Two thirds of all Mars missions have been failures.
I have spent a lot of time rumaging through the investigative reports. Not once, in any of these records from Mars 2 to Beagle 2 did I find any evidence anyone had completed an analysis of all Mars missions for a common failure mode. Not once did anyone even whisper, Newton might be wrong.
Bad Engineering, Bad Physics, or both?
Mars 2 Mars 3 and Mars 6 (Soviet missions)
Both probe insertions were steeper than planned a problem blamed on Software:
On November 21, 1971, the space automatic positioning system was used to make the third correction to the trajectory of the Mars 2 spacecraft trajectory. The Mars 2 spacecraft lander was directed to the planet. At the same time, the spacecraft was placed in a Martian orbit with a pericenter height of 1,350 kilometers and a period of rotation of 18 hours. This operation was not successful. Apparently, after the second correction, the Mars 2 spacecraft trajectory had been close to the predetermined trajectory.
Nevertheless, the onboard computer issued a wrong command to decrease the height of the pericenter of the flyby hyperbola. As a result, the lander entered the Martian atmosphere at the big angle and hit the Martian surface before the parachute system was activated.
A third correction to the Mars 3 spacecraft trajectory was made on December 2, 1971. The lander separated from the spacecraft and entered its predetermined trajectory for a Mars encounter. After 4 hours and 35 minutes, the lander entered the Martian atmosphere and landed at a location with coordinates 45¡ S and 158¡ W.
The spacecraft was placed in a Martian satellite orbit with a pericenter height of 1,500 kilometers and a period of rotation of 12 days and 19 hours. The predetermined period of rotation was 25 hours. The discrepancy between the real and the predetermined periods of rotation could again be explained by the time limitations, which did not allow for proper testing of the computer programs developed for the space automatic positioning system.
Mars 6 apparently crashes after successful entry, reason unknown:
Viking AnomaliesAn analysis was performed after the flight and showed that the Mars 6 craft landed in the vicinity of the Valley Samara, which was characterized by a V-shaped crosssection profile. The coordinates of the landing site were 23¡ 54Õ S and 19¡ 25Õ W.Perhaps the landing occurred on a steep hill?
...9% margin of error in Mars gravity - kinda wide, butAttitude rates observed on both VLCI (Viking 1) and VLC2 (Viking 2) in both pitch and roll exceeded the predicted maximum rates by a factor of 2 after maneuvers and before 0.05 g.
Predicted pitch –0.85 Observed: -1.81, -2.08
Predicted yaw 0.85 Observed: -1.09 -2.07
Both vehicles trimmed at higher negative angles of attack than expected and they had higher L/D than planned. Basecover heating exceeded expectations but was acceptable. Real Gas effects caused several parameters to vary from predictions based on wind tunnel tests using air, but these variations did not have a significant effect on the missions.
Dynamic pressures derived from the flight-measured stagnation pressure differ significantly from those derived and used in LTARP. The LTARP values are based on accelerations measured in flight and aerodynamic characteristics, particularly the axial force coefficient, (CA) determined from wind-tunnel tests using air.
Flight data points show sizable percentage errors at the higher and lower extremes of the altitude range…It is in the low range area of operation that the unexpected discrepancy occurs.
…the angles of attack and lift-to-drag ratios for both VI an dVL2 were greater than predictions…The causes of higher trim angles of attack on both vehicles cannot be positively identified but are believed to be some combination of off-nominal cg position and out-gassing/ablation effects.
P37 Because flight trim angles were higher, a slight reduction in Ca would be expected. However, the figure shows that flight axial force coefficients were significantly higher . This was particularly true early in the entry when the Mach number was greater than 10.
…Very early in the entry near the 0.05-g point, accelerations are very small and the accuracy of the force coefficient calculation is poor. No obvious explanation is available for the high Ca values later in the entry between Mach 2 and 5.
…during the early high-speed portion of the entry, the pitch stability was about 505 greater than the prediction derived from the air wind-tunnel tests. This is a real gas effect and was expected and discussed in Tef w.
p38 – the Acceleration Vector inclination was expected to be ~ -1.0m but it was measured at –1.12 and –1.13
P47 Data from the entry of the first capsule showed a peak temperature on the fiberglass inner cone that exceeded expectations. Further, the temperature sensor on the aluminum outer cone failed well before it attained its peak value.
Expected Temperatures Observed (F?)
369 inner 406 394
416 outer lost 498
p57 …somewhat higher drag obtained on Mars is reminiscent of the drag increase that occurred at low speeds in the Earth flight test.
Expected drag: 0.68
Observed drag 0.74
P59 All criteria were met except the time at CVD setting. The longer times were caused by the higher CVD phase ---approximately 63-64 ft verses 55 ft targeted.
Time at constant-velocity descent (CVD)
Criteria: 5.9 to 7.9 Experienced: 8.1, 7.9
P 69 One phenomenon, common to both landers, was that the constant velocity descent phase was approximately 1 sec longer than anticipated, which converts into about
2.4 lbs of terminal descent fuel. Because we had about 30 lbs of fuel left over in both missions, this was of no consequence.
On both missions, approximately 8 ft extra were spent on the constant velocity descent phase over nominal. This means that either there was a bias in the REA at the end of updates or there was a navigation error accrued between 135 and 63 feet. Examination of the TDLR, velocity navigator, RAE and the altitude navigator rates at 135 ft show no reason to believe there was a navigation error in the altitude rate estimate. Thus the cause must be in the RAE bias error.
As both landers were about 8 ft high in the CVD start altitude, the error seems systematic Three possible causes exist. The bias error compensation in the software is computed as follows:
P103 Descent performance was normal for both landers until just before touchdown, when some channels showed an increasing indicated velocity.
P108 An anomalous increase in velocity indications occurred just before touchdown on both landers….Analysis and previous tests showed that material displaced over a smooth surface would not cause measurement errors. The displaced material, as seen on the landing pads and high-gain antenna, appears quite cohesive and seems to contain a relatively high content of magnetic material. Both conditions can contribute to a relatively high radar cross section. (Radar error)
128…It can be concluded that thrust levels were higher than expected, and for VL2 the end of de-orbit burn thrust level was slightly above PD specification limits. Frpm acceptance test data VL1 engines were nominal and VL2 wer averaged 1% higher than nominal. The flight data indicate a 5% difference rather than 1%…
A final site-dependent bias was observed
in the last 35 days of tracking data between the DSS 65 ranging passes and the 15 and
45 passes. The cause of this last bias was unresolved before landing and a set of s i t e -
dependent range biases was added to the final OD solutions to account for it. The apriori
sigma on the DSS 15 and 45 bias parameters was set at 7 RU; while the DSS 6 5
parameterÕs sigma was set at 30 RU. The estimated value of the DSS 65 bias in these
runs ranged from -15 to -20 RU relative to the DSS 15 and 45 passes.
ft V1 263.319 259.636 error 3,683 ~ 14%
ft V2 263,878 257,666 error 6,212 ~23.6%
P151. Mars gravity 12.20099-12.24428 ft/sec^2 0.354%
Errors in Viking Lander Atmospheric Profiles Discovered Using MOLA Topography
Withers, Paul; Lorenz, R. D.; Neumann, G. A.
NASA Center for AeroSpace Information (CASI)
Lunar and Planetary Science XXXIII, LPI-Contrib-1109 , 20020401; April 2002
Each Viking lander measured a topographic profile during entry. Comparing to MOLA (Mars Orbiter Laser Altimeter), we find a vertical error of 1-2 km in the Viking trajectory. This introduces a systematic error of 10-20% in the Viking densities and pressures at a given altitude. Additional information is contained in the original extended abstract.
Pathfinder ParachuteFlight reconstruction of the entry using MPF flight accelerometer data revealed that Pathfinder decelerated faster than predicted based on the estimated value of the MPF parachute CD of 0.50; a value which was determined from low altitude Earth flight tests and wind tunnel data during the development of the MPF parachute (see Ref. 3). An explanation of this underperformance of the MPF parachute system from that which was predicted is still not known.
The next major activity was cruise stage separation, which occurred at 9:30 am, or about 30 minutes prior to entry. Peak atmospheric deceleration occurred two minutes after entry, when the vehicle sustained a peak load of about 16 g’s. Parachute deployment occurred at an altitude of 9.4 km, 134 seconds before landing. Although the spacecraft carrier signal was not detected during peak deceleration and parachute deploy, it was reacquired at Madrid shortly thereafter. Heat shield release, lander bridle separation, and radar altimeter acquisition all occurred at the expected times (relative to parachute deploy). The radar altimeter acquired data starting at about 1.6 km from the surface and maintained lock through rocket ignition. The descent rate on the parachute was somewhat higher than expected, but well within the design envelope.
The rockets were ignited about 98 m above the ground, and lander separation occurred at an altitude of 21 m. Landing occurred at 9:56:55 am PDT at a vertical impact speed of 14 m/s and a peak initial deceleration of 18.6 g’s. At least 15 subsequent bounces occurred before the lander came to rest on the base petal about 2 minutes later.
The spacecraft carrier signal was not observed during the bounces, but was obtained again after the lander came to rest. The airbags were then retracted and the lander petals opened. The EDL phase ended officially at 11:34 am, when the petals were fully deployed and the lander moved on to the surface phase.
Mars Climate OrbiterFlight reconstruction of the entry using MPF flight accelerometer data revealed that Pathfinder decelerated faster than predicted based on the estimated value of the MPF parachute CD of 0.50; a value which was determined from low altitude Earth flight tests and wind tunnel data during the development of the MPF parachute (see Ref. 3). An explanation of this underperformance of the MPF parachute system from that which was predicted is still not known.
The more detailed Mars Climate Orbiter report is no longer available on the network, but I think I can salvage my case from this summary:
MCO Mishap Investigation Board
That sounds all well and good, but as always, the devil is in the details:During the 9-month journey from Earth to Mars, propulsion maneuvers were periodically performed to remove angular momentum buildup in the on-board reaction wheels (flywheels). These Angular Momentum Desaturation (AMD) events occurred 10-14 times more often than was expected by the operations navigation team. This was because
the MCO solar array was asymmetrical relative to the spacecraft body as compared to Mars Global Surveyor (MGS) which had symmetrical solar arrays. This asymmetric effect significantly increased the Sun-induced (solar pressure-induced) momentum buildup on the spacecraft. The increased AMD events coupled with the fact that the angular momentum (impulse) data was in English, rather than metric, units, resulted in
small errors being introduced in the trajectory estimate over the course of the 9-month journey. At the time of Mars insertion, the spacecraft trajectory was approximately 170 kilometers lower than planned. As a result, MCO either was destroyed in the atmosphere or re-entered heliocentric space after leaving Mars’ atmosphere...
Root Cause: Failure to use metric units in the coding of a ground software file, “Small Forces,” used in trajectory models Contributing Causes:
1. Undetected mismodeling of spacecraft velocity changes
2. Navigation Team unfamiliar with spacecraft
3. Trajectory correction maneuver number 5 not performed
4. System engineering process did not adequately address
transition from development to operations
5. Inadequate communications between project elements
6. Inadequate operations Navigation Team staffing
7. Inadequate training
8. Verification and validation process did not adequately address
The Climate Orbiter had three navigational solutions: Only one of the three contained the unit conversion error (The Earth-based calculation.) All three solutions predicted a safe orbital trajectory. The final 24 hour solution is most telling:Throughout spring and summer of 1999, concerns existed at the working level regarding discrepancies observed between navigation solutions. Residuals between the expected and observed Doppler signature of the more frequent AMD events was noted but only informally reported. As MCO approached Mars, three orbit determination schemes were
employed. Doppler and range solutions were compared to those computed using only Doppler or range data. The Doppler-only solutions consistently indicated a flight path insertion closer to the planet. These discrepancies were not resolved.
The twenty four hour solution integrated the latest ranging data. Where the orbiter had been was known and the small force error during the final 24 hrs should be insignificant. The graphs estimate the entry periapse altitude was 54 km.During the 24 hours preceding MOI, MCO began to feel the strong effects of Mar’s gravitational field and tracking data was collected to measure this and incorporate it into the orbit determination process. Approximately one hour prior to MOI, processing of this more accurate tracking data was completed. Based on this data, the first periapse altitude
was calculated to be as low as 110km. The minimum periapse altitude considered survivable by MCO is 80 km.
Why were all three solutions off by a factor of at least two and since conversion error involved only one of the three solutions, why was it assigned as the primary cause for the failure? What was wrong with the other two sets of calculations? Why did they all predict a safe entry? Nowhere is the reports have I found the words: “A correct calculation would have predicted an unsafe periapse”.
In fact, in the Investigative Report from the Mars Polar Lander Failure, they almost state the opposite:
In other words, even after the small forces calculation error was corrected for, there was still a great deal of uncertainty in telemetric navigation.The second unanticipated event occurred after the navigation team was augmented with additional workforce and expertise following the loss of MCO. At this point, it was discovered that the ability to model the effects of small forces (already more numerous than previously planned because of the daily slews required by the Star Camera anomaly) based on telemetry received from the spacecraft was not as good as was previously assumed.
Originally Posted by http://centauri.larc.nasa.gov/discoveryAgain, a flaw was identified and even though there is no directly supporting telemetry, the investigators, with little more than a set of blueprints to go on, are confident they have identified the failure mode. The problem with this confident posture is this: What about the Impact probes that were part of the Lander mission? Now we have four landing events, each following similar trajectories, and no less than three failure modes.While the probable cause of the loss of MPL (premature trigger of touchdown sensor) was identified as a potential failure mode by this fault-tree analysis prior to EDL, the description of the software design and testing provided at that time by LMA did not leave any concerns in the mind of the MSST.
From the same report:
Based on this paragraph alone, it is difficult to see how the investigative team became so certain of their analysis. (Ironically, after reading the entire analysis, I agree with their conclusion, though I would have worded it with much less certainty.)The decision to use pulse-mode control for the descent engines avoided the cost and cost risk of developing and qualifying a throttle valve in exchange for a somewhat more difficult terminal descent guidance system algorithm. This introduced other risks in the propulsion, mechanical, and control areas. Although the risks in the mechanical and thruster areas were dealt with satisfactorily, the risks in the dynamics and control area were not completely retired and should have been more fully addressed through analysis and test.
If Mars is 15% more dense than Newtonian predictions, the engineering mass margin was only 1% above failure status. It is important to note that there are no corroborating flight data to support this finding, so other failure modes cannot be ruled out, including unwarrented faith in Newtonian mechanics.At PDR, the resulting MPL mass margin was only 15 percent for the chosen launch vehicle, with significant mass liens yet to retire. Given the state of maturity at that point, a prudent mass margin should have been at least 25 percent.
In reference to the Impact Probes:
It is worth noting that no common failure modes for the probes and the Lander were considered plausible, and certainly not a miscalculation due to a failed law of gravity!Unlike the case with MPL, there was no one failure mode that was identified as being most probable. However, there were four failure modes that were determined to be plausible and they are listed
below. Refer to Section 8 for a more detailed treatment of the DS2 failure modes.
Both probes bounce on impact due to unanticipated surface effects. (See Section 8.1.1,
Both probes suffer electronic or battery failure at impact (See Section 8.1.1, FLAG 2)
Probes fail due to ionization breakdown in Mars atmosphere. (See Section 8.1.1, FLAG 3)
Probe lands on its side, interfering with antenna performance. (See Section 8.1.2, FLAG 4)
By eliminating gravity as a source of mission problems without even a cursory review looking for a thread of commonality is Bad Astronomy!
More on the probes:
And the lander...The aeroshells were designed to fall to the surface, shattering on impact and releasing their probes. The probes would then penetrate the surface by as much as a meter, first separating into two parts at impact — an aft-body (which would stay at the surface) and a penetrator (which would come to rest below the surface) — connected with a flexible cable. The probes were expected to strike the surface with an impact velocity of about 200 meters per second. The aft-body was designed to withstand a peak rigid body shock of about 60,000 g’s; the penetrator, a shock of about 30,000 g’s. The aft-body could operate in temperatures from 0 to –80 degrees C; the penetrator could operate in temperatures as low as –120 degrees C.
Now that the crash site has been discovered, it will be interesting to see how closely it matches this scenario. At first glance, I would have to say it is entirely plausible, even without Newtonian gravity.At 40 meters altitude, the lander has a velocity of approximately 13 meters per second, which, in the absence of thrust, is accelerated by Mars gravity to a surface impact velocity of approximately 22 meters per second (the nominal touchdown velocity is 2.4 meters per second). At this impact velocity, the lander could not have survived.
[ur= herel]Spirit and Opportunity[/url]
Opportunity's release was adjusted one more time...Note, due to an observed dust storm on Mars just weeks prior to arrival, the targeted parachute deployment dynamic pressure was increased from the 700 N/m2 to 725 N/m2 for the “Spirit” entry and to 750 N/m2 for the “Opportunity” entry to raise the deployment altitude. This modification was made to hedge against the possibility of encountering a lower density profile than predicted which would reduce the parachute deployment altitude, and thus, the descent timeline from parachute deployment to RAD firing.
Returnin to the tech report:Analysis of Spirit's descent through Mars' atmosphere for its landing at Gusev has contributed to a decision by flight controllers to program Opportunity to open its parachute higher than had been planned earlier, said JPL's Dr. Wayne Lee, chief engineer for development of the rover's descent and landing systems.
Edit - Properly bracket direct quotesFor both the “Spirit” and “Opportunity” entries, the time of parachute deployment was later than predicted because a lower density atmosphere was experienced. Based on preliminary atmosphere reconstruction estimates, approximately an 8% lower density profile (correlating to roughly a 1-low profile) was encountered in the maximum deceleration region during the “Spirit” descent, while approximately a 12% lower density profile was encountered during the “Opportunity” descent. This greater reduction in the density profile for “Opportunity” (as compared to “Spirit”) is consistent with the observed later time of parachute deployment.
The landing locations for both “Spirit” and “Opportunity” were within the pre-entry predicted footprint ellipses.
“Spirit” landed 13.4 km downrange from its predicted landing location, while “Opportunity landed 14.9 km downrange from its predicted landing location.
The reconstruction work is ongoing in order to gain a better understanding of what transpired during the “Spirit” and “Opportunity” landings.