First posted March 28 2010
Completely re-written Jan. 18 2011
Five Altimeter Bombs
It was a well-placed bomb detonation in the forward hold of PA103 that sealed the fate of 270 people on the evening of December 21, 1988. To be precise, the UK Air Accidents Investigation Branch (AAIB) decided it had detonated at 19.02:50 (7:03pm) – almost exactly thirty-eight minutes after it left the ground from London’s Heathrow Airport at 6:25.
This time span, officially decided on by Libyan plotters for unclear reasons, is an uncanny fit with another technology, controlled by another group entirely – Ahmed Jiibril’s PFLP-GC. To recap, Marwan Khreesat is a Jordanian intelligence operative and long-practicing maker of airliner-killing bombs for Jibril (with exact allegiance unclear in a chicken-and-egg sort of way). Khreesat’s bombs were triggered by altitude (air pressure) and hidden inside consumer electronics – usually radio-cassette players. In the fall of 1988 he was called in again by the PFLP-GC on a mission for Iran – avenge the US destruction of Iran Air flight 655 by destroying an American airliner themselves.
Khreesat was hired to make or modify five bombs, and contrary to GID orders (as we've heard them), compelled to make them fully armed. Only one radio-IEDs of the five was intercepted on the arrest of Khreesat and the whole cell on October 26 1988. Three more of these were found later by German authorities and studied, given the numbers below. All seized Khreesat devices, in order of discovery:
1) Toshiba BomBeat 453 radio-cassette
2) Sanyo computer monitor
3) Ultrasound radio
4) Ultrasound radio
The fifth device, which Khreesat says he only soldered two wires on after someone else constructed it, was a different Toshiba radio (not the BomBeat seized nor the Libyan BomBeat SF-16 officially accepted). The someone else who made it was likely "Abu Elias," a relative of Jibril’s and a senior PFLP-GC operative. Khreesat thinks Abu Elias worked by copying his own work, perhaps recycling some parts from the fifth test model he made for Jibril back in 1985.
This one hasn’t been studied, as it officially never turned up. Many, however, still suspect it did surface – on Pan Am flight 103. So we’ll return to it at the end, after grasping the four that were definitely never used.
Gobel and Khreesat explain the time aspect
[Note: a separate post details the hardware end of these bombs in a little more detail.]
David Leppard’s curious 1991 book On the Trail of Terror describes the work done by German investigators on the four bombs found in Germany, and summarized the findings of BKA scientist “W. A. Gobel." This looked at how the altimeter and time delay modifications worked. Here he refers only to unit 1, the seized BomBeat 453:
“Gobel had tested the performance of the pressure pick-up gauge in a vacuum chamber. The circuit would close at a pressure between 940 and 950 milibars - equivalent to an altitude of about 2,400 feet. The contact point was connected by a wire to a condenser which acted as a time delay switch. The condenser was an insulator, absorbing the electrical charge until it fired. At this point the circuit closed, detonating the bomb.”The condenser here is what’s elsewhere called an ice-cube timer – a simple metal circuit, coiled-up, I believe, and encased in a cube of translucent resin (looks like an ice-cube made with dingy water).
"When the necessary operating height has been reached the fall in pressure connected with it will start the timing mechanism, and when the delay period has elapsed the detonator will be activated. […] The time delay of the electronic component fluctuates over a wide margin since the structure of the circuit is relatively simple. Time delays between 35 and 45 minutes were measured." [p 11-12]This 35-45 minute range is often cited, wrongly I think, as the range of times to be expected in field usage. But a passage from Khreesat's interview with the FBI in November 1989 explained that the variability comes from repeated tests on the circuit in a lab setting:
"Khreesat advised that the times are not exact and the time changes depending upon how long the timers have been tested after last being used. They usually reset to zero after a day. He used to test the timers three times in a row before installing the timer in a device. He found that in each test the time decreased. When this happened, he put the timers aside, and the next day when he tested them, they would run for the same time as when he had first started them." [Marshman report, page 32, PDF p 19]The variability then, with repeated testing, is downward from a long baseline time. Absent repeated tests, which might be unwise in an assembled explosive device, the timers would all have, as the designer described them as having, one set time each, and it would be the long time. So in use, the device Gobel analyzed would detonate about 45 minutes after triggering. This would be more than 50 minutes, not 38, from takeoff.
That certainly doesn’t mean it’s not a fit. Khreesat also told the FBI that the timers were not the same from one unit to the next:
"One of the timers was a half-hour timer, one was for three-quarters of an hour, and one was for one hour. Khreesat does not recall what time the fourth timer was set for. None of the timers were for more than one hour." [Marshman report, p 32]These seem to go in nice 15-minute increments - 30, 45, 60. His early bombs in 1970 and 1972 blew up too early - about 15 minutes after triggering - leaving two of three planes able to safely re-land. Thus for "stronger medicine" in the 1980s, he went for increments at or higher than 30 minutes. By deduction, the fourth unit was a repeat of one of the other three.
Three more to consider
Herr Gobel at the BKA, as we’ve seen above, made a close study of Khreesat device number 1, apparently housing the (or a) 45-minute ice-cube timer. He also looked at number two, the IED hidden inside a Sanyo computer monitor. This was armed with Semtex-H like the others, and rigged with the same basic altimeter-timer arrangement. “Calculated on the basis of the values of the built-in components,” Gobel wrote, “the delay time is put at between 30 and 35 minutes.” [Zeist transcripts p 8769] It's unclear if he means test variability or range of estimates for its one field-use time.
Leppard bemoans the German’s presumptions based on only one bomb sample to study, but they had units 1 and 2 and also, at one point, two other such devices. Numbered 3 and 4, these were both housed in Ultrasound radios, both found after the Sanyo monitor in a third raid, and were both destroyed under grim circumstances.
In April 1989, one of the Ultrasound IEDs inexplicably detonated while in the hands of a highly experienced BKA bomb technician, killing him and maiming an assistant. The other of the two (it’s not clear which is 3 and which is 4) was then destroyed rather than be studied, for obvious safety and psychological reasons.
But despite the British alarm over this German “sloppiness,” it wasn’t a complete scientific loss. Aside from what was missing in the rubble, which was safe enough to examine, the BKA’s forensics section ST33 noted on what remained from one the two:
“[T]he accompanying capacitor is of is of the same value as in 1 and 2, but has however, jumped out of circuit. […] it can be assumed from the remains of the circuit that the time delay was in the same region as 1 and 2." [p 144]From four intercepted IEDs, we have three capacitors, all apparently of the vague 30-45 minutes “region.” This comports reasonably with Khreesat’s list of 30, 45, 60, and one repeat. It seems to be lacking anything with a longer time like one hour, but one device, likely the one that killed a bomb tech, was not studied like this. Perhaps that was the one-hour unit.
The results are not 100% conclusive with the information I have, but it remains consistent enough between Khreesat and the Germans to say we have a double-corroborated explanation, more-or-less, of what to expect with each of the four seized devices.
And the Fifth?
Khreesat felt that all four bombs he was tasked with re-used the timers and altimeters from his 1985 test run. But that involved five radios, one of which he says was kept intact by Jibril to study. It’s worth wondering if these same parts, fitting the same patterns, were later used to build Abu Elias’ fifth device. But wondering aside, we can’t be certain.
We can examine the facts of December 21 to see whether or not it’s compatible. Below is a graphic from the AAIB's report on the bombing, based on Flight 103’s data recorder. The key line here is altitude, of course labeled “ALT.” The other is some measure of airspeed. The left bar gives feet above sea level (add three zeroes) and the bottom gives time in minutes from data start.
But inside the plane, where the suitcases were, the pressure is controlled and drops more slowly. The work of Rainer Gobel, as related by Leppard, is also confirmed by Paul Foot, writing in 2001 in Lockerbie: The Flight from Justice, as addressing the expected rate where it matters:
"A clearer description of this ice-cube timer was given the same day by Rainer Gobel, physicist for the German police, the BKA. He had a remarkable insight into how long such a timer would take to set off an explosion. “Pressure change in the cabin,” he said, “is regulated automatically. It happens more slowly than the drop of pressure in the air through which the plane is flying, which means, according to the documents which were provided to me, in a plane of the aforementioned class, about seven minutes after take-off the pressure within the cabin will have dropped to such an extent that 950 millibars or hectopascals will have been reached at which the barometric gauge would be involved and a circuit, a current, would be activated.” [p 19]Some would protest, and fairly I think, that we can't be certain that even if the PFLP-GC's bombs were involved, that this one used the same altimeter setting, or altimeter, or even timer style, as the units examined. But considering the group's reliance on Khreesat’s formula and available materials (the fifth device from 1985), it seems possible enough that it would. And Gobel's "about seven minutes" is quite a nice fit with the roughly eight required to make this fit into Khreesat’s 30, 45, 60 framework. Thus the strong possibility that cannot be reasonably discounted here is that Flight 103’s ascent triggered another ice-cube timer, of the 30-minute size, about eight minutes after wheels-up.