Total Stations in Archaeology
John W. Rick
My point in writing about total station-type transits is
to provide current information about a technology of great utility to
archaeologists. Total stations are best described as very accurate,
distance-measuring electronic theodolites capable of diverse mapping and
position-measuring tasks. I have considerable experience using total stations
but am hardly a "power user," having little formal training in surveying. Thus,
my target audience is primarily those archaeologists not currently using these
types of instruments, but interested in their potential.
Conceptually, total stations are different from most measuring systems used by archaeologists because they are effective over a great range of scales and have an accuracy that is unusual in our experience. They encompass a range of about five to six orders of magnitude of accuracy. For example, you might be measuring the position of a point 1 km away from the total station and be accurate at least to the centimeter. The equivalent in a more common measuring system would be to use a tape to measure the distance to an object a meter away with .01 mm accuracy. To achieve an error of millimeters in measurements made over kilometers conflates scales we usually associate with different measuring devices. However, with a total station, the tool we use to make very large-scale maps can be used for precisely measuring objects in small excavations.
Archaeologists have probably been slow to use total stations, with notable exceptions, because of their cost and the complexity of their use and maintenance. I want to demonstrate that total stations are becoming cheaper and easier to use, and offer a greater range of applications than they have had previously.
Angle and Distance Measuring
Total stations combine a number of technologies to achieve their remarkable
accuracy. The first, an extension of traditional transits and theodolites, is
an ability to register very fine angular divisions. Accuracy varies with price,
but total stations are capable of measuring to the thousandth of a degree.
Obviously, the error of radial measurements increases with distance from the
measuring instrument. The angular precision for commonly available instruments
ranges from 20" (60"=1't; 60'=1deg.) to less than 1". To
give an idea of how well accuracy is conserved at distance with these levels of
angular precision, a rule of thumb is that 1" is 1 cm at 2000 m of distance,
so the maximum angular error of a 1" total station would be 1 cm when
shooting 2 km. A 10" instrument would achieve the same accuracy at a
distance of 200 m.
The second, more novel aspect of total stations is that they measure the distance to the target point with an infrared laser emitted by the EDM (electronic distance measuring device) and reflected by a prism held vertically above or below the actual point of interest. The actual accuracy is determined by the wavelength of the light used, and errors can range from as little as 1 mm plus 1 ppm (part per million; e.g., 1 mm in 1,000,000 = 1 mm/km) to around 5 mm plus 5 ppm. Thus, at 2,000 m, distance error would vary from 3 mm to 15 mm over this range of accuracies. The total distance that the EDM can measure depends on a number of variables, including atmospheric conditions, quality of the EDM, and the number of prism reflectors used as targets (generally one, three, or nine). The least powerful instruments are limited to 300-m maximum distance with one prism, while the most powerful, readily available instrument with nine prisms can exceed 5-km distance under ideal conditions. The telescopes of the theodolites vary in power from 24x to 43x, corresponding to the magnification necessary for the EDM's maximum range. Even the least accurate short-range total stations generally exceed the abilities of optical survey instruments. The angular accuracy matches that of the distance measuring, so that radial, lateral, and vertical errors are similar. Typical configurations are shown in Table 1. Anyone considering the acquisition of a total station will need to balance these factors, along with other features mentioned below.
Total stations have one or two LCD (liquid crystal) displays, which range from
1 to 8 lines of 16-40 characters, and some are capable of simple graphical
display. On high-end Nikon models, the user can display a crude map of points
to be laid out.
Total stations produce the same basic spherical measures as optical survey instruments--horizontal and vertical angles and a radial distance measure. Total stations differ, however, by taking additional data and then calculating additional measures. Most importantly, most total stations are capable of simultaneous trigonometric conversion of spherical survey coordinates into Cartesian orthogonal measures--usually east, north, and altitude. The coordinates of the instrument setup position can also be input (or found within a data register) and consistently offset from the measured points. This allows all measurements to be taken with reference to a single datum, eliminating the need to manually adjust the data from different local datums to the overall site grid.
Computational abilities go far beyond these simple transformations, and most total stations carry a number of useful programs in their memory. For instance, with the free station function, a total station set up at an unknown point can calculate its current position after measurements are made to a couple of known points. Another useful function for archaeologists is called "setting-out." The coordinates of desired points can be input manually or from data registers, and then the instrument will direct the user to the points. The screen will indicate the horizontal and vertical alignments of the point to be found, and then will report how far the prism target must be displaced out or in along the radial line of alignment. In this way the instrument can be used to lay out uniform grid systems on uneven ground, quickly locate the widely dispersed units of a random sample, or stake out the corners of perfectly square or rectangular excavation units.
The diversity and capability of available programs is increasing through new input and linkage devices that permit the use of complex code. While most total stations have small program storage areas accessible only by instrument dealers, we are beginning to see DOS-compatible operating systems and user-programmable abilities. These will allow custom design of the instrument's function and the expansion of available software. The standardization of PCMCIA-type cards and the inclusion of such slots on the total stations now allow personal computer programs to be easily transferred to the instrument.
Data Storage and Transfer
To be effective, the data the total station gathers and transforms must be
input into a computer. The transfer of information from computer to theodolite
is useful as well. Previously shot points, perhaps from an earlier project, can
be very useful reference points in the field. Complex grid system points can
also be transferred to the theodolite for the setting-out process. These upload
and download transfers can be done directly or through an intermediary storage
device. I will review these systems, going from the most established to the
For many years the most common data transfer method was through data collectors, which resemble oversize calculators that typically hang on theodolite tripod legs. Not only do the collectors store data in fairly large amounts (they may exceed 1 megabyte of memory), but they also have the ability to manipulate and display the survey data. In effect, they are in competition with the total stations themselves for being the brains of the survey instrumentation. They are time-tested, durable, shock-resistant, and increasingly they accept removable memory or program cards, giving them much greater potential than they had previously. For a lower end instrument, they continue to be a very viable solution.
A second transfer medium, which probably will decline in usage, is the proprietary memory module, usually of modest capacity (approximately 64 to 128 kilobytes), which is inserted into a receptacle on the total station. The memory modules require a specially designed reader that is attached to the data-receiving computer through either serial or parallel ports. Although now surpassed in capacity, economy, and convenience by other transfer methods, in my experience the memory modules are nearly indestructible; I have never experienced a module-related data loss of any sort. Time will tell whether the PCMCIA cards, a design not intended for rough outdoor use, will prove to be as durable.
PCMCIA data cards are rapidly becoming an industry standard for data transfer. They are relatively inexpensive depending on the capacity--which ranges into multiple megabytes--and they are easily moved between the slots on a theodolite and a computer. Almost all laptop, IBM-compatible computers now have such slots, and a slot interface is readily available for desktop machines. Given the large capacity, entire field seasons of data can easily fit on a single card, reducing the equipment necessary for field data collection. The need for computer transfer while in the field may be reduced to backup functions. Many advances can be expected in the arena of the PCMCIA interface; programs for the total station and data itself can all reside on one form of media, so that input, output, and program control are greatly simplifed. One powerful Nikon total station even carries two PCMCIA slots, allowing segregation of program and data cards.
Direct transfer of data from total station to computer appears at both ends of the price spectrum. Less expensive instruments often have an internal memory, around 256 kilobytes to 1 megabyte, which can store in the range of 500 to over 2,000 measurements, depending on the particular data recorded per shot. The data can then be transferred when the total station returns to the computer at home base; an RS-232 serial connection is standard on almost all electronic theodolites. The alternative form of direct transfer puts the computer in the field linked to the total station. Each shot is recorded directly to hard or floppy disk, circumventing any necessary on-board theodolite storage.
Storage is only one aspect of field computer direct links. When attached to the total station, the computer's greater memory and computational capability become available. Control of the total station usually passes to the computer, and complex annotation of points (types of shots, feature numbers, contextual information, etc.) can be input through menus, pointing devices, or keyboard. This alphanumeric information is not easily added from the cramped and limited keypads of theodolites. Immediate data reduction and display are another obvious outcome--the larger screens of notebook computers are much more effective than the best of total station displays in showing data point positions as they are shot. In this way, data relationships can be identified during fieldwork, and grievously aberrant point positions can alert the survey team to errors in instrument setup or alignment before significant amounts of erroneous data are collected.
I am familiar with LISCAD Plus, a Leica product, that I have used with a small IBM notebook computer, the Thinkpad 500. The software has a number of CAD (computer-aided design) features, such as surface modeling, and works within MS Windows. Output of CAD files allows direct export to AutoCAD or Microstation programs. The greater readability of the menus of such programs and the immediate graphic feedback allows crews with little field experience to be productive right away and reduce error rates while at the same time recording much more detailed information than before. Our data files are ready for mapping and CAD use immediately after error scanning.
One weak link in this field setup is the computer, which may have problems in the field, especially in dusty or wet settings. But my experience in the dry and relatively calm highland valleys of Peru suggests that with care, a notebook computer can be reliably used alongside the total station, taking into account three problem areas. First is that the computer needs an adequate platform that places it conveniently close to the total station operator and well above the ground. For a light computer, I have found that a strong wooden platform attached to tripod legs is quite adequate. Mine was produced by a helpful surveyor-tinkerer with archaeological interests; I have had little luck turning up any sort of commercially available equivalent.
Second is the problem of screen readability in bright light--this is particularly a problem with color screens, which are rapidly becoming the standard on portable computers. Shading the screen is helpful but clumsy. The best monitor for daylight readability and battery savings is a monochrome LCD screen whose backlight can be completely turned off in sunlit conditions. Another solution is a pen-based computer designed with readable screens and less vulnerability to contamination and dampness, although these are often expensive features.
The final problem is not unique to field computers but aggravates an existing total station problem: battery life. Most total stations run on removable, rechargeable, proprietary NiCD battery packs. Their duration depends on the amount of time that they are in use and the number of shots taken, but with recent improvements many battery packs are capable of lasting a full day of intense measurement. Battery chargers usually require line voltage, making them difficult to use in remote locations. We are beginning to see some instruments that accept standard alkaline batteries, and solar charging systems could be assembled but are not generally available for total station situations to my knowledge. A partial solution to both computer and total station battery problems can be found with rechargeable lead-acid sealed batteries, which also require solar or line-current charging but have much greater capacity than on-board batteries. Total station batteries, and many laptop ones as well, range around 1-3 amp hours, compared to typical lead-acid cells of 7-12 amp hours. With the right cabling and conversion, both the computer and the total station can be run off a single large battery. Two such batteries, judiciously used, could run a system for a week in the absence of line voltage.
Some Things to Consider
Total stations are not for everyone, and they do have some often unforeseen
liabilities. They are still fairly heavy, averaging around 5-7 kg for the
instrument, and another 3-5 kg for the very necessary protective travel case.
They must be treated with great care, although most product lines are built
with durability in mind. An inopportune spill of a tripod-mounted theodolite
may put you out of business for some time, given repair availability and costs.
Calibration, testing, and cleaning of the instrument is advisable before every
major season, which may cost two hundred dollars. I feared relying on a single,
electrically based system that was not field-repairable, but have not had a
failure of any part of my total station system in six major field seasons.
Apple computer users may face difficulties, given that much of the survey hardware and software has traditionally been only compatible with the IBM-type personal computers, but this seems to be slowly changing toward greater flexibility.
For those working outside the United States, a total station is an expensive piece of equipment to put at risk and one that may be difficult to pass through foreign customs. Yet I have found that it is not a recognized target of thieves or zealous inspectors because the value of the instrument is largely unknown, if not indeterminable, in many countries. Passing through customs has therefore been relatively easy in my experience.
Finally, to make use of the long-distance measuring abilities of the total station, quality two-way hand-held radios are essential. When measuring distances over 50 m, and especially at distances beyond 200 m, voice communication is not realistic and hand sig-nals are of limited utility. Radio communication gets to the heart of the matter and pays off many times in temper and efficiency. We have also found that simple signals, such as sending quick clicks with the radio's send button, can greatly speed routine messages.
The Latest and Greatest
As with most technologies, the most exciting new features are at the top and
bottom of the price range. Decreasing prices at the lower end should allow
cash-strapped archaeologists into the market. Street prices for lower end total
stations (5"-20" angular accuracy, 3 mm plus 3 ppm to 6 mm plus 6 ppm)
should fall in the $5,000-7,000 range. Tripod, reflector, and reflector rod
need to be purchased as well, but with some competitive pricing, you can work
with an excellent instrument for far under $10,000, something unthinkable just
a couple of years ago. Think carefully about your accuracy, distance, data
storage, and special program needs, and choices will narrow quickly.
At the high end, in addition to ever-increasing accuracy, there are some very useful features emerging. Most outstanding are the motorized and automatic target recognition instruments starting to appear on the market. A number of manufacturers are making total stations that have motor drives to seek specific vertical and horizontal points, something useful in setting-out or finding previously recorded points. While this increase in speed and ease of sighting has value in archaeology, it may not justify the considerable additional cost. Automatic target recognition allows one-person operation, because the total station will track the reflector as it moves across the landscape, and its functions can be radio-controlled by the person holding the reflector. This technology is new and expensive (expect $30,000 plus), but will undoubtedly mature and become more accessible in the near future. For CRM projects, the reduction in costs (one person less and greater speed of measurement) may help justify purchasing the instrument. Although even at this expense the instrument has a limited tracking range generally not exceeding 1,000 m, in my experience it is not usually effective to work at this distance.
Another new innovative feature helps in setting-out. Normally, one of the most time-consuming tasks is bringing the rod person into line with the theodolite's orientation. However, a number of higher-end instruments have point guides that project a split, low-power visible light laser, half of which is intermittent. The reflector-holding person knows they are in alignment with the theodolite telescope when both the flashing and constant beams are visible. This feature has a working range of about 100 m, beyond which it is difficult to distinguish the guide.
Because of dropping prices, increasing reliability and numbers of features, and
decreasing weight, total stations are increasingly likely to enter the
archaeologist's toolkit. For mapping of all kinds, the additional speed and
accuracy offered by total stations demand our attention. The reduced error
rate, greater annotation potential, and higher efficiency of electronic data
transfer contribute to timely, successful data processing. For local area
surveys, total stations could be quite useful in place of or in conjunction
with GPS instruments. Visibility of the target reflector will always be a
problem over significant distances of broken or heavily vegetated terrain, but
the accuracy of measurement would be difficult to match with current GPS
technology of equivalent cost. For example, in open country, a high-power total
station on an elevated point could potentially record points within an
approximately 50 km2 circular area, with even the farthest points
not exceeding an error of about 2 cm. Total stations can also work efficiently
and accurately on a very local scale, such as profiling excavation features and
recording three-dimensional artifact locations. With the additional notation
abilities of theodolite-linked computers, excavation materials can leave the
site fully cataloged, with excellent numeric control and the potential for
immediate spatial display. Now is a great time to assess whether a new total
station or an upgrade of an old one would make sense in your field project.
I would like to thank representatives of Leica (800-645-9190), Nikon (East Coast 516-547-4200, West Coast 310-516-7124), Sokkia (800-476-5542), and Topcon (201-261-9450). My Leica distributor, Haselbach Surveying Instruments of Burlingame, Calif. (415-348-7247), has been instrumental in keeping me equipped and updated on trends in total stations.
John W. Rick is in the Department of Anthropology at Stanford University.