The following is a version of a paper submitted for presentation at the 25th Mars Society convention in Arizona, in late October of this year (2022).
Update 10-25-2022: this paper was presented, and well-received.
---------------------------------------------------------
There are two common issues among the many modern plans for
manned missions to Mars, that are as yet
totally unaddressed by all the probes and landers and rovers we have
sent there over the last nearly 6 decades. Those are (1) a need for a large, relatively smooth, and rather flat landing area, with a surface strong enough to support what
we send there, and (2) a source of ice
from which to make drinking water, breathing
oxygen, and even rocket propellants for
the return flight home. See Fig. 1
Figure 1 – Site Requirements For A Manned Visit or Base of
Any Kind
#(1) a need for a large, relatively smooth, and rather flat landing area…..
Only part of this can be adequately addressed by photography
and other sensing from orbit: “large”. That issue becomes less important after there
is an equivalent to GPS in orbit about Mars,
plus appropriate beacons installed at the landing sites. None of that is there now, so “large” it has to be, to accommodate landing errors on the order of
at least 5-10 km! Orbital sensing can
identify large sites.
The remote sensing enthusiasts will protest otherwise, but in my opinion being absolutely sure about
“relatively smooth, and rather
flat” is still not a possible thing to do reliably from orbit. What we are worried about is meter-scale
variations in slope, meter-scale
boulders, and meter-scale erosion gullies. At a larger scale, that is the same sort of thing that nearly
aborted or killed the Apollo 11 landing on the moon, averted only by going to full manual
control to search for an adequate touchdown site.
Some of the proposed vehicles to be landed, and/or to be flown back off of Mars for the
Earth return, are quite tall and narrow
in their proportions. SpaceX’s “Starship”
is one such. As little as 10 degrees out
of plumb can be enough to cause a tall,
narrow vehicle to topple over,
which is a guaranteed fatal explosion!
I have never seen anyone report 10 degree localized slopes from
orbital sensing data!
And having a landing leg come down on a boulder, or down in a gulley, by around meter or so, is the same sort of risk, except it is actually worse. I’ve seen claims of being able to see from
orbit boulders under a meter in size,
although I remain rather skeptical of that. I have never seen anybody identify
from orbit a gulley of only a meter scale!
See Figure 2 for how these risks calculate.
Figure 2 – Topple-Over Limits From Elementary Statics
Just for an example,
let us assume a vehicle with a pad span s near 11 m and a
center-of-gravity height h near 25 m.
Let us assume the critical minimum dimension of the pad footprint d is
4.5 m. The critical slope angle for
tip-over is then just about a = 10 degrees,
which really isn’t such a large localized slope. Assuming a typical dimension x = 1 m for
boulders and gullies, then for that same
vehicle, the resulting angle b is also
just about 10 degrees. Separately, these are critical. Combined:
even smaller is fatal! Could that
combined problem happen? Absolutely!
So, these are very real
risks! And, on-site photography from a rover can
identify them, and precisely locate them. That right there should tell you something
about the need for this proposed mission!
Most of the successful landers on the moon and on Mars have
had a span between the landing pads that was larger than the height to the
center of gravity of the vehicle. That
pretty much eliminates the topple-over risk,
but it is not a feature of the SpaceX “Starship”, nor any of the “Earth return vehicle”
proposals I have seen from other companies.
Those are all tall and narrow.
…..with a surface strong enough to support what we
send there
Regardless of the topple-over risk, there is also the risk of a landing
leg-and-pad stabbing into soft local regolith and burying that landing pad (or
pads). Usually, landing pad size is quite limited, and the static pressure exerted by any pad
upon the local soil is the vehicle weight divided by the total pad area. For the dynamics of touchdown, the usual engineering practice is to at
least double that static pressure! If
the vehicle in question is an Earth return vehicle, you have to use the much larger
fully-fueled weight for takeoff, but
you need not double it for the dynamics of touchdown. There is another factor of 2 needed for
modeling the effects of coming down out-of-plumb, so that one leg hits first. The higher factored pressure applies only to
that pad.
If this exerted pressure at any given landing leg exceeds
the bearing failure stress for the soil,
the leg and pad will sink into the soil,
and be buried. If all legs exceed
this critical failure pressure, all will
sink into the soil, and the pattern of
it will always be uneven, leading
also to a vehicle sitting out-of-plumb!
If tall and narrow, such soil
failure could then lead to topple-over,
and the resulting fatal explosion.
If the vehicle is an Earth return vehicle, pads sinking into, and buried by, the local soil are also a very serious problem
for takeoff. There is both the weight of
the soil atop the landing pad adding to the vehicle weight, plus the added friction force required to
extract the leg-and-pad buried in the soil.
Summed over the landing legs,
these can be large numbers,
comparable to the takeoff weight of the vehicle! Digging the legs and pads out, is thus a required activity, and one with serious associated risks.
The problem on Mars is that well over 90% of the
planetary surface seems to be fine, wind-blown
dry sand, sometimes with rocks in
it, and sometimes not. The critical failure strength of similar
materials here on Earth is about 0.25 to 0.30 MPa, with the maximum “safe” bearing pressure
nearer 0.10 MPa. That difference reflects
the usual design safety factor of about 2.5 to 3 between failure and allowable.
Why is this important?
Consider the crude estimates for SpaceX’s “Starship” at Mars, that are given in Figure 3. The vehicle has a bigger payload, but little remaining propellant at
landing. It has a smaller payload and a
full propellant load at takeoff for Earth return, as shown.
Figure 3 – Sizing Landing Pads
The static weights on Mars are quite different for landing
and takeoff, but the landing must be
factored-up for the dynamics of touchdown, and for the possibility of touching
down out-of-plumb, where one
pad hits first. Such factoring is
unnecessary for takeoff. Using the
soil failure pressure leaves no margin for design error, giving an absolute minimum pad area
requirement. These pads are
quite large for “soft fine sand”, being
about 2.55 m in diameter if there are 4 circular pads, and about 2.08 m diameter if there are 6
circular pads. We have seen nothing even
close to such large landing pads on any of the “Starship” prototypes flown so
far.
There is just no way to determine the actual safe or
failure bearing pressures on Mars,
from orbital sensing! This can
ONLY be done in-situ, with some sort of
a soil tester on a lander or a rover!
Period! End of issue! Nothing like that has ever been flown! Not in all these decades! That should also tell you something about the
need for this proposed mission!
#(2) a source of ice from which to make drinking
water, breathing oxygen, and even rocket propellants
Mars is quite dry,
because liquid water is not stable at such low atmospheric pressures. There does seem to be a little moisture
content to the regolith. I have heard
various numbers bandied about; most
commonly it is asserted that there might be 1 to 3% water content in the
soil. What that means is if you dig up and
process a metric ton of regolith,
you will at best recover 10 to 30 kg of water. That a lot of effort for a rather small
return.
And here is what that really means: if you need one ton of recovered water over
the course of a year, you will have to
process at least 30 to 100 tons of regolith.
For ten tons of water, it’s 300
to 1000 tons of regolith. For a hundred
tons of water, it’s 3000 to 10,000 tons
of regolith. For a thousand tons of
water, it’s 30,000 to 100,000 tons of
regolith. If you are making propellant
for an Earth return vehicle, your water
requirement is closer to 1000 tons than it is 1 ton! So, you
must deal with that!
This is not something you do with a shovel or a small
backhoe. This is a major effort with
major pieces of machinery, almost no
matter the water requirement. Soil
moisture is not a very attractive resource.
By way of an alternative,
there are strong indications from orbital sensing that Mars might have
some large deposits of buried ice, with
at least some of these deposits located near the lower latitudes that are feasible
for a manned visit or a base. Orbital
sensing is not real ground truth,
but the odds seem good that the buried ice really is there at those
sites. What orbital sensing cannot
tell you is how deeply buried the ice is,
how thick the deposit is beneath its cover, or anything at all about the quality and
purity of that ice.
What that lack tells you is that you need some sort of a
rover at each of those promising sites. The
rover needs a drill rig capable of at least 10-20 m depth, and equipped for many repeat drillings! That’s how you get real ground truth
for the presence/absence of the buried ice,
how easy or difficult it will be to reach and extract it, and what quality it is. These properties affect drastically how much
processing is required to utilize the resource,
and also what equipment you should bring to extract it. See Figure 4 for a simple notion
of why different equipment might be needed with varying site conditions.
Without that real ground truth in-hand, it is entirely unethical to send
a crew there and bet their lives on their being able to live off that uncharacterized
resource!
Figure 4 – Different Methods For Different Depths To A
Massive Ice Deposit
What should be on the rover
The rover with the highly-capable drill can also carry a
soil tester. Probably the easiest to
implement would be a dropped-ball rig.
This rover should have a camera,
to do the photographic evaluations of local slopes, boulder sizes, and gullies,
and also to support the dropped-ball soil tester. It should probably have a push blade on
it, to move boulders off of
otherwise-nice spots. It wouldn’t hurt to have a beacon or two to deploy, to help reduce the landing error for
subsequent craft.
We already know from our experiences with the two
nuclear-powered science rovers that aluminum wheels are a bad idea for
moving across the sharp rocks of Mars. This engineering rover thing probably
ought to have treads, or at least tough
steel wheels with big lugs on them. It
will have to “go-and-do” regardless of how daunting the sharp rocks might be!
It does need to be quite powerful, especially if it is going to push boulders
around and out of the way with that blade. It does not have to be really large, just big enough to carry the drill, soil tester,
and camera, and to push fairly
substantial rocks with the blade.
Depending upon just how much site preparation work this thing is
actually expected to perform, it might
be wise to nuclear-power it.
You will note that no scientific instruments are
proposed. This is an engineering
rover, not a science rover! It is intended to do the real engineering
tasks required for a crew to safely visit any given site. None of the science rovers have ever been
asked to do that, nor have they ever been
equipped to do it.
Drill rigs we might “adopt, adapt,
and improve” for this mission
The kind of drill needed for this mission drills quite deep
(10 to 20 m) and multiple times. There
are no drills like that. There are
multiple drills being developed or proposed for lunar missions. There is even one supposed to be aboard the
ESA ExoMars mission, since delayed. NASA had one named ARADS they tested in the
Atacama desert. Most of these are quoted
as able to drill 1 maybe 2 m deep. There
was once one able to drill 5-10 m deep,
but the Canadian company that created it no longer exists. That was Deltion, and the drill they were developing was
commonly known as the “Canadrill”. See
Figure 5. This is the technology we need to resurrect, and to develop further: it needs to go deeper, and it needs enough “reloads” to do this deep
drilling several times.
Figure 5 – The Canadian Mars Drill Technology Called
“Canadrill”
The ExoMars mission may or may not ever actually fly. It has a drill rover named Rosalind
Franklin, said be capable of 2 m depth. The entire rover is said to mass 310 kg. It carries some science instruments that this
engineering rover may not need, but
would require a larger heavier drill rig to go much deeper. All in all,
by the time we upsize the drill and add the soil tester arm, plus a push blade for moving rocks, and stout steel wheels with big lugs on
them, a good guess for the
engineering rover is about 500 kg.
The dropped-ball soil tester notion
This is based on one of two test methods commonly used at
construction sites to estimate safe soil bearing strengths for adequate
foundation design. The other is a press
technology requiring more equipment.
This one is very simple, but not
as precise. That doesn’t really
matter: even imprecise ground truth is
far, far better than no ground truth at
all.
As illustrated in Figure 6, the tester is a dropped iron sphere, held to an arm by an electromagnet. The arm is articulated, and can be moved. The held ball and a camera are near its
tip, as shown. The idea is to drop the ball from a known
height, and then measure the impact
impression it makes in the soil. The
size of that impression allows you to estimate the impact force (and pressure)
that created it.
The iron ball has a known diameter D and a known mass
m, so its weight W on Mars is precisely
known. Dropped from known height H, it has a potential energy WH. The work done Fh by the average impact force
F upon the soil, depends on the depth h
of the impact impression. That work must
equal the potential energy of the ball,
so the average impact force is F = WH/h.
The diameter d of the circular impact impression allows one to estimate
its area A = 0.25*pi*d2.
Because the ball is spherical,
there is a relationship between d and h:
if you measure one, you can
determine the other. The average impact
force divided by the area is the estimate of the soil bearing failure pressure
that you desire. The safe bearing pressure
is that failure pressure, divided by
some appropriate factor of safety (2.5 to 3).
With the camera on the end of the arm, you look closely at the dropped ball. The impression diameter d can be scaled in
the image from the ball diameter D, as
indicated in the figure. That determines
the impression depth h from the spherical ball geometry. And so the calculation proceeds. The arm then recovers the ball, guided by the camera, by activating the electromagnet in close
proximity.
Figure 6 – The Dropped-Ball Soil Bearing Strength Test
What the rover might actually look like
This will resemble other rovers, being a wheeled body with instruments
within, and equipped with appropriate
manipulation arms. It will be different
in that tough steel wheels are used,
which have big lugs on them for added traction, plus a sort of mini-bulldozer blade on one
end. It needs fold-out solar
panels, but these need to be movable in
order to dump off dust.
My best guess is that the drill rig should operate off one
side of the vehicle, and the soil tester
off the other side. The solar panels
should be at the other end from the push blade.
There should be spares on deck for drill bits, and spares on deck for extra drill stem
pipe, plus spare soil drop test balls. Any site beacons also go on the deck. The only instrumentation should be that for
analyzing the core samples from the drill rig. See Figure 7.
A key item for the soil tester would be to make scaling the
impact impression diameter from the ball diameter in the camera image an
automated process. That will require
programming skills that I lack. But it
would seem intuitively to be possible.
Figure 7 – What The Rover Layout Might Look Like
What the landed package might actually be
The MER missions (Opportunity and Spirit) had smallish
rovers of about 185 kg, released from a
lander that massed some 348 kg.
Together, the landed package was
533 kg, pretty close to the 500 kg rover
mass guessed above, for this proposed mission. What that says is that we need to delete the
lander that had airbags and a righting arrangement, in favor of something more like the
“Skycrane” used by Curiosity and Perseverance,
except perhaps not so complicated.
Here is the MER mass statement:
The original MER missions had a cruise stage for course
corrections, all the way to Mars arrival
(and direct entry off the interplanetary trajectory). The package at entry interface was about 820
kg. After the hypersonic aerobraking
ended, the heat shield was jettisoned
and the parachute deployed, so that the
package was (just barely) subsonic as it neared the surface. At 10-15 m altitude, solid braking rockets attached to the
parachute harness just above the backshell decelerated the package to a stop in
midair, and the lander/rover was
released at that 10-15 m altitude. It
fell to an impact attenuated by airbags,
and the lander had means to deflate the airbags and then extend petals
for self-righting. Once righted, the lander then released the rover.
What we might do here is add a few more small solid
cartridges to the braking assembly on the parachute harness, with a total thrust sized to simply hold up
the package weight, at the braking
altitude, for several seconds, after the braking rockets burn out. There is no lander, only the rover. It is released 10-15 m above the
surface, but on a cable-on-a-reel-and-escapement
mechanism, that limits its downward
travel to about 0.5 m/s. So, we need the “weight-holding” thrust for about
30 seconds worst case, to land the rover
on its wheels at a safe speed. No
lander is needed, and the rover thus can
be near 500 kg, for a package that
otherwise falls within the MER mass statement. Quite frankly,
I am surprised no one has thought of this simpler version of the “Skycrane”
landing approach.
See Figure 8 for a comparison of these landing
strategies.
Figure 8 – Comparison of Landing Strategies
What the cruise stage might be
I would suggest making the propellant tanks of the cruise
stage slightly larger, just to get a bit
more delta-vee capability out of it,
maybe closer to 0.2 km/s delta-vee.
In that way, we might get
slightly-better landing accuracy, and
spend less time traversing the rover to exactly where we want it to be. We might end up nearer 1200 kg than the MER
1063/1068 kg value, but it would be
worth it. Payloads that we can send to Mars are larger now.
See Figure 9 for a comparison of what the
cruise stage designs might be. Note that
we are roughly-estimated at slightly under 1200 kg per vehicle sent to Mars.
Figure 9 – Comparison of Cruise Stage Design Data
What launch vehicles might send this to Mars
Given a custom payload shroud of sufficient volume, one Falcon-9 launch could easily send three
of these 1200 kg probes to Mars in a single flight, for which the listed payload capability is
4020 kg. The website says the vehicle
must be flown expendably to do this. The
Falcon-Heavy is listed as 16,800 kg to Mars flown expendably, and so could carry 13 or possibly even 14 of
these 1200 kg probes in a single launch.
Both rockets are flying right now.
Launch prices are still something of a guess. If the Falcon-9 were priced at $120M per
expendable launch, then with 7 such
launches you could cover some 21 potential landing sites on Mars, with real ground truth. That’s real ground truth for a launch cost of
about $40M per site! If the Falcon-Heavy
were priced at $240M per expendable flight,
that’s some 13 sites evaluated for something like $18.5M launch cost per
site.
Folks, that’s
quite cheap!
Now, the MER program
spent something like $744 million on spacecraft development and launch, and another $336M on some 15 years of mission
operations. We won’t need 15 years
of operations with the engineering rovers. But we will need to do at least some of the
spacecraft development work again, since
we are eliminating the airbag-equipped,
self-righting lander in favor of a cable-reel-escapement mechanism, to land a heavier rover without a lander at
all. And we will be building more than just
2 spacecraft to send to Mars. Just for
the sake of argument, call it a $900M
program.
7 Falcon-9 launches cover some 21 sites. We need 21 spacecraft sent to Mars. That’s somewhere close to $43M per site for
spacecraft development and production. That
plus $40M per site launch cost is some $83M total per site for ground truth, at some 21 sites!
1 Falcon-Heavy launch covers some 13 sites. We would need some 13 spacecraft sent to
Mars. That’s about $69.2M per site for
spacecraft development and production. That
plus $18.5M per site for launch costs is some $87.7M per site for real ground
truth, at some 13 sites.
Either way,
it’s still incredibly cheap! See
Figure 10.
Saving lives sent to Mars,
and saving money at the same time?
There is NO EXCUSE not to do this proposed mission!
Figure 10 – Estimated-Cost Data Per Site For Ground Truth
As a wild guess,
assume it will take 3 years to decide which drill can be scaled up for
20 m depth, and then get that job done. Assume also that it will take 3 years to
develop the 500 kg rover with the lugged steel wheels, the drill,
the soil tester, and the
instruments and controls that make it autonomous. That can be done in parallel. Assume also that it will take 2-3 years to
dig out the MER cruise stage plans and update them for more delta-vee, as I suggested above. That can also be done in parallel.
We cannot make a 2024 launch window, but we could make a 2026 launch window. That means we start getting on-site
ground-truth data in 2027. That’s
just in time to aid anything going to Mars in 2028 or later.
Somebody really needs to get on with this war! See Figure 11.
Figure 11 – Wild Guess Program Schedule
----------------------------
A final thought: this really needs to be done on the moon, too. And, pretty much anywhere else we might go.
