NASA - MAVEN - updates and discussion

[LouScheffer]

I'm not a rocket scientist but trying to image the spacecraft directly does not give me confidence that they have received much tracking data 🙁

The JPL update today makes it clear that they've not received any since Dec 6th. 

The update before that said "Further, the frequency of the tracking signal suggests MAVEN’s orbit trajectory may have changed."

Let's say the spacecraft changed velocity by 0.5m/sec - that would put it more than 600 km out of place after 2 weeks.  It's not a significant leap to infer it could be 10 minutes out of place in its orbit - and that has an impact on when it goes in and out of occultation, when uplink attempts might or might not be blocked by Mars etc etc.

Without more information, the "orbit might have changed" accuracy covers a wide range.   If the signal was locked to the DSN uplink, and their "brief" contact was 100 seconds, then they would know the frequency to about 1/100 of a Hertz, which implies an accuracy of roughly 0.2 mm/sec.  Even if this velocity was not quite what they expected, after extrapolating the new orbit a million seconds (12 days) they'd be only a few km off, and imaging would have likely worked.

If MAVEN was not locked to the DSN uplink, the accuracy would be far, far worse.  MAVEN does not have an ultra-stable oscillator (USO), so it's back to a plain old crystal oscillator in the non-locked case.  The crystal in the SDST has a spec of 5PPM, or about 40,000 Hz at X band.  Even if was calibrated perfectly and recently, it varies with temperature, which would be expected to change in a spacecraft with a non-expected orientation.  As a completely wild guess, I'd imagine a 1 meter/sec change might be detectable.  Then after a week or so without contact they'd have very little idea of where to find MAVEN, as you point out.

So the lack of imaging implies not only that the spacecraft was not where expected, but that the spacecraft was also not locked to the Earth signal.  Then likely the tracking data from the contact they did have will not be of much help.  So MAVEN is indeed lost in space.

[djellison]

they'd be only a few km off, and imaging would have likely worked.

You're missing this piece from the update
" As part of that effort, on Dec. 16 and 20, NASA’s Curiosity team used the rover’s Mastcam instrument in an attempt to image MAVEN’s reference orbit, but MAVEN was not detected"

If you look at the dates of the updates put out, the dates of the MSL imaging, and add the lag of planning cadence for MSL - it's entirely plausible that the imaging attempts were of the reference (unpertubed) orbit while the analysis was ongoing of the limited tracking data to establish that it appears to not be on that reference orbit.  Put another way, with the benefit of hindsight, MSL could have been looking in the wrong place.

There's a huge ammount of public data out there to be looked at if one is motivated enough to do so.

Looking at what's been delivered to the operational folders of NAIF SPICE data ( https://naif.jpl.nasa.gov/pub/naif/MAVEN/kernels/spk/ ) one can make an assumption that HORIZONS is pulling the latest trajectory interpretation provided to NAIF ( https://ssd.jpl.nasa.gov/horizons/app.html#/ ) as both the latest individual SPK file, and the maven_orb file that HORIZONS cites have a delivery time of 2025-12-17 14:44.

One can then look up the timing of imaging the MAVEN updates describe as happening one Dec 16 ( one frame for example - https://mars.nasa.gov/raw_images/1540406/?site=msl ) and 20  (https://mars.nasa.gov/raw_images/1542473/?site=msl) and use HORIZONS to get an observation table in local Az El for the MSL site.

The Dec 16th timing ( 2025-12-16 08:22:52 UTC ) puts MVN at Az 177.807407  El -4.020712 ( below the horizon )

The Dec 20th timing (  2025-12-20 02:24:24 UTC ) puts MVN at  Az 47.514890 El -61.398422 ( further below the horizon )

And a recently posted mosaic ( https://science.nasa.gov/resource/sol-4671-right-navigation-camera-cylindrical-projection/ ) shows the horizon at Az 177 is about 12ish degrees and Az 47 is about 5 degrees.   That mosaic seems valid for looking at the distant horizon for the current rover location as the frequently updated map ( https://science.nasa.gov/mission/msl-curiosity/location-map/ ) shows the rover is currently not far from that Sol 4671 location.

All that implies that the 'reference trajectory' mentioned in today's update, and the most recent trajectory data delivered to NAIF on 12/17  are different, and diverging.  This supports the hypothesis that MSL was looking in the wrong place while MVN was still figuring out the best fit of the limited tracking data.

[LouScheffer]

they'd be only a few km off, and imaging would have likely worked.

You're missing this piece from the update
" As part of that effort, on Dec. 16 and 20, NASA’s Curiosity team used the rover’s Mastcam instrument in an attempt to image MAVEN’s reference orbit, but MAVEN was not detected"

No, I didn't miss that - this is what I meant.  The MastCam instrument has a pretty big field of view - about 20 x 15 degrees.  So trying to image MAVEN where it should be, and failing, tells you that the delta-V imparted was above a certain amount - enough so MAVEN is now out of the frame entirely.

For example, suppose MAVEN's reference orbit was 500 km from the camera when they imaged it. The field of view is then about 500 x sin(10^o) or about +- 86 km from the center.   So to get entirely out of the field of view, MAVEN must have been at least 86 km from where they expected.  If the imaging was 1 million seconds (12 days) after the trajectory change, then the delta-V must have been at least 0.08 m/s.  So they can firmly rule out small delta-V changes.

[djellison]

The MastCam instrument has a pretty big field of view - about 20 x 15 degrees. 

MastCam Right ( the one used in this imaging ) has a much smaller field of view than that

https://www.researchgate.net/publication/318223117_The_Mars_Science_Laboratory_MSL_Mast_cameras_and_Descent_imager_Investigation_and_instrument_descriptions_MSL_MastcamMARDI_Descriptions

"Mastcam-100 f/10, 100 mm, f/10 takes < 1600 x 1200 pixel images in broad and narrow band color over a field 6.8° x 5.1° at 74 µrad/pixel scale"

So trying to image MAVEN where it should be, and failing, tells you

It could also tell you that its apparent magnitude, when smeared across the sensor due to motion is below the floor of MastCam's sensitivity.  The chance of a false negative, even if MVN is in the MCAM FOV - is significant.

suppose MAVEN's reference orbit was 500 km from the camera when they imaged it.

You don't need to suppose anything - you can go figure out the exact range using the spice data and the HORIZONS system.  You can also measure it using the toolbox in EotSS ( https://eyes.nasa.gov/apps/solar-system/#/home ) or bring the SPICE data into Celestia or Cosmographia and measure it yourself there.  You'll find the range for a sensible overflight you might try and image is typically 700-1500km.

So they can firmly rule out small delta-V changes.

The delta between the the reference orbit and the most recent one posted to NAIF already shows that to be the case.

[Kaputnik]

What would the chances of imaging MAVEN using Mastcam be expected to be?

[LouScheffer]

What would the chances of imaging MAVEN using Mastcam be expected to be?

If it was where it was expected, the odds should be good.  In 2004, a photo of the sky from the rover Spirit showed a streak.  Analysis showed it was likely a comet, but also "based on its location and motion, researchers could not rule out the possibility that the streak may have been the defunct Viking 2 orbiter, which still orbits the Red Planet."

If it's plausible to see the Viking orbiter, the odds should be good the similarly sized MAVEN, if it was where they think it was.

[djellison]

.....the rover Spirit .....

Different camera optics, different sensor.

Analysis showed it was likely a comet,

Not a comet - a meteor - a shooting star.....from the paper you linked "We therefore suggest that the streak is a grazing meteor that passed 200–300 km from Spirit, with an observed travel of 13–24 km (4.0° of arc)."

the odds should be good the similarly sized MAVEN

One doesn't need to infer this based on performance of a different spacecraft 20 years ago.    You can look at the night time imaging already done this month by MastCam R in pursuit of MAVEN and look for stars in the raw JPGs already released ( here's an example image - https://mars.nasa.gov/raw_images/1542473/?site=msl ) ......I did, the dimmest star I could find via mean image subtraction to find stars out from the sensor noise was the mag ~2.7ish Iota Centauri.  There are probably a few dimmer than that - let's be gracious and go Mag 5 or 6 as the floor for a star. 

Once can also estimate the brightness of MAVEN itself. There's many places to look at equations for this....I plugged a crude estimation into a python script and got something between Mag 5 and Mag 7.  That seems reasonable given that medium sized LEO satellites will typically be Mag 4 at Earth and we must account for MVN being 50% further from the sun and further from the observer.  We also need to account for the fact that it's moving more quickly than the rotation of the planet will smear across pixels, thus MVN will be dwelling for shorter periods across each pixel.

So it's reasonable to suggest that MVN is right on the limit of what MSL M100 can detect.  It's not a slam dunk by any means, even if looking at the right time.  As I said earlier -  the chance of a false negative, even if MVN is in the MCAM FOV - is significant.

[LouScheffer]

[...]
Once can also estimate the brightness of MAVEN itself. There's many places to look at equations for this....I plugged a crude estimation into a python script and got something between Mag 5 and Mag 7.  That seems reasonable given that medium sized LEO satellites will typically be Mag 4 at Earth and we must account for MVN being 50% further from the sun and further from the observer. 

So it's reasonable to suggest that MVN is right on the limit of what MSL M100 can detect.  It's not a slam dunk by any means, even if looking at the right time.  As I said earlier -  the chance of a false negative, even if MVN is in the MCAM FOV - is significant.

Thank you for your fact-based analysis. 

Following the same train of thought leads to a brightness towards the higher end of your range.  A commonly suggested tactic for estimating satellite brightness is to find a comparable satellite which is commonly seen.  Here one might be Iridium.  It's near the distance you calculated (it's in a 780 km circular orbit) and similar in size (Iridium has 18 m^2 of solar panels, Maven about 12^m).

Iridium is expected to be about magnitude 4.2 on Earth (from The Heavens Above predictor).  Then we need factors of 1.5^2 for Mar's distance from the sun, and a factor of 1.5 for satellite size.  That adds up to about 1.3 magnitudes.  So this would predict Maven at about 5.5.  Still not great, but not as bad as it could be.

[djellison]

Here one might be Iridium.  It's near the distance you calculated (it's in a 780 km circular orbit) and similar in size (Iridium has 18 m^2 of solar panels, Maven about 12^m).

The first Iridium one I found is Iridium 82.  Which is not in a 780km circular orbit but a ~370x 490km orbit ( Perigee: 391.7 km
Apogee: 469.8 km  - see https://www.n2yo.com/satellite/?s=25467 )

A 90deg El pass - giving a close approach of 418km reports a brightness of Mag 4.2 - the number you cite ( but clearly significantly closer than 780km circular orbit )

https://www.heavens-above.com/passdetails.aspx?lat=0&lng=0&loc=Unspecified&alt=0&tz=UCT&satid=25467&mjd=61034.7778190417

By the time it's at 10deg elevation it's  at ~1500km and Mag 6.8 - 7.3

Split the difference to get ~700km and that's Mag 5.5 - adjust for 1.5 AU and you're at Mag 7.

Again - the floor of MCAM sensitivity and the likely brightness of something like MVN are probably very close.  A false negative or to be more specific, an negative result without knowing if it's a false negative or not is quite likely.

Another set of what would appear to be search images just went live -  ( https://mars.nasa.gov/msl/multimedia/raw-images/?order=sol+desc%2Cinstrument_sort+asc%2Csample_type_sort+asc%2C+date_taken+desc&per_page=50&page=0&mission=msl&begin_date=2025-12-24&end_date=2025-12-24&af=MAST_LEFT%7CMAST_RIGHT%2C%2C  ) - so far images from 2025-12-24T05:24:44.000Z to 2025-12-24T06:10:54.000Z are posted - covering 45mins.

HORIZONS reports a pass at ~Az 232 El 19 at 2025-Dec-24 06:05  at a range of ~1900km - with Azimuth going right to left - those images cover that time range.

[DaveS]

Well, they're certainly going for it today, first DSS-25 at Goldstone had a multi-hour long session for MAVEN alone and now it shifted to DSS-36 at Canberra. Both doing X-band uplinks at 20 kW.

[DaveS]

Now shifted over to S-band at the same power. Also just noticed that the 70 m dish (DSS-43) at Canberra is assigned to MAVEN.

[MickQ]

DSS-36 at Canberra is now back on X band.

[MickQ]

And S band once more.  Certainly trying hard.

[ccdengr]

And S band once more.  Certainly trying hard.

I don't know what the S-band is about.  MAVEN is X-band only AFAIK.  I don't think S-band has been used at Mars since Viking, has it?

[djellison]

And S band once more.  Certainly trying hard.

I don't know what the S-band is about.

I'd wager that's just an idiosyncrasy of whatever unusual emergency commanding the DSN is doing that's spoofing the data pipeline feeding DSN Now into reporting it as S-Band.

https://deepspace.jpl.nasa.gov/files/820-100-H.pdf (page 13) and https://deepspace.jpl.nasa.gov/dsndocs/810-005/104/104N.pdf  (page 22) suggests that DSS-35 can only transmit at 200W in S-Band, not the 20,000 watts DSN Now reports.  Both https://www.cdscc.nasa.gov/Pages/Antennas/dss35.html and https://pdfs.semanticscholar.org/18aa/4a14a04eee561210ae7c1a7a15b5929f5758.pdf suggest it doesn't support S-Band at all.

Using Chrome developer tools one can see the source xml files driving DSN Now - such as https://eyes.nasa.gov/dsn/data/dsn.xml?r=353344760 which contains

upSignal active="true" signalType="data" dataRate="0" frequency="0" band="X" power="20" spacecraft="MVN" spacecraftID="-202"
upSignal active="true" signalType="data" dataRate="0" frequency="0" band="S" power="20" spacecraft="MVN" spacecraftID="-202"


This further suggests it's a DSN Now data pipeline idio rather than actual S-Band uplink to MVN.

[LouScheffer]

[...]

Another set of what would appear to be search images just went live -  ( https://mars.nasa.gov/msl/multimedia/raw-images/?order=sol+desc%2Cinstrument_sort+asc%2Csample_type_sort+asc%2C+date_taken+desc&per_page=50&page=0&mission=msl&begin_date=2025-12-24&end_date=2025-12-24&af=MAST_LEFT%7CMAST_RIGHT%2C%2C  ) - so far images from 2025-12-24T05:24:44.000Z to 2025-12-24T06:10:54.000Z are posted - covering 45mins.

These appear to be much more sensitive observations.  Looking at the first one (https://mars.nasa.gov/msl/multimedia/raw-images/?order=sol+desc%2Cinstrument_sort+asc%2Csample_type_sort+asc%2C+date_taken+desc&per_page=50&page=0&mission=msl&begin_date=2025-12-24&end_date=2025-12-24&af=MAST_LEFT%7CMAST_RIGHT%2C%2C ) shows many thousands of stars. 

With a field of view of 6.8 x 5.1 degrees, that's a total of about 35 square degrees, about 1/1000 of the whole sky.  If you can see thousands of stars in 1/1000 of the sky, you could see millions in the whole sky.  Then according to this table, that would require seeing stars down to magnitude 10-11.

If this calculation and interpretation are correct, MAVEN should be easily visible, if present.  Of course MAVEN is moving, and the stars are not, but if MAVEN is in or close to the nominal orbit, then summing along along the expected motion should help.

[djellison]

.....shows many thousands of stars. 

Those are not stars. Those are hot pixels.   By all means try and run them through astrometry.net and have it identify which stars they are. (spoiler alert - it'll fail that analysis - they're not stars).

Infact - I'll help you - https://nova.astrometry.net/user_images/14350311#original.  I'm not going to wait for it to finish processing, I'm pretty confident it'll fail to plate solve.   (update - it did indeed fail)

If they were stars they would moved about 10 degrees in the 45 minutes between this image at 05:25 UTC - https://mars.nasa.gov/raw_images/1543451/?site=msl and this one at 06:10 UTC https://mars.nasa.gov/raw_images/1543785/?site=msl - they would have moved from bottom to top, twice.    You'll note the hot pixels don't move in 45 minutes.

Try taking all ~150 in the set on 2024-12-24 and run them through a mean image subtraction process and see what sticks (spoiler alert - basically nothing)

that would require seeing stars down to magnitude 10-11.

Definitely not. 

If this calculation and interpretation are correct, MAVEN should be easily visible, if present.

This is I think the third time I'm going to say this sentence. The floor of MCAM sensitivity and the likely brightness of something like MVN are probably very close. 

I would find it incredibly unlikely that you'll find anything but a Mag 3 or brighter star in the raw JPG images after significant processing. You'll definitely not see them in a single JPG.   If one takes the time to find likely star fields in the PDS MCAM deliveries of calibrated data, you might get down to Mag 6 or 7.

I hate to use the appeal to authority angle and I've tried to not pull that card, I've steadfastly avoided sharing privileged information and i've demonstrate purely via trivial analysis of publicly available data why the floor of MCAM sensitivity and the likely brightness of something like MVN are probably very close and you've refused again and again to take that info onboard. You've come back repeatedly with inaccurate retorts, cited incorrect scientists, incorrect instrument specifications, incorrect orbits and now you're interpreting hot pixels that don't move as stars.    I've spent significant time explaining carefully, via analysis of public data, what you're getting wrong and why and you're simply refusing to believe what I'm telling you.

So I'm going to pull the authority card out....  I was on shift on Monday 12/22 as ECAM PUL when this observation was planned. I was up late Sunday night trying to help find a viewing opportunity (this one was the first I found) for my colleagues on the MCAM team to target this specific observation before we downed tools for solar conjunction and others on the team were doing the same work and found the same opportunity.  I worked with the rest of the planning team that day to make sure after the MCAM observation was complete that the RSM was pointed safely at our conjunction change detection target as this was the last planning shift before solar conjunction starts for MSL.  I am in a position of some authority to state - the floor of MCAM sensitivity and the likely brightness of something like MVN are probably very close.

[Kaputnik]

Doug... I just want to say that, as an interested lay person, I hugely appreciate that you take the time to share your expertise with us in this forum.

People like you help make NSF a very special corner of the internet.

[ccdengr]

If one takes the time to find likely star fields in the PDS MCAM deliveries of calibrated data, you might get down to Mag 6 or 7.

FWIW, the best processing of MSSS rover camera star images I've seen is by Simeon Schmauß (see https://www.flickr.com/photos/semeion/54831807799/in/photostream/ for an example.)  This shows a limit of about mag 8, but these were 30-second exposures with stacking and AI denoising.  (And this was from M2020 not MSL, but the MSL M100 and MCZ at 100mm are not too different in terms of sensitivity.)

[djellison]

but the MSL M100 and MCZ at 100mm are not too different in terms of sensitivity.)

Yeah - pretty heckin' close.  I think MSL's M100 is f10 to M20's MCAMZ@110 is f9.5.  I think M20 would just get the nod.

Simeon's processing given the source material is off the charts good

https://mastcamz.asu.edu/galleries/zr0-1643-0812830223-519rad-n0790870zcam05201-1100lmf/?back=%2Fmars-images%2Fimages-videos%2F%3Ffrom_sol%3D1643%26to_sol%3D1643