How did Phoenix manage to orientate itself E-W when Landing Options

[Sandro]

Hi together

During the hot phase of the EDL-sequence with thrusters active, phoenix made automatically many different attitude controls. Among them was the orientation of phoenix's arm to the north. I wondered how phoenix found out his actual orientation and where north was. Since Mars does not have a strong dipol magnetic field like earth, I assume there must have been some other way to figure out the current orientation?!

I thought of maybe using the sun? But this would have been not easy at all!

Does anyone know how this was done?

Many thanks in advance!

Regards,
Sandro

[jmknapp]

I wondered how phoenix found out his actual orientation and where north was.

That pirouette was a nice flourish! At one stage during the EDL coverage it was mentioned that PHX was turning off the star tracker and going on gyros for guidance, so that would be a good guess.

[ugordan]

Are you sure propagation of IMU data is good enough to land within 0.6 degrees of E-W alone after several minutes of tumbling through the atmosphere? I'm curious as to how it was oriented as well. I thought a Sun sensor of some kind was used.

[MahFL]

I am pretty sure the IMU's contolled the position, oh and Phoenix is a she not a he......

[jmknapp]

Are you sure propagation of IMU data is good enough to land within 0.6 degrees of E-W alone after several minutes of tumbling through the atmosphere?

No!--I'm amazed any of it works. But she wasn't exactly tumbling.

[nprev]

I can see IMUs providing that degree of accuracy, Gordan. What I don't get is how they were apparently switched from a free-space reference (the Sun and Canopus as per usual, I assume) to a Martian geoid reference.

Aircraft IMUs/INUs align by sensing the rotation of the Earth given an input latitude and figuratively thinking "Okay, I have to correct the matching amount of degrees of precession for this latitude in so-and so direction to keep the platform level with respect to local gravity, so north must be that way..." and the other directions follow.

Only thing I can think of is that some sort of forced IMU alignment (or a translational database to generate another coordinate set) to a fixed set of coordinates based on spacecraft attitude with respect to the initial entry angle with respect to the surface at the time was uploaded before EDL. This would demonstrate tremendous confidence in knowledge of the spacecraft's attitude with respect to its model of Mars' shape (to say nothing of confidence in the model!) If they only got an 0.6 deg error, we must understand the shape of Mars very well indeed.

(Sorry for using the term "with respect to" so much, but it's kind of hard to talk about inertial navigation without citing what the intertia of interest is measured against as an assumed standard!)

[dvandorn]

Yep -- this is what's called a stable matrix. You can change the stable matrix out from under an inertial system without re-aligning the platform(s) and expect accuracy to remain within less than a tenth of a degree.

Remember, changing the matrix is something that is done in software, not in hardware. The platform(s) remain inertial, with the same angles relative to the spacecraft body, before and after a change to the stable matrix. The only difference is how the "body angles" are intepreted relative to the stable matrix. I would expect greater deviations based on "gyro drift" than due to changing the stable matrix.

In Phoenix's case, they had two platforms, so the gyro drift should have been minimal (and easily predictable) after the final star alignment. Regardless of the amount of movement by the spacecraft. Heck, the deceleration of atmospheric braking probably introduced more drift to the system than motions on 'chute. That's why the event triggers were based on deceleration rates and atmospheric pressure, and why final maneuvers were based primarily on radar data. The landing site stable matrix was likely a relatively crude approximation of the actual landing site/spacecraft geometry, especially in terms of exact location of the ground relative to the spacecraft.

-the other Doug

[Sandro]

oh man, didn't think the answer is going to be this "complicated"! I didn't understand everything but I think I got the main idea. If I understood you correctly they used the stars as long as possible, then switched to the IMU. The IMU registered EVERY motion of the lander and therefore always knew the exact orientation of phoenix?!
This is just amazing.....

Many thanks for your explanations!

Regards,
Sandro