Last up dated: 11 March 2004
Posted on: 17 February 2004




by Glenn Chapman

When Kelly Johnson designed the U-2, it was imperative that anything not absolutely essential to flight not be part of the design. This included an ejection seat, high-frequency radio system, automatic direction finder systems, radar, in-flight refueling, and many other systems. Although most other American military aircraft had many of these systems, it was essential to stay within the “one pound of weight equals one foot of altitude” rule. Still, because of the pilot’s inability to see anything directly below him during flight, and because the U-2 had only rudimentary navigation systems, and to assist in some systems platform’s operation, some sort of viewfinding apparatus was needed. However, it couldn’t be too big, must be very, very light, and be simple in operation.

A company in Norwalk, Connecticut was awarded the contract to devise such a system. The company was called Perkin-Elmer and was experienced in designing and manufacturing optical equipment. They designed a thing called a “Driftsight” which would be operated with a fairly simple electronic hand control device capable of looking around the underside of the aircraft from about 85 degrees horizontally to the vertical “nadir” position. (Nadir is a term defined as “perpendicular to the horizontal axis.”) The hand control could also provide a “tracking” function that initially was thought could help in controlling the trip rates of the camera (or Nephographic Systems as it was originally called) while in flight. The pilot could crank the altitude and calculated ground speed into the hand control and a V/H or “V over H,”, signal could be computed. This is the quotient of V, or ground speed, divided by H, or altitude, giving a ratio to the camera platform. This was the standard then, and remains ;the standard today, that provides the camera system with the information needed to allow the shutter to trip at a rate that provides a 56% overlap of film coverage. He would be able to check the accuracy of the tracking by viewing the ground movement through the driftsight.

The driftsight system designed by Perkin-Elmer was designated as the “Mark I Driftsight” and the driftsight hand control as the “Mark II Hand Control.” Connection of the hand control to the driftsight optics mechanism was fully electrical through small servo systems that provided an “A”, or azimuth, and “A+E”, or azimuth plus elevation, movement of the driftsight optics. Both had to work together in order to move the optics around properly. The driftsight also had a X1 and X.4 (Times 1 and times .4) magnification of the area being viewed below the aircraft that was controlled by a switch on the hand control. At the top of the driftsight was a “viewing knuckle” that was bent about 85° degrees from the driftsight itself, which was positioned vertically behind the instrument panel. Inside this knuckle was a fixed upside-down inverted “V” mirror that reflected image to the viewing glass at the front of the knuckle This was was an optically-ground glass about six inches in diameter that allowed the pilot to view non-distorted image through the driftsight. In front of this glass was another viewing plate called the “drift reticule plate” with vertical lines on it. At the top of this plate was a small rectangular plastic tab that was marked in degree increments of up to 7°. Inside the driftsight itself were a series of optically-correct lenses and a “crosshairs” reticule that was aligned with the forward axis of the aircraft. By positioning the tab on the plastic drift reticule plate and aligning it up with the absolute direction in line of flight, the pilot was able to determine the exact amount of drift he was experiencing with the aircraft. There was no “drift string” atop the nose of the aircraft as many in those days had, because the pilot would not be able to see it because of obstruction due to the placement of the viewing knuckle. In fact, this knuckle created a blind spot that prevented almost any forward sight from the cockpit whatsoever. The movable optics inside the lower part of the driftsight itself was housed in what Perkin-Elmer called a “hyperhemispherical” glass bubble about five inches in diameter and optically ground to prevent viewing distortion. This, in a nutshell, was the original driftsight configuration in the original U-2 aircraft. A long rubber hood called a “boot” was attached to the viewing knuckle to prevent any shadows or extraneous light from obscuring the image through the driftsight. This was in effect doing the same thing that early photographers did by throwing a hood over their shoulders while viewing an image through a camera lens.

In addition to the driftsight, an optical sextant was installed at the top of the aircraft nose just forward of the windscreen. This was the main navigational system used in the early U-2 aircraft whereby optical celestial sightings could be made. The sextant, although not an integral part of the driftsight, was attached to it via a “roof mirror” attached to a sextant pull knob at the upper right-center of the instrument panel just below and to the right of the viewing knuckle. This roof mirror was comprised of two rectangular front-surface mirrors attached in a “V” configuration. A rectangular access in the top rear section of the driftsight housed this roof mirror assembly. Sightings through the driftsight or sextant could be made, but not both at the same time. For viewing through the driftsight, the sextant knob was pushed inward all the way, moving the roof mirror out of the driftsight’s optical path. For viewing through the sextant, the sextant knob was pulled fully outward, pulling the roof mirror into the driftsight optical path, allowing the image from the sextant to be reflected onto the roof mirror, thence upward through the driftsight viewing knuckle. Instrument technicians had responsibility for maintaining the sextant and roof mirror while the driftsight and hand control were the responsibility of the Nephography Shop technicians.

In order to prevent the driftsight and sextant optics and hyperhemispherical glass bubbles from fogging at altitude, a purging of all air from the two bubbles needed to be accomplished. This was performed by connecting a bottle of -96° ultra-dry nitrogen dew point to a connection that allowed the nitrogen to enter the bubbles under pressure and push the air from the bubbles. This took about forty-five minutes immediately prior to engine fire-up and lasted until approximately ten seconds after the engine had come to idle speed. Originally, the sextant and driftsight were purged separately, but a modification was soon made that allowed both to be purged at the same time, allowing a much more perfect purging cycle to be accomplished.

As the U-2 design progressed, and the systems platforms began to take shape, it was determined that the camera platforms that the U-2 would be using were so simplistic themselves that no V/H signal would be required at all. The camera systems designed originally for the U-2 were the Hycon A-1, A-2, and 73B configurations, and a Perkin-Elmer Mark II 70-mm Tracker Camera. The A-1 and A-2 configurations used a motor-and cam switch system to provide a 12-second or 15-second camera trip rate, depending upon whether a 4- or 5-digit cam was used and the 73B had a “programmer” unit of relays that used a base of 70,000 feet to set trip rates internally. The Tracker Camera used a “relay logic” binary intervalometer for scan rate, normally set for 32-seconds. None of these systems required the usual V/H signal, although future systems might be developed, and ultimately were, that could use it.

The Mark I Driftsight was a very reliable unit, but required a lengthy amount of time to install and align it with the line of flight. This was done by first aligning and stabilizing the aircraft laterally and dropping a plumb bob from the pitot tube under the forward part of the nose, which was about as close as possible to exact centerline of flight of the fuselage. Using masking tape, an “X” would be marked on the ground directly under the point of the plumb bob. Then the plumb bob was aligned with the forward part of the ARC-34 UHF Radio Antenna directly behind the fuselage equipment bay and again a masking tape “X” was marked. The antenna, like the pitot tube, had been determined also as an exact centerline mark. Then a piece of chalk line or length of masking tape was used to align the two marks made with the tape to provide a “line of flight” marker under the aircraft. The driftsight optics were then manually adjusted under the cockpit floorboards so that they were pointing at the nadir position. With one technician in the cockpit looking through the viewing knuckle and a second technician on the ground at the driftsight bubble, the driftsight was positioned so that the fore-aft lines inside the driftsight reticule were perfectly aligned with the line of flight mark on the ground. After securing the driftsight, the optics were checked again to ensure that line of flight had been properly accomplished. The roof mirror inside the viewing knuckle also had to be aligned to the cross-hair reticule inside the driftsight. This was done by viewing as closely as possible through the center of the knuckle and moving the knuckle right or left until the “swinging image” of the roof mirror line matched up with the reticule line-of-flight line.

The Mark II Hand Control, as well as the driftsight itself, could be a monster to work with and very labor-intensive. The driftsight was pretty reliable, but not so the Mark II Hand Control. The biggest problem seemed to be keeping the servos inside the hand control properly aligned. It always seemed like little gremlins would crawl inside it and loosen any security the servos had. This was a 115 VAC, 400 Hertz system along with the normal 28 VDC to accomplish some functions. The 400 Hertz power was solely for the servos, while the DC power was for indicator lights, camera operation, driftsight optical magnification, and other systems operations. If the aircraft electrical inverter system varied by only a few Hertz, the servos would not position right. There were many problems with this hand control, but these were the most important and caused the biggest problems.
Connection of the Mark II Hand Control to the Mark I Driftsight was in itself a major job. At the lower starboard position of the driftsight underneath the aircraft floorboards were the two mechanical drivers for the driftsight optics. Access was through a panel hole in the right floorboard about five-inches in diameter.

The drivers were nothing more than two circular metal discs about the size of a nickel that each had a slot machined across them. These attached directly to the optics drive mechanism. These slots had to be aligned vertically in a near-perfect position in order for the Mark II Hand Control electronics adapter to be positioned. Inside this adapter were two discs identical to the optics driver discs except that in place of a slot, a tab across the disc was affixed. These tabs also had to be positioned in a near-perfect vertical position in order to align the adapter to the driftsight. An electrical Cannon plug was attached at the exterior of the adapter that allowed the connection of the hand control cable. Inside the adapter were two servos that were attached to the positioning discs. The entire adapter had to be installed perfectly onto the driftsight with three Allen head 10-32 machine screws. And this all had to be done by feel, underneath the floorboards, bending over on hands and knees, with other cables and stuff in the way. It was possible to secure the adapter onto the driftsight with the discs not aligned properly. There was no way of knowing until power was applied to the hand control and carefully moving the handle around to see if the optics tracked properly. If the discs were mis-aligned, the optics would not move or would move erratically. More importantly, the entire driftsight optics driver assembly as well as the electronics adapter could be severely damaged or, at the very least, need re-alignment in the shop, requiring even more time. Usually the adapter was initially secured only “finger-loose” so that if alignment was not right, the chance of damage might be less. Sometimes it took hours just to get this adapter properly positioned before the rest of the job could be completed. Alignment of the hand control was not required because the servos had already been adjusted in the shop. Once the driftsight and knuckle were installed, properly aligned, the electrical adapter connected properly, and the hand control connected and powered up, the system could be checked operationally. A checklist was accomplished in which the optics were moved around to check for proper positioning and operation, the driftsight magnification worked right, and tracking was done as it should be.

After only a couple of years, it was determined that a simpler hand control system could be used that would cause less maintenance efforts to be expended. A company called Baird-Atomic developed a purely mechanical unit that used no servo systems whatsoever. No tracking was done. The unit was about 25% as big as the original Mark II Hand Control and the only things on the panel were a Mode Switch (for operation of the camera platforms), a Master Switch (provided power to the tracker camera or other systems), a “X1-X.4” Switch (for driftsight magnification), four indicator lights (”A”, “B”, “C”, and “D”) that indicated camera modes, and a triangular plastic handle about four-inches long that was used to position the driftsight optics mechanically. This hand control was designated as the Mark III Hand Control and was extremely reliable, simple to use, and much less difficult to maintain and align.

The Mark III Hand Control used two 12-tooth splined cables connected between it and the driftsight optical drivers to provide optical positioning. In place of the Mark II Electronic Adapter under the floorboards, the Mark III Hand Control used two small mechanical adapters that connected to the optics drive assembly much the same way as the Mark II Adapter had. A splined cable connection was at the exterior of each adapter. Installing these adapters and aligning them with the driftsight optics drivers was much easier than the Mark II Adapter. These adapters were secured to the driftsight with three 10-32 Allen-head screws much like the Mark II had been. It was learned early-on that the easiest way to install these Mark III adapters was to use an old piece of the splined cable about two-inches long as an alignment tool. Again, this was all performed underneath the floorboards on hands and knees.. The splined tool would be put into the adapter, one screw with the Allen wrench already attached to it was put into one of the three holes in the adapter, the whole thing secured with the right hand, and placed through the hole in the floorboard onto the .lower optics driver disc. While holding onto the adapter, Allen wrench, and screw, the splined tool would be moved around slowly with the thumb and fingers until the alignment tab fell into the alignment slot on the driver disc. This was easy to feel, unlike the Mark II adapter, which one could not feel at all. Once the tab and alignment slot were aligned, the Allen wrench would be turned until the screw was “finger loose.” Then the other two screws were installed and all three were tightened down onto the driftsight properly. The top adapter would be installed the same way. The two splined cables were now connected and it was time to get back to the Mark III Hand Control.

Aligning the Mark III Hand Control was much easier than the Mark II Hand Control. Again, a line-of-flight line was installed underneath the aircraft. While looking through the viewing knuckle, one of the splined cables was turned clockwise while the image in the driftsight was being observed. If the image only went around in circles, that cable was the “A”, or azimuth cable. If it went up and then turned right and kept going, it was the “A+E”, or azimuth plus elevation cable. While turning the “A+E” cable, the image moved upwards and to the starboard side of the aircraft until it was approximately 90° in respect to line-of-flight. While holding the triangular handle on the Mark III Hand Control fully vertical and pointing about 90° to the right, the “A+E” cable was moved counter-clockwise until the image intersected the line-of-flight marks on the ground. The “A+E” cable was connected to the rearmost splined connection under the hand control and the “A” cable to the other, foremost, part of the hand control. Now the hand control was positioned into the right cockpit console directly aft of the Auto Pilot control unit but not yet secured into the console. Again, the handle was rotated and moved up and down carefully to check operation. Then the handle was pulled full downward and rotated until the detent in the hand control was felt. The “A” cable was dis-connected and the image aligned by rotating the cable until the pitot tube was aligned with the driftsight reticule line. The “A” cable was re-inserted and the hand control was checked again for full operation up and down and full around the underside of the aircraft. If fine-tuning of the cables was needed, it was usually only one tooth plus or minus. A last check was made whereby the handle was put into the detent position and straight up with the optics pointing at nadir position. While holding the handle full up, the handle was pulled in a full 360° circle. The image at nadir was checked to move in either a very tight circle or, better yet, no circle at all. At this point, the hand control was considered to be in alignment.

Eventually, the line-of-flight alignment was discontinued and the “quarter method” began being used. This was highly illegal at the time, but saved a lot if time and was much easier. I was the one that initiated the quarter method way back around 1960 or so, scared to death I would be caught, and then was caught by none other that Colonel “Big John” DesPortes, our Wing Commander. All he told me to do was re-write the checklist and be sure to “---put a FOD (foreign-object damage) phrase in it.” It finally evolved into a small brass Hand Control Alignment Tool with a “Remove Before Flight” streamer.

The Mark II Hand Control was very precise, but was extremely over-designed for use in the U-2. It was a relatively unreliable unit, was difficult to install and align, and very time consuming for both flight line and in-shop maintenance and alignment, and was expensive. It was used much more by the CIA pilots than with the Air Force U-2s.

The Mark III Hand Control was not as precise as the Mark II, but was extremely reliable, could see almost a full 90° up and down, and could see 360° around the underside of the aircraft. It was not difficult, except for the adapter installations, to align, did not require a large maintenance and alignment time on the flight line or in-shop, and was inexpensive. It was used almost exclusively with the Air Force U-2 program.

Eventually, six of the Mark III Hand Controls were modified with a special detented collar around the handle. This collar was easily positioned by the pilot to any of the four positions. The reason for this modification was to allow the pilots to use the detents while the hand control was positioned to one side or the other. Each detent corresponded to a specific camera angle, especially the three Hycon 73B camera “look angles,” so he could determine whether he was properly on track to get the best oblique photography. When the hand control was positioned fully to starboard or port, a detent notch in the collar could be selected, the handle pulled down into this slot, giving the pilot an extremely good view of his track, allowing him to move the aircraft as desired to get optimum photo coverage with the camera platform. This hand control was designated as the Mark III Modified Hand Control or as was commonly called, “Mark III Mods.”

As the U-2 evolved, so did the driftsight. Eventually the Perkin-Elmer Mark I Driftsight and Mark II Hand Control and the Baird-Atomic Mark III and Mark III Mod Hand Controls were replaced with a much better driftsight system manufactured by the McDonnell-Douglas company. As the U-2R aircraft were modified into the U-2S digital “Glass Cockpit” models, the driftsight system was removed. Technology had evolved to the point where Global Positioning Systems and other high-tech systems were installed, relegating the driftsight and hand control into antiquity. For it’s time, however, it was a useful photographic and navigation aid for U-2 pilots. There are a few pilots, and Driftsight Dragons also, out there today that will bemoan the loss of this “old stovepipe” and reminisce endlessly over it. I am one of those Driftsight Dragons and miss it, sometimes with tears in my eyes, and other times, as I remember my times with it, like I would miss a toothache. The old “Driftstick” is now in Optical Heaven for sure. Or would it be Optical Hell?

Who knows, and who really cares? Only us old Driftsight Dragons and a few old Dragon Pilots.


Perkin-Elmer Mark I Driftsight in Early U-2 Aircraft
Driftsight Part Number Was PE-151-0001


Baird-Atomic Mark III Mechanical Driftsight Hand Control
Hand Control Part Number Was BA-501-0005


©2004 Glenn Chapman

Many thanks to Glenn for writing down this excellent history and set-up of the driftsite used on the U-2A....He is a truly the "Driftsight Dragon" -John


Back to the main U-2 page