< Back to Blogs

Sol 4: A good day on Mars

by Mark Lemmon

May 30, 2008 - If you’ve followed the mission closely, you’re aware we had the first instrument problem, an intermittent short-circuit in a part of TEGA, on sol 4. If you look at the images on the Phoenix web site, you perhaps have noticed some of the sol 4 images were not up to the same quality standards we had been maintaining for the first few sols. So, why did I title this post “a good day on Mars”? Frankly, any day with a tomorrow is a good day on Mars. And even with some small problems, the data we got back were spectacular: a look partly under the lander with RAC, maybe of ice; the previous sol’s SSI movie of lidar firing; the first movie of the telltale swinging in the breeze; a successful MECA optical microscope checkout. On top of that, we’ve been using manual exposures with the camera rather than auto-expose, and haven’t made a single significant exposure mistake. So I’m almost as happy as I could be with the Phoenix performance so far.

That said, we’re all learning. Phoenix is a complex robot with complex instruments. We have a different operating system than the Mars rovers, which changes the way things are commanded. With the rovers, if you want to take a picture, you issue the command to take a picture with all the parameters set to determine the picture you want. That would go badly on Phoenix, for
various technical reasons associated with VML2, message queue sizes, and such.

What we do when we want a picture with Phoenix (as SSI lead, I have a very visual bias in describing what we do) is write “blocks” essentially programs and subroutines that control the camera. In a sense, the “commands” we use share many properties with flight software, and are not just simple commands. So, we put a lot of effort every day into validation -- making sure the
cameras won’t do anything bad.

A year ago, when we were struggling to internalize all the hazards and opportunities we get with VML2, I was asked by a systems engineer how hard it could be to command SSI. “It’s a camera. Point … and shoot.” So I went home and wrote a block that takes pans. It doesn’t do anything else (like take pictures), but it goes through the motions. At each spot, though, it looks
for a plug-in subroutine that handles the “shoot” part. I asked one of the people writing blocks for SSI to do another that, with one command could take any subframe with any combination of SSI filters. His can be plugged into mine. Now I can write a command that tells SSI to take a full site pan (in super-resolution if I want) with all filters, using only two lines (plus standard block beginnings and endings). Yikes! Phoenix can only hold about 14 megabytes of data overnight, and send only 10 or 20 in a single day. Plus, it takes lots of time to take that many pictures. So we have to be very careful. That scenario is easy to check for and exclude, but many, many scenarios need to be checked.

OK, so why this digression? Well, some may have argued with my claim about not having exposure mistakes, since there 8 released images that are saturated. Mars was exposed well. Mars was supposed to be in the picture. The pictures were accidentally aimed at the RA shoulder, which is vastly brighter. Another color image was unnecessarily duplicated, and aimed too far out of the workspace, while a third image (of the calibration target) was blocked by the robotic arm.

So, we’re going back and shooting these things again. Really, this boils down to communication. We’ve practiced having people looking over each others shoulders, but the stress of doing this for real is taking more time than a dry run operation in the PIT. Things are getting smoother, and
reminders to fly exactly as we test are good. The good news is that communications have been awesome (excepting the Electra failure on sol 2). We have the chance to take more data than in any of our tests, and making up these observations will be easy. The other good news is that all the core science and characterization went exactly as planned, and the imaging problems only affected what we affectionately call “drop-in” science. This means that we lost very little, and that we are looking forward to possibly a better return from the mission than we ever planned for (but let me tell you, as an imaging guy, I’m ready to fill up as much data volume as they’ll let me).

So, truly, a good day on Mars.

Figure: The image below shows our sister camera, RAC. Congrats on the great under-lander pictures! (This image was targeted just right by the way, showing the last step in the RA unstow sequence.)