Phoenix Final Descent Trajectory Options

[djellison]

If someone can convert these into CSV's or something - that would be usefull for all I'm sure.

Doug

[jmknapp]

If someone can convert these into CSV's or something - that would be usefull for all I'm sure.

Here's a CSV:

PHX EDL CSV

The fields are:

1. UTC date & time
2. spacecraft ephemeris time
3. PHX subpoint longitude
4. PHX subpoint latitude
5. PHX altitude (km)
6. PHX speed (km/s)
7. PHX acceleration (g)
8. X coordinate (IAU_MARS)
9. Y coordinate (IAU_MARS)
10. Z coordinate (IAU_MARS)
11. X velocity (km/s)
12. Y velocity (km/s)
13. Z velocity (km/s)

Note that the altitude is relative to the idealized tri-axial ellipsoid model of Mars, & since the northern plains are below that model, Phoenix lands at "altitude" -2.74 km!

[djellison]

Awesome - I'll have fun with that. Is this all DSN tracking-reconstructed do we think? If you plot Altitude-time, it's a bit \_ if you know what I mean.

Doug

[jmknapp]

If you plot Altitude-time, it's a bit \_ if you know what I mean.

I see what you mean--after the parachute deployed, Phoenix only traveled horizontally the length of the red line below:



The black line shows the approximate MRO/HIRISE line of sight.

[kwan3217]

Thank you. I'll try it in the morning. Do you know where to find the Phoenix mass model, i.e. moments and products of inertia, center of gravity, etc.? Is there an aerodynamic model specific to the Phoenix 70 deg. aeroshell including Cm, Cmq, etc. accessible on the web? This information would be in the Lockheed-Martin databook for Phoenix but they're probably keeping it close to their vest. Thanks again.

Except for reference area, the 70deg aeroshell is exactly the same as Pathfinder, which is reasonably well documented at ntrs.nasa.gov . It's the same shell all US missions have used since Viking. I remember readng presentations comparing all of them, and like I said, aerodynamically they all act identically when using the proper reference area. Unfortunately, the best curve I saw for Pathfinder was of actual Cd (only) as a function of mach number, but only along the entry corridor it actually experienced. Cd really is the most important thing. As far as Cm and such go, the center of pressure at hypersonic speeds is always a bit behind the center of mass, so it is stable.


Search for these on ntrs.nasa.gov

2007 Mars Phoenix Entry Descent and Landing Simulation and Modeling Analysis
Mars Pathfinder Atmospheric Entry Trajectory Design (Contains accel vs time curve, Cd vs mach curve, and mass, cg, and inertia tensor for Pathfinder, among other things)
Mars 2007 Scout Phoenix Parachute Decelerator System Program Overview (For aerodynamics of parachute)
Mars Exploration Entry Descent and Landing Challenges

Also, the reconstruction is incredible in its detail. The comments attached to the spice kernels claim 5ms resolution, and from what I see I believe it. At this resolution, the accelerometers act like microphones, so you can see loud events in the acceleration curves.

Attached thumbnail(s)

 

[Juramike]

Thank you. That graph is so cool...

I think I just got myself seasick imagining being at the end of a tether yo-yoing back and forth from the yank of parachute deployment.

[djellison]

Oh wow - that IS good stuff....

For excel-monkeys like me - keep CSV's coming

Doug

[kwan3217]

Is there a good way to get the actual orientation and pixel scale of the color Phoenix on the ground MRO image? If there is, it becomes possible to plot the final latitudes and longitudes of the last bit of the landing and answer the question posed by the first picture in this thread.

[gallen_53]

Here's a CSV:

PHX EDL CSV

The fields are:

1. UTC date & time
2. spacecraft ephemeris time
3. PHX subpoint longitude
4. PHX subpoint latitude
5. PHX altitude (km)
6. PHX speed (km/s)
7. PHX acceleration (g)
8. X coordinate (IAU_MARS)
9. Y coordinate (IAU_MARS)
10. Z coordinate (IAU_MARS)
11. X velocity (km/s)
12. Y velocity (km/s)
13. Z velocity (km/s)

Note that the altitude is relative to the idealized tri-axial ellipsoid model of Mars, & since the northern plains are below that model, Phoenix lands at "altitude" -2.74 km!

I can do the conversion myself but it might already be out there on the web. Is there a table of the above trajectory in the following coordinates:

altitude above a stated reference (MOLA)
free stream velocity (state if inertial or relative)
velocity angle with respect to the local horizon (state if inertial or relative)
azimuth angle (state if inertial or relative)
latitude (state if areocentric or areographic)
longitude

kwan3217 said:

"Except for reference area, the 70deg aeroshell is exactly the same as Pathfinder, which is reasonably well documented at ntrs.nasa.gov . It's the same shell all US missions have used since Viking. I remember readng presentations comparing all of them, and like I said, aerodynamically they all act identically when using the proper reference area. Unfortunately, the best curve I saw for Pathfinder was of actual Cd (only) as a function of mach number, but only along the entry corridor it actually experienced. Cd really is the most important thing. As far as Cm and such go, the center of pressure at hypersonic speeds is always a bit behind the center of mass, so it is stable."

The problem with the 70 deg. aeroshell is there are too many aerodynamic models. Martin-Marietta developed an extensive model for Viking but unfortunately that model was based upon wind tunnel results in air and NOT carbon dioxide (real gas effects were bogus). The aerodynamic model has evolved considerably since Viking. NASA Langley has the best model based upon CFD work. However there were several published variations of the NASA Langley model. There must be a canonical model that JPL believes in.

The aeroshell center-of-pressure doesn't change much when Mach > 6 [hypersonic limit] but begins to move forward as Mach number drops below Mach = 6. By the time Mach = 1.8, the center-of-pressure is almost of top of the center-of-mass and the vehicle is dynamically unstable (you need to pop the parachute before then). When Mach < 1.8, a 70 deg. sphere-cone will become statically unstable. This phenomena almost caused the Viking aeroshell to tumble. The Viking RCS was blowing lots of propellant prior to parachute deployment because the aeroshell was dynamically unstable.

The Phoenix EDL is extremely interesting in the context of trajectory software and aerodynamic model validation. Please correct me if I'm wrong but it's my understanding the Phoenix was despun prior to hitting the atmosphere and flew a lifting trajectory without RCS. MER and Pathfinder were both spin stabilized and flew ballistic entries. Viking flew a lifting entry and was 3-axis controlled but constantly banging its RCS. Phoenix should have a perfect trajectory for code and model validation. I'm sure there are lots of interesting real gas aerodynamic effects that can be extracted from the flight data.

[jmknapp]

For excel-monkeys like me - keep CSV's coming

Here's one with 5msec resolution (warning: 90,000-line CSV file):

PHX EDL CSV (4MB zip file)

[djellison]

90,000-line CSV file)

Sweet.

[elakdawalla]

Goody goody goody....keep the awesome graphs coming...I love that you can see the three individual leg deployments!!

--Emily

[kwan3217]

On the CSVs with velocity, make sure you specify the coordinate system, particularly whether it is inertial (J2000 or MME) or rotates with Mars (IAU_MARS). This is the difference between inertial velocity and airspeed (discounting wind). To get geodetic latitude, longitude and altitude, use IAU_MARS coordinates and run it through the code described here (fortran, but short and easy to translate) http://www.astro.uni.torun.pl/~kb/Papers/geod/Geod-BG.htm . The ellipsoid used with this spice kernel is a=3396.19km, b=3376.20km, centered on Mars center of mass, b axis=polar axis

Once you have the geodetic latitude and longitude, you can set up a local vertical coordinate system at each point and convert airspeed into north, east, up coordinates. Also, from the inertial velocity, you can do a numerical differentiation (acceleration=change in velocity/change in time) which is what I did to get the parachute graph above. You know the distance to the center of Mars and can therefore subtract off gravity and you are then left with non-gravitational accelerations generated by such things as drag, lift, engines, and resting on the ground. This non-gravitational acceleration is what the accelerometers feel. Also since you know the inertial velocity it is possible to set up a lift/drag local coordinate frame at each point to get lift and drag acceleration. With the inertial attitude provided by the attitude kernel, you can get angle of attack in both directions (and confirm that the entry vehicle is not spinning) and with the mass of the vehicle, you can get lift and drag force as a function of velocity and angle of attack along the entry corridor. You still need a Mars model atmosphere to get all the way to Cd and Cl, and this cannot be derived solely from the kernels.

So yeah, this data is enough to do some interesting aero models.

Event times, based on acceleration transients in most cases

CODE

Time from start of SPICE kernel, Event
0, SPICE kernel starts, 2008 May 25 23:30:57.920 UTC SCET, 3522.2km from center of Mars, (Entry interface, 125km above spherical reference surface)
15.202, 125km above ellipsoid (23:31:13.122 UTC SCET) (Interpolated)
122.955, Peak deceleration, 84.2403m/s^2
227.825, parachute firing
228.935, first peak parachute deceleration, 82.962m/s^2
242.825, heat shield jettison
252.985, leg deploy
253.485, leg deploy
253.980, leg deploy
404.940, lander separation
405.516, First thruster pulse, low thrust
408.005, First thruster pulse, high thrust
429, transition to constant velocity
431, constant velocity achieved, ~2.65m/s down
446.005, peak of touchdown transient, 52.666m/s^2, 23:38:23.925 UTC SCET, 430.804s after entry interface
452.860, SPICE kernel ends, 2008 May 25 23:38:30.780 UTC SCET

Edit: Entry interface is 125km above a spherical Mars reference surface, or 3522.2km from the center of Mars. I originally had entry interface 125km above ellipsoidal surface. It turns out that the spice kernel starts within 5ms of entry interface as properly defined. It is obvious that the kernel is intended to start exactly at entry interface.

[Juramike]

From the parachute deployment, the tether seems to have a natural 0.7 sec oscillation wavelength. Looking at the leg deployment on the acceleration graph above, it almost seems that the leg deployment sequence timing might've augmented the oscillation and caused more of a yo-yoing than necessary.

Kinda like when you pump your legs on a swing set to go higher.

Probably not a big deal, but a future landing might could be made smoother with a different leg deployment delay to dampen the oscillation - like 0.3 sec.

-Mike

[fredk]

fortran, but short and easy to translate

Who would ever want to translate code from such a wonderful language as fortran?

Thanks for your work on this, guys!