Shough Report

1. Abstract

2. General Description of the Radarscope Photographs

3. Frame-by-Frame Description of Radarscope Photographs

4. Radar Specifications and Mode of Operation

5. Reconciling Time and Distance Data

6. Interpreting the Unidentified Echoes

7. Conclusions



Download full
report as PDF

Anomalous Echoes Captured by a B-52 Airborne Radarscope Camera

Martin L. Shough


1 UK Research Associate, National Aviation Reporting Center on Anomalous Phenomena (

2 There are small uncertainties in these measurements due to inherent limitations both of the photography and of the PPI display. The geometric scope centre can be determined from the bearing ring, but the range rings show a perceptible distortion from perfect circular symmetry and are not perfectly concentric (see below). Also the print focus appears to be of variable sharpness, and/or the video gain is varied. There is also variable noise speckling in the altitude hole. Thus the rings thus have some perceptible thickness and often appear faint and broken, sometimes vanishing, which introduces uncertainty. Original range and azimuth values used were based largely on careful measurements made by Brad sparks, using the nearest parts of what were believed to be 1-mile range rings to minimise distortion effects. But further research has required revision of the PPI range scale from 1-mile rings to 0.5 mile rings, and moreover because the display trace time is triggered not at zero range but at a range corresponding to the edge of the TR (transmit/receive) hole, given as at least 2000 ft in training manual CDC 32150K, Vol.4, ranges measured from the scope centre are therefore increased by a "hidden" radius. The effective zero-point of range is 0.25 NM, such that (for example) the third, brighter, 1/2 mile range marker is actually not at 1.5 NM but at 1.75NM, as shown in Fig.3. Values cited are believed accurate within error bars of +/- 0.05 mi and +/-1 deg.

The cause of the display eccentricity is probably optical. The range rings are electronically produced by brightening the scope trace and may be subject to instabilities in the video voltages or changes due to local electromagnetic or geomagnetic fields; but the direction of the eccentricity, diametrically opposite to the anomalous bright patch at around 26 degrees on every photo, suggests that the distortion is due to the angle of photography and convexity of the tube face. This would be consistent with Richard Haines' identification (unpublished private report) of the bright patch as an offset ghost image of the bright centre-spot, doubly reflected via the camera lens.

There is a complication, however, because the camera optics is understood to consist of a system of prisms and a lens mounted inside the CRT, the optical axis passing through the middle of the drive gear that rotates the tube inside its yolk to maintain "north-up" or "heading-up" presentations. (A second optical system combines an image of the clock, counter etc. onto the same film frame.) The image is therefore a reversed view of the inner, convex, phosphor-coated surface of the tube face, not of the outer, convex glass surface. Haines' theory therefore perhaps needs reinvestigation. Nevertheless the bright patch in question is almost certainly an artefact of an analogous kind.

3 As mentioned, the radarscope clock shows 0906:14 GMT at the time of frame 771, or 0406 local time. However, a transcript of the radio transmissions between RAPCON and the aircraft commander is at variance with this, indicating that the B-52 lost its UHF transmitters at 0358 local time, simultaneous with the first appearance of the radar target. Compounding this, the investigating officer's timetable of events linked to the B-52 flight track has the aircraft some 16 miles NW of Minot at 0406 by the clock; yet the transcript of the pilot's communications with RAPCON would place the B-52 some miles SE of Minot by 0406 and perhaps already turning back on a NNW heading after executing its missed approach.

These inconsistencies at first sight suggest a possible 7-8 minute systematic error in the times recorded either in the RAPCON transcript or by the radarscope clock. The photographed clock may seem the more reliable record of the two, given that there is evidence that RAPCON transcript is at least incomplete. However clock error cannot be ruled out, since the clock is a mechanical timer which is required to be re-set by hand prior to each mission. Setting the clock accurately (to GMT) was a checklist item, but according to former B-52 navigator Richard Sessler it was nevertheless "easy to forget" and he admits to doing so himself on occasion.

The possibility that the clock may run inaccurately has been considered. An inquiry by Jim Klotz to the manufacturing company, Bulova, disclosed that typical variation acceptable in a comparable instrument watch might have been in the region of +/- 1/2 minute over 3 days, and that a variation of 7-8 minutes in 10 hours (the duration of the mission) would indicate a mechanism in bad need of an overhaul. Although this has to be considered less likely than a setting error, as mentioned above the photos do indicate a small discrepancy between a scan/photo rate of 3 seconds and the clock time of around 0.5 - 1.0 second over 36 seconds which, multiplied out over 10 hours, would accumulate to approximately 8 minutes. In other words the clock is "slow" relative to the nominal 20 RPM radar scan rate, which, if the latter could be relied upon, might be explained by inadequate winding or a weak spring, for example.

However the discrepancy from the indicated Control Tower time is in the other direction, and would require the clock to be fast, not slow. So this phantom 8 mins appears to be a coincidence. The small photo discrepancy is probably due to imprecision in the radar antenna's hydraulic motor and/or rotation control mechanism (the ASB-4/9 Tech Order gives the nominal 20 RPM setting as "17.5 - 22.5 RPM") and there is no internal evidence in the photos that the clock is not accurate.

4 Blue Book documents and maps indicate that Werlich located the photo position at the end of the radar event at least partly on the basis of contemporary crew testimony to the effect that the camera was not switched on until the end. McCaslin confirms Werlich's contemporary report. This is also consistent with psychological and operational factors. When the UFO first appeared the B-52 was already heading home at this point, there were no bomb-nav exercises in prospect and all that remained to complete the Instructor pilot's Standards & Evaluation exercise was the approach to Minot. When reports were received from RAPCON of a UFO in the area, McCaslin asked Ritchie to switch the radar into Station Keep mode. It was not uncommon to do this during an approach, but camera operation is not automatic and has to be manually selected. No-one thought of the camera until the pilot suggested it near the end of the event.

5 An STC filter (not to be confused with Sensitivity/Time Control) is also sometimes known as Fast Time Constant, or FTC, a name more descriptive of its function: It is a circuit in the video amplifier whose purpose is to suppress echoes on the display in inverse proportion to the rate of rise of the input signal, or in other words to the steepness or sharpness of the leading edge of the envelope. The effect if used in a terrain mapping radar mode would be to eliminate or reduce the intensity of echoes with slower rise and decay times which return from gently varying terrain, whilst preserving sharper echoes from features like buildings, riverbanks or coastlines.

6 Power reflection coefficients at normal incidence in excess of 10^-14 or -140db attenuation are thought possible on occasion (Atlas 1959), so given a peak power in the order of 100 kW then for a spillover lobe 30db down on the main lobe this gives about 10^-17 of 100 kW or 10^-12W, which is about the power commonly reckoned to be in one Just Detectable Echo or a faint blip at the limit of detectability on a typical PPI. Evidently much higher efficiency than one JDE would be indicated in this case, given expert testimony of an echo significantly stronger than that from a very large jet at close range. Since the ratio of normal signal intensities on a PPI might easily range up to 10^4 (a single aircraft at constant range could vary by as much as 10^3 depending on aspect) we should probably assume an echo of at least 10^4 or 10^5 JDE.

7 The UHF frequency in question was around 270 MHz. Given dark sky conditions one would have to assume a considerable depth of weakly-ionised air in the line of sight in order to achieve a significant radio opacity at the same time as zero effective optical emission from recombination. It seems conceivable that a certain critical electron density might create a "one way" radio mirror, if the aircraft UHF transmissions were of lower power than the ground UHF transmissions, explaining the loss of all transmissions from the aircraft to RAPCON during the incident whilst the aircraft could receive RAPCON transmissions perfectly. There is presently no information on radio power outputs however.

The main problem with this model would then be how to account for the origin of a large volume of ionised air in the absence of any atmospheric electrical activity. It would be reasonable to speculate that this could be related to the presence of the unidentified radar target and associated visual reports of a luminous object or objects; but as already discussed there is presently no available physical model of such a phenomenon.

Also the ionisation theory does not explain in a very natural way why the UHF transmission cut out suddenly in the "middle of a word", implying a very rapid rise in the electron density along the radio line of sight some 90 degrees of azimuth and some 6000 ft away from the presumptive source. It would be interesting to investigate the decibel attenuation required for signal strength to go from receiver saturation to receiver noise level. Some estimate of the rate of rise of electron density in the line of sight could then be derived, leading possibly to limits on ionising particle energies and an overall mean power output.

Such speculation aside, this model is fairly consistent with the report. Short waves are less subject to scatter from ionisation than are long. This dependency has been measured as proportional to the cube of the wavelength for electron densities typical of meteor trail ionisation, ~10^12 cm^-3, and proportional to the power 7.7 of wavelength for auroral ionisation with electron densities around 10^5 or 10^6cm^-3. It is reasonable to assume that densities in the present case would be lower still and the wavelength dependency of a higher power than 7.7. This fits the sequence of events in the RAPCON transcript: Ground reception of the B-52 was initially weak on 271.3 MHz but immediately loud and clear when they switched to 326.2 MHz. Conservatively assuming a power 7.7 wavelength dependency, ionisation would be more transparent by a factor of about 4.0 or better at the higher frequency. Meanwhile the B-52's IFF signal and raw radar paint (probably nearer 3 GHz) were both apparently detected without difficulty on the ground; and the B-52's radar also continued to paint a constant ground echo at 3cm (10 GHz), because attenuation at these wavelengths would be in the order of 10^10 less than at one metre and therefore negligible.

If the asymmetry between the two radio responses remains too much of a problem to be explained by a power differential and a "one-way radio mirror", one other possibility that might be investigated is icing-up of the UHF transmitter power relay. Unlike some other types of relay failure this could be spontaneously self-correcting. Radiosonde temperature readings given in the file (for Glasgow, Mont., Table 4) show 1.0 degree C above freezing at the highest level recorded, 11,500 ft above terrain. Subfreezing temperatures are very likely at higher levels of the B-52 approach. However one would have to demonstrate the unusual vulnerability of this crucial power relay to icing (unlikely given the sealed components used at this date); explain why no other systems on the plane appear to have been affected; and explain the fact that the relay had been operating perfectly normally for some time at FL200 and higher, and would be expected to be warm, which seems inconsistent with sudden freezing in the middle of a transmission just when the B-52 started descent. In summary, there seems to be no wholly satisfactory explanation of the UHF failure.

However Claude Poher has proposed a very interesting hypothesis to explain the asymmetry of this situation. He proposes that a region of weak air ionisation is associated with the presence of the unidentified object, just below a threshold electron density at which it would begin to noticeably affect UHF transmission and reception. However each time the copilot depresses his transmit switch to energise the radio transmitter the electric field strength around the antenna is increased and free electrons are shunted along the field lines. Obviously the energy used in this redistribution of charge is energy stolen from the transmitter power, so the output radio signal is effectively 'shorted'. It's possible that this could neatly explain why only air-to-ground UHF transmissions were affected.

8 The RAPCON tape transcript contains the following message to the B-52 timed at 0852 CDT by the tower clock (03:52Z): "The UFO is being picked up by the weathers radar also, should be at your 1:00 position 3 mile now." The aircraft responded that they "do not have anything on airborne radar and we are in some pretty thick haze right now and unable to see out that way." The time of this message would be before the 30/180 turn onto the TACAN initial approach fix and so prior to the start of the ASB-9 radar episode discussed here. The file shows that Lt.Marano from FTD made several attempts to get further information from Col.Werlich about the weather radar but was either rebuffed or ignored. It has not proved possible to find information about the characteristics or location of the weather radar involved. Generally one would expect this sort of radar to operate at X-band (similar to the 3 cm airborne radar) or S-band (up to 10 cm). Weather radar is by no means generically inferior to other types of radar, requiring good accuracy and resolution, and most importantly radar height indication as well.

Another puzzle is a report of a rapid unidentified echo on the B-52 gunnery radar. According to the gunner, at some point during the ASB-9 radar episode a target was picked up on the scope of the rearward facing AN/ASG-1gun control radar. It was aft of the aircraft 30 degrees to port of the centre-line (i.e., at about 7 o'clock, behind the left wing) and moved from 1000 to 12,000 yards range in a few seconds (less than 10). It was a "brilliant target". The gunner was impressed by its size and speed, and noted that there was no clutter or other echoes on the scope. Little information is available presently about this radar, but it appears to have been a multilobe tracking radar employing sum-and-difference circuits for ranging and and target following. Probable wavelength would be under 3cm. Azimuth coverage was 320 degrees, 160 degrees left and right of the tail. But there is no reference at all to this incident in the official file and other witnesses have no memory of it.

There is also uncertainty about the role of ground radars in addition to the weather radar. Col. Werlich states that RAPCON at Minot AFB did not detect the UFO at any time, but he observes that "IFF equipment was operating in the airplane. It’s a fairly good size blip. Every time it sweeps it shows the blip. The object would have been covered by the blip." It is unclear whether he is referring to the surveillance radar or the precision approach GCA radar or both. The ADC radar south of Minot "do not remember" seeing any unidentified targets says Werlich, but this is a fairly meaningless statement. FTD's requests for more information on this, and on the RAPCON radars, were no more successful than their requests for details of the weather radar report. Werlich's only response is the inaccurate reassertion that only the B-52 bomb-nav radar was involved. This is very unsatisfactory. Werlich himself indicates that he was denied the technical assistance he requested from SAC to further his investigation, and in the context of SAC's sensitivity about some security implications of the incident this is certainly suspicious.

It has been possible to find the following information on radars operational the the Air Defence Command SAGE [Semi Automated Ground Environment] radar site 16 miles S of Minot in October 1968:

FPS-26 Height Finder
2 x 2.5Mw
pw 4.5mS
pps 333-328?

Made by AVCO. Dual Channel at 2.5 Mw each or one channel at 5Mw. SAC 42A Klystron.

FPS-26A Height Finder
2 x 2.5Mw
pw 4.5mS
pps 333-328?

Made by AVCO. Said to be similar to FPS-26 with extra ECCM features.

FPS-27 Air Surveillance
pw 6mS
pps 333
2322- 2670 MHz

Made by Westinghouse. Stacked beam system using 10 vertical beams.

Note the "diversity operation" of the heightfinders - two transmitters of 2.5MW each are operated in tandem with the pulse repetition rate slightly out of phase and the delayed signals recombined at the receiver to give an effective 5MW peak power.

All these radars were also part of the ADC "Frequency Diversity Radar Program". Frequency diversity means that as well as being slightly delayed, multiple channels are assigned slightly different frequencies - usually about 5% or less - which increases probability of detection by receiving different scattering patterns from a single target. The SNR increases like the square root of the number of channels and the practical range performance similarly. Whether this freq diversity applies to all 10 of the surveillance beams isn't certain. The vertical stack probably also gives the FPS-27 a heightfinding capability as well.

Note that the range dimension of the resolution cell, in the surveillance set in particular, is quite poor at 6 microsec., about 0.5NM (0.38NM for the heightfinders). No information is available on beam width. The surveillance peak power of 15MWatt converts to a mean power of 30kW, so this is a powerful transmitter designed for long range. The unambiguous range allowed by the interpulse time would be about 240NM.

A target at the altitudes indicated near Minot should have been above the ADC radar horizon out to a range of well over 100 miles if the ADC antenna is about the same height ASL.

9 For completeness we should note that rocket missiles conceivably could exhibit these rates, and the area was dotted with Minuteman ICBM silos. But it is hard to think of a reason why missiles would be in flight in this area at this time - barring accident or mischief with potential consequences so serious that one can scarcely imagine the cause remaining undiscovered - and of course the extended Phase B episode does not remotely suggest a missile. None of the witness testimony on file is remotely consistent with missile launches or impacts.

10 Complete NCDC raw insonde dataset for 0000 hrs and 1200 hrs, Oct 24 1968, Bismark, ND:


Bibliography ››