PUSSES, commercially available aluminium pressure tubes were chosen for the pressure housing, the most suitable type to accommodate the existing sonobuoy electronic circuits being 153mm internal diameter and 25.4mm wall thickness tube of the widely available, ‘normal’ HT30 alloy which gives a failure depth of 6300m and a safe working depth of 4000m, just about good enough for the proposed experiments. Flat endcaps were used, with a tube length of 1.22m. Buoyancy was provided by 4 glass spheres in a polypropylene tube and two blocks of syntactic foam and the release and acoustic command system was that developed by the Institute of Oceanographic Sciences to operate around 10Khz.
A prototype was tested in March 1979, but ‘unfortunately, it failed to record data on any of these stations’. A further test in September – ‘unfortunately due to an airgun failure during the first recording window no shots were fired at a suitable range for recording refracted arrivals‘, however, later on the same cruise good data was at last obtained. A further comment from Martin’s thesis is illuminating – ‘The major factor in the time required to lay the OBH is the need for a satellite fix to accurately locate its position on the bottom’. In spite of the setbacks, this OBH seabed Hydrophone was successful and justified its cost of £2740 each, including £1500 for the IOS acoustic release system.
Martin Sinha was followed by Jeremy Duschenes, who started from the premise that it was desirable to record full ground motion as well as the pressure signal, and that instruments with geophones integral to the main housing were unlikely to couple effectively to the seabed so as to accurately record true ground motion. Having worked on the Massachusetts Institute of Technology
OBS in the US, Jeremy was able to build on the work he had done there, and started with a well tested geophone design by Ray Davis of USGS/Woods Hole that he was able to marry to Martin Sinha’s OBH design to produce a new OBS. This geophone was a heavy cylinder 300 mm long by 230 mm diameter. In Duschenes system it was deployed on its side by an ingenious mechanism using a 1.8m pole as part of the link between disposable anchor and instrument tube, configured with a removable pin so that when the anchor hit the seabed, the geophone pulled the pole almost horizontal, which withdrew the pin and deposited the geophone on the seabed some 1.5m from the anchor anchor weight, after which the buoyancy of the instrument housing pulled the pole vertical again. Since the geophone itself weighed 25Kg in water it was not feasible to provide sufficient buoyancy to bring it to the surface, so a cutter was designed as part of the release link for the anchor weight that severed the geophone electrical connections to the instrument housing. One comment he made indicates the state of knowledge in relation to geophones – ‘In effect this meant that the shape and weight of a the geophone housing was far less important than its size for achieving good coupling’ – an error that persisted for some time and took some dramatic data to refute.
Initial trials were carried out in Biscay Nov 1980 aboard RRS John Murray – although it was planned to have all 5 instruments ready for testing, ultimately only one was finished. Ultimately a total of 29 deployments were made, with good data recorded on 21 of them, giving a 72% success rate – not bad as these things go.
One interesting observation made was that there was very good correlation between the vertical geophone and the hydrophone records, and Jeremy states after discussing these data in detail that ‘It is taken that the horizontal geophones, by implication, for which there is no independent test, are also recording ground velocity with fidelity”. He then goes on to speculate that what we now know was almost certainly rocking of the cylindrical geophone on the elastic seabed was in fact P to S converted energy. Duschene’s work
Around 1980 major changes in Cambridge Marine technology occurred; – the sonobuoy systems were abandoned in favour of seabed seismometers, and research student led instrument development of seismic systems technology gave way to longer term development by lab staff. By this time Ocean Bottom Seismometers – OBSs – had become almost as reliable as sonobuoys but with much greater receiver location accuracy, allowing higher spacial resolution surveys. The lab staff had for many years been developing electronics for seismic systems on a modular basis, and this had enabled students to put recording systems together without detailed technical knowledge, but after the Sinha OBH system the lab began a program of continuous development of OBS instruments.
From 1980 the lab was known as the Bullard laboratories after Teddy Bullard, following the amalgamation of the departments of Geodesy and Geophysics, Geology and Mineralogy and Petrology into the Department of Earth Sciences.
In 1981 a new generation of analogue OBSs was developed with microprocessor controlled timing and clock, still using four analogue cassette recorder modules for data storage, with an IOS pyro release, HE 30 aluminium pressure tubes with a 4000m depth rating and glass spheres for flotation. The geophone was a large aluminium cylinder containing a gimballed 3 component array of 4.5 Hz geophones immersed in thick oil for damping and to couple horizontal motion to the sensors. The geophone package was a massive 230mm diameter by 300 mm long. The complex deployment arm used by Duschenes was abandoned, and several simpler systems were tried. During a Martin Sinha experiment in the Atlantic on old/new crust comparison several of these geophones were deployed on end and several on their sides. The horizontal arrivals on the two differently deployed sets of geophones were of markedly different frequencies, indicating strongly that the frequency of the recorded signals were likely to be a function of the way that the geophone coupled to the ground, and indeed might bear little relation to the true ground motion. This revalation led to a number of lab based experiments and theoretical analyses of horizontal coupling, mostly undertaken by Part 3 Physics undergraduates, which led to the abandonment of gimballed sensors and the design of new geophone packages based on a better understanding of the coupling of sensor packages to soft, elastic sediments. This generation was the ‘standard’ Department marine seismic refraction system for a number of years.
Although we had been using microprocessors for programable recording window timing inOBH/Ss for a few years, this wasn’t our first microprocessor instrument. In 1978 Tim Owen and Ian Hutchinson developed a digital heat flow recorder using the newly developed RCA 1802 CMOS microprocessor with 256 Bytes (yes, 256 Bytes!) of low power static ram for both program, monitor and a couple of variables, with digital recording on cassette.
Our first digital recording system for seismics was developed around 1985, based on an RCA 1805 processor using an 8 bit analogue to digital converter with automatic binary gain ranging, giving 16 bits of effective dynamic range for 8 bit recorded data plus gain change flags. Data was recorded as MFM digital coding onto audio channels on high quality Sony Walkman professional cassette recorders at 9600 baud. The resulting system, known as the DSR ( Digital Seismic Recorder) was used both in land instruments and in a new generation of OBH/OBS instruments known as DOBS – digital Ocean Bottom Seismometers. The DOBS were built in tubes of high strength aluminium alloy – Hidraminium 48 – that enabled them to be deployed down to 6000m, and used the same IOS pyrotechnic release as the previous generation of instruments. The development of these instruments coincided with the switch from explosive sources to air gun arrays, so that it became necessary to record many more shots at frequent intervals, and it was therefore no longer possible to use discreet programmed shot windows, one for each shot, as had been done for explosive shooting. The requirement for continuous windows meant that much longer data spans had to be stored on the stereo cassette recorders fitted into the OBSs, meaning that it was no longer feasible to routinely record four channel data, which was not, in general, viewed as a major setback since very little use had been made of horizontal components of past data. Since the data handling and recording were all digital the instruments were designed to be programmed to record any number of channels up to 4. The data writing system had four cassette recorders which wrote on two channels in 24KByte blocks (12K per channel) in a stop- start mode taking 11 seconds to write a block that took 24 seconds to record 4 channel data, or 128 seconds to record single channel data. Overall recording time was therefore about 96 minutes per recorder in 4 channel mode, or about 6 hours per recorder in single channel mode, giving a full 24 hours of continuous single channel air gun data at a sample rate of 256 samples per second. DOBS was used from about 1986 until the arrival of the miniDOBS, although in its later years it suffered from trouble decoding the MFM encoded signals for reasons that were never understood, but were thought to be related to changes in tape characteristics.
It was clear by the time the DOBS were getting towards the end of their life that the experimental requirement was increasingly towards larger arrays of sea bottom seismometers and, following oil exploration practice, towards 3 D surveys. As a consequence cumbersome and relatively expensive instruments like the DOBS that were very labour intensive to make did not seem a very sensible model for the next generation, and while many OBS design teams, particularly in the US were elaborating their design ideas towards bigger and more complex instruments with high levels of electronic and mechanical reliability, Tim Owen at Cambridge adopted a different strategy. There were thus at this time two divergent approaches – the minimalist approach typified by the OBSs of Yannis Makris at Hamburg and XXX Japanese at University of Texas and the maximalist approach taken by a group of US institutions for the instrument sponsored by the U.S. Office of Naval Research. The argument in support of the minimalist approach was that for a given cost and on board space you could afford many more intruments – at least twice as many – and that this would compensate for a higher failure/loss rate and still give better data coverage.
Cambridge knew what it wanted to build, but did not have the funds or ongoing research grants to pursue the project until approached by AMOCO who suggested that we put in a proposal to develop our new OBS with a view to its commercial use. The proposal was eventually funded, but by this time AMOCO had lost interest in the project and we were left to develop the new instrument – the miniDOBS – to suit our own requirements. The major parts cost in any OBS are the acoustic release system and mechanical actuator, the pressure vessel and buoyancy, the recording system and clock and the sensors. To meet the miniDobs draft specification of a target parts cost of around £5000 it was necessary to avoid the use of a packaged proprietory release system (typical cost about £6K at that time) and instead use separate components and a custom designed plastic burn wire release mechanism, plus a 19 inch diameter glass sphere as instrument housing and primary buoyancy with lab designed and made recording electronics and sensors. A completely new recording system was designed using components that were newly available, including the Motorola 68332 processor and initially 16 bit analogue to digital converters, later upgraded to Crystal 24 bit analogue to digital converters, using a SCSI hard disk drive as the recording medium. At the start of development the largest hard disk we could buy was 20 Mbytes, but we gambled on larger ones being available in time for first use, and the instrument did indeed go into use with 120 Mbyte disks – enough for X days of recording. To minimise the required disk size and to reduce data transfer times from disk the miniDOBS used a data compaction algorithm that only recorded the significant bits of data values by using variable wordlength data packets to give lossless compression to about 11 recorded bits per 24 bit sample for an average experiment, including packaging overhead and time marks.
Early instruments in use used an initial version of the Crystal A to D delta sigma converter with voltage input, which was changed to an upgraded chip using current input that never achieved the same high signal to noise ratio in spite of extensive development work. The miniDobs recording system was used in many configurations in addition to its original OBS form, including use in Iceland as a land seismometer, and in Mexico as a moored seabed hydrophone. A number of miniDobs were built for Portuguese and Spanish scientists, and the miniDobs design of plastic burnwire release and the deployed geophone design and the geophone drop timed release were used in a number of other Ocean Bottom Seismometers. MiniDobs OBSs were built by Cambridge for Juanjo Danoieta in Barcelona and for a group at the University of Lisbon, and continued in use for several years after the Cambridge systems were retired. Later SARTI, a technology centre of the University of Vilanova i La Geltru, adapted the miniDobs software for its new OBS system using an updated processor.