Summary
 

In a discussion with a NASA aerospace engineer familiar with the space shuttle reaction control system, I learned that the thrusters never generate any light while operating, but they always emit a small cloud of unburned propellant just before the thruster fires and a much larger cloud immediately after the thruster shuts down. The post-burn cloud may be visible, but only when reflecting sunlight. The pre-burn cloud is never visible to the human eye but might be detected by a light-sensitive camera. Any light flashes seen in space shuttle videos cannot be from a thruster unless they coincide with the beginning or end of a rocket burn. The consequences of this information in regard to two videos of apparently anomalous objects taken by shuttle video cameras are described. 
 

Introduction
 

As described in previous articles here and elsewhere, several objects in the STS-48 video of Sept. 15, 1991 seem to react to a flash of light by changing course. According to James Oberg and others associated with NASA, the flash of light was caused by the firing of a small reaction control system (RCS) thruster on the space shuttle. Oberg has asserted that: 
 

The RCS jets usually fire in 80-millisecond pulses to keep the shuttle pointed in a desired direction, under autopilot control (usually once every few minutes). These jets may flash when they ignite if the mixture ratio is not quite right. Propellant also tends to seep out the feed lines into the nozzle, where it accumulates, freezes through evaporative cooling, and flakes off during the next firing. The ejected burn byproducts travel at about 1000 ft/sec. One pulse usually emits about a quarter pound of propellant in a fan-shaped plume. 
 

While I've written several articles pointing out flaws in the interpretation of the light flash in the STS-48 video as the product of a thruster firing, I had no reason to question the above description of thruster operation because Oberg had a long career as a flight officer in the Mission Control Center at Johnson Space Center, and frequently appears as an expert on spaceflight for TV news and on the lecture circuit. But it turns out his description is wrong or at least misleading on several important points. 
 

I recently had the opportunity to discuss various aspects of the space shuttle's RCS propellant supply system with a NASA aerospace engineer who was involved in the design, testing, and performance evaluation of the RCS from the nearly the beginning of the shuttle program. Unlike Oberg, this engineer observed tests of thruster firings close up on a routine basis. Most of our discussions were related to work I am doing and was focused on the ingenious system that dependably supplies fuel and oxidizer to the rockets under weightless conditions in space as well under Earth's gravity during reentry. However, he also described what happens on the business end of an RCS thruster when it fires. 
 

RCS Propellant Behavior Before, During, and After a Thruster Burn
 

All of the space shuttle's rocket engines use the same "hypergolic" propellants, meaning two different chemical compounds that ignite on contact without the need of any ignition source such as an electrical spark. The fuel is monomethyl hydrazine (MMH) and the oxidizer is nitrogen tetroxide (NTO). This is fairly common knowledge and can be found on many web sites 
 

The important fact about the RCS thrusters that I learned from the NASA engineer is that the two valves that simultaneously open to admit MMH and NTO into the rocket combustion chamber are not located immediately next to the combustion chamber walls. The valve mechanisms cannot withstand the intense heat generated by combustion(about 3500¡ Celsius). So the valves are mounted on a "thermal standoff" plate some distance from the combustion chamber and are connected to it by tubing, as shown in the diagram of a vernier RCS thruster in Figure 1. 
 
 

Figure 1: Diagram of a vernier reaction control thruster of the type alleged to have fired to cause object motions in the STS-48 video. The green regions indicate the "dribble volume," which is the section of tubing between either of the propellant valves and the combustion chamber. 
 

When the thruster starts firing, the propellants are briefly exposed to the vacuum of space after flowing out of the opened valves until they reach the combustion chamber and ignite.While exposed to vacuum, some of the liquid propellants boils off into space and then immediately freezes into "microscopic snow," as this engineer called it. In the case of the small thrusters, this happens so quickly over the short distance from the valve to the combustion chamber(about 2 inches) that the amount of "snow" generated is too small to be seen. 
 

When the valves are closed to shut the thruster down, small amounts of propellant are trapped in the tubes between the valves and the combustion chamber.The engineer called this volume of trapped propellant the "dribble volume," perhaps because it was observed to just dribble out of the thruster after shutoff during ground testing under atmospheric pressure. But in the vacuum of space, the "microscopic snow" also forms after shutoff just as it does at startup. But the dribble volume is large enough that the snow generated can be seen as a white plume in reflected sunlight. It is totally invisible without some external source of illumination. 
 

The claim that significantly more unburned propellant is expelled at the end of a thruster burn than at the beginning appears to be supported by the telemetry records for the combustion chamber pressures of the Discovery's thrusters during the STS-48 mission, which I obtained from the NASA FOIA office.Figure 2 shows a plot of the combustion chamber pressure versus time for the thruster burn proposed as the cause of the light flash in the STS-48 video. At the beginning of the burn, the pressure quickly rises from 0 (vacuum) to 110 psi, but after the thruster is shut off, the pressure stops its rapid decline and tails (or dribbles) off again to zero. This tail region evidently is the pressure generated by the unburned liquid propellant trapped in the "dribble volume" evaporating into space. The record of STS-48 thruster firings I have is for a time period of 45 minutes with many burns recorded for all of the shuttle's vernier RCS thrusters, and every single one of them has the same steep rise at ignition and the same small "tail" after shutoff. Since it is a feature of every thruster firing, an associated emission of a visible plume of propellants is most likely to be present, too. 
 
 

Figure 2: Plot of combustion chamber pressure versus time for the firing of the L5D vernier thruster proposed as the source of the light flash seen in the STS-48 video on Day 2, 21:28:19.5 Mission Elapsed Time. 
 

The engineer likened the plume of small ice particles to the smoke that pours out of the barrel of a gun after the muzzle flash. Then he added "but there is no 'thruster' flash." Quite to the contrary of what might be construed from Oberg's assertion, there is no "flash" in the sense that the propellant itself generates little if any light at all during a burn. The unburned liquid propellant can be said "flash" to a vapor after thruster shutoff, but this refers only to the rapid phase change from liquid to gas, not to light emission. While Oberg may have been aware of this fact, his description was unclear about the meaning of the word"flash" and I doubt the meaning was apparent to many of his readers. 
 

Mr. Oberg has asserted that: 
 

It should also be pointed out that as all experienced observers of shuttle TV images realize, the visible flare of these jet firings is only an occasional and sporadic feature of their actual firings, which at other times -- especially in periods of smooth, stable propellant flow -- can be invisible. 
 

The evidence of the combustion chamber pressures indicates that there is nothing at all "sporadic" about this phenomenon. The numerous photos of shuttle thrusters emitting plumes provide more evidence that they are predictable result of a thruster shutoff. I have been puzzled for a long time about why the rocket combustion gases were so easy to see when they are supposed to be nearly invisible. The puzzle is apparently solved: these photos show jets of microscopic snow at the end of the firing cycle in reflected sunlight. In Figure 3, plumes from two primary RCS rockets on the left can clearly be seen. The dull reddish glow on the right is apparently from an aft-firing primary thruster, making a total of three thrusters simultaneously emitting plumes of what must be unburned propellant – an unlikely coincidence if such emissions were sporadic rather than routine consequences of thruster firings. 
 
 

Figure 3: Photograph showing plumes of unburned propellant emitted by three RCS thrusts simultaneously. 
 

In the event that there is any instability during a thruster firing that causes the fuel-to-oxidizer ratio to be out of balance, the unburned propellant would be unlikely to form snow as it exits. According to one reference, "Hydrazine ... is technically stable to about 250 C."At the 3500 C temperatures in an RCS combustion chamber, any excess MMH fuel would be converted to a gas, decompose into simpler component gases such as water and methane, and finally exit the thruster nozzle unseen along with the combustion products. 
 

A blockage of propellant flow long enough to intermittently halt combustion and cause "snow" to be generated in mid-burn could be extremely dangerous, according to the NASA engineer. If the thruster were not shut down immediately, such a malfunction could potentially lead to an explosion and even loss of the vehicle. 
 

In the light of this new information, it seems that several things that have been written about the STS-48 video have to be reconsidered concerning the behavior of space shuttle thrusters. Another video taken during the more recent STS-102 mission also is reexamined here because of its similarities with the STS-48 video. 
 

STS-48
 

A time-lapse composite of video frames before and after the light flash is shown in Figure 4. Several objects move almost immediately when the light flash occurs. Their position at the time of the flash is indicated by the red dots in the image. The light flash and movements of the objects has been attributed to the orbiter's L5D thruster (left side firing down). 
 
 

Figure 4: Time-lapse image of objects in the STS-48 video created from a composite of video frames spaced at 1-second intervals. 
 

Since there was no serious malfunction during the STS-48 mission, if the light flash observed were from the L5D thruster, it could only mark the end of the thruster burn, not the beginning. Figure 5 shows a graph I made for an earlier article concerning this light flash. There is a faint "pre-flash" that might seem to fill the role of a very brief jet of "snow" at the start of the thruster firing, assuming the camera, which was set to for low light levels, could detect light that is too faint to be perceived by the human eye.The pre-flash occurs 0.4 seconds before bright "main flash" that seems to have caused the objects in the video to move. But the L5D thruster firing supposedly responsible for the flash was 1.2 seconds in duration, so the pre- and post-burn flashes should have been 1.2 seconds apart, not 0.4 seconds. Worse for the thruster theory, the exhaust exits the nozzle at a speed of 3500 meters per second. If it is assumed that the exhaust plume hit the objects just as the thruster shut down, the objects would have to be over 4 kilometers away from the shuttle, since that is the distance the exhaust plume would travel in 1.2 seconds before impinging on the objects. 
 
 

Figure 5: STS-48 video average frame brightness over seven seconds plotted at 1/30-second intervals. Start time corresponds to the video display clock time of 20:39:23.Objects in the video changed course at the time of the highest intensity peak in this graph. 
 

There is a third pulse that occurs roughly 1.3 seconds after the main flash that might be suspected to be the rocket's shutdown flash, but again it is much fainter than the main flash, which shouldn't be the case if the third pulse was the light from the dribble volume of fuel escaping after thruster shutdown. And there should be only one or two light pulses, not three. 
 

However, a post-burn jet of "snow" might explain one seemingly anomalous characteristic of the light flash I noted in my previous article: the three light pulses occurred during a period of elevated background brightness in the image that lasted for at least 5 seconds, which is much longer than the 1.2-second duration of the thruster firing. The jet of snow at shutdown probably travels at a much slower rate than the 3500m/sec speed of the (invisible) plume of combustion products during the thruster burn. Moving at a slower rate, it would disperse much more slowly than combustion products and thus linger as a diffuse cloud, perhaps reflecting enough sunlight to raise the background brightness for several seconds. So this one characteristic of the post-burn plume from the thruster might support the thruster hypothesis. But all its other characteristics make the thruster hypothesis seem even more preposterous than it did when I wrote the earlier article, which assumed that thrusters can "flare up" sporadically, as Oberg would have it. 
 

STS-102
 

The relationship between thruster firings and "flashes" of light reflected from unburned combustion products also has relevance to a video taken during Space Shuttle Discovery's STS-102 mission in March 2001.As in the STS-48 video, a light flash occurs in the STS-102 video and a slow-moving object abruptly changes course. This is shown in the time-lapse composite of Figure 6. Another object appears at the top of the frame and seems to pursue the slower-moving "main" object. A high-speed object also appeared to be pursuing Object M1 in the STS-48 video, but was moving too fast to be captured in the time-lapse image of Figure 4. 
 
 

Figure 6: Time exposure overlay of frames at 1-second intervals from STS-102 video. The red shows the position of the object when the light flash occurs.
 

In a previous article I noted that the time display in the Mission Control Room, which appeared briefly in the STS-102 video, indicated that the light flash occurred at least 5 seconds before the closest thruster firing, which occurred at 12:30:39 GMT.Since it is impossible for the thruster firing to be the cause of an event that preceded it in time, it could not have been the cause of the light flash – assuming the mission control time display was accurate. 
 

Another event occurred in the STS-102 video that has been suggested as support for the "prosaic" thruster hypothesis.This event was a second flash of light that occurred 1.3 seconds after the first flash. As it happens, 1.3 seconds was also the amount of time that elapsed between the start of the 12:30:39 firing suspected to be associated with the light flash and the next firing at 12:30:40.3. While I assumed this was probably a random coincidence, the argument has been made that the elapsed times are the same because the flashes came from the thruster at the beginnings of the two firings. But as previously noted, the NASA engineer said that the visible plumes of unburned propellant come at the end of a thruster firing and not at the beginning. In this case, the elapsed time between the end of the first and second firings was 0.96 seconds (12:30:39.548 GMT and 12:39:40.508 GMT) – significantly less than the 1.3 seconds between the observed light flashes. 
 

The elapsed time between the end of the first and second thruster firings does not agree with the elapsed time between the first and second light flash, assuming the visible flashes mark thruster shutoff, as the NASA engineer asserted. In that case, a rocket firing can be ruled out as the cause of the object's motion in the STS-102. 
 

Even if it is assumed that the light pulses come at both the beginning and the end of a thruster burn and that the startup pulse can be brighter than the shutoff pulse, Figure 7 shows the actual brightness curve still cannot be matched to a thruster firing. The first peak at time A (when the "main object" in the video changes course) might correspond to the startup pulse of the thruster burn and the second peak at time B might correspond to the "shutoff" pulse. But the peak at B would be 0.37 seconds into the 0.48-second burn, assuming it started at time A in Figure 7. This would indicate the burn was prematurely interrupted and that a "snow" of unburned propellant was being expelled from the thruster rather than the hot gases of combustion more than 1/10 of a second before the intended thruster shutoff. Such a premature interruption of the firing could signal a potentially serious problem with the propellant supply, such as the ingestion by the thruster of a large bubble of helium. Helium is the gas that is used to pressurize the propellant so that it will flow to the thruster immediately when the fuel and oxidizer valves are opened. A small amount of helium is always in solution with the liquid propellants, but it is never supposed to form large gas bubbles in the propellant lines. 
 
 

Figure 7: STS-102 video average frame brightness over 2 seconds plotted at 1/30-second intervals. The length of time from A to D is close to the 1.28-seconds between documented thruster burns. The length of time between A and C is the duration of the first thruster burn assumed to start at time A. The rise in brightness to the peak near B might seem to correspond to the first burn's "shutoff" light pulse but that peak comes a tenth of a second before the assumed shutoff time. 
 

Conclusion
 

There are only two cases that I know of in which objects in a space shuttle video react to a light flash with a radical change in course: STS-48 and STS-102. The information I've received from an expert in the shuttle's RCS propulsion system provides a compelling refutation of Oberg's argument that thruster firings were the cause of the objects' behavior in both cases. 
 

Another assertion by Oberg that is incorrect is that propellant seeping out of propellant lines and freezing in the nozzle is a routine occurrence on the orbiter. To the contrary, it is an indication of a potentially serious leak according to the NASA engineer. If such a leak is detected the caution and warning system notifies the crew that a "Fail-Leak" condition has occurred. The leaking thruster is then isolated by shutting valves upstream of the leak. There is no record that I know of that any such failure occurred during either STS-48 or STS-102, at least during the time when the videos were taken. Based on Oberg's statements, it had seemed to me that in both the STS-48 and STS-102 videos the high-speed "projectiles" seemingly pursuing slower-moving objects might be explained as ice chunks expelled from the nozzle during the rocket burn. But as the presence of such ice in the thruster would indicate a possibly serious problem with the shuttle propulsion system, this explanation no longer seems feasible. 
 

Finally, Oberg stated that the speed of the RCS rocket exhaust gases is about 1000 feet per second. Their actual speed is 3500 meters or 11,482 feet per second. That is ten times faster than the speed he cited. This error is of no great importance to the question of apparent anomalous objects in shuttle videos. But it is one more indication that while Oberg may well be an expert on many aspects of space flight, he evidently has no particular expertise or experience with the RCS propulsion system. 
 

While the NASA engineer I spoke with was a legitimate authority on the space shuttle's RCS rockets, it should be apparent from reading this article that I have not simply made an appeal to authority here just because what he said severely undermines the "prosaic" explanations for these space shuttle videos. Instead, I've tried to verify his assertions to the extent feasible by checking independent references and data sources such as the RCS combustion chamber pressure records for STS-48. Everything I've found was consistent with what he told me. 
 

The brightness curves for both the STS-48 and the STS-102 videos do not match what would be expected for reflected light from unburned RCS propellant if it is assumed that the light flash comes only at the end of a thruster burn as asserted by the NASA engineer I discussed the question with. Even if it is assumed that a light flash can be detected at the beginning of a thruster firing by a light-sensitive camera, the brightness curves in both videos still do not match what would be expected for a thruster burn.