Sawppy Issue: Portability

Sawppy’s top speed is approximately 40 times faster than the scaled top speed of Curiosity rover. Even then, it is the pace of a very slow walk and I frequently pick up my rover to walk faster. Which leads to the next problem: there’s no good way to carry Sawppy. The cool rocker-bogie suspension gets in the way of using any piece of aluminum extrusion beam as a carry handle, and even when I try to do so, Sawppy looks really undignified when carried that way. It looks better for me to slide my arms under the main body and lift. Sawppy remains upright, but my arms get tired very quickly. And the dangling rocker-bogie still keep bumping into things.

Sawppy is also pretty awkward when I need to carry further than walking distance. I built Sawppy to be roughly the same size as the JPL Open Source Rover, but more precisely I built Sawppy to be roughly the largest size that I can still carry in my car. I’m not ready to buy a larger vehicle for the sake of carrying a rover.

For walkable distances, SGVHAK rover is ferried about by a collapsible wagon. I thought that would be a good idea for Sawppy but have yet to get one of my own. And that still doesn’t solve the problem of taking up a lot of space in a car.

The real Mars rovers have a provision to fold up compactly for their journey to Mars. The tall camera mast folds down, and so does the rocker-bogie suspension. Sawppy has not duplicated either of these capabilities, and I think it might be time for Sawppy to learn to be a little more flexible for long trips. Thanks to some helpful friend-of-friends networking, I found out that the suspension folding mechanism is the “Rocker Deploy Pivot”. (Thank you, @LongHairNasaGuy) Knowing the official terminology is the key to start researching for more details. Studying the real thing aids contemplation for adapting a simplified form for Sawppy.

For walking distance trips, several options are possible. I could get a wagon like the one for SGVHAK rover, or I could build some kind of a skateboard for Sawppy to ride. I should also think about adding a carrying handle to Sawppy somewhere. So many things I can add to Sawppy, but doing too much risks making an overly complicated rover.

Sawppy Is Both Too Fast And Not Fast Enough

Sawppy was designed to be a motorized scale model of Mars rover Curiosity and Perseverance that faithfully replicates their rocker-bogie suspension geometry and articulation, which is really cool for rough terrain but overly complicated for flat ground. Sawppy also have wheels heavily inspired by Perseverance, which is designed for Mars but rather less great here on planet Earth.

Rolling around on those wheels is a signification deviation from source material. Sawppy crawls around fairly slowly, just under 40 centimeters per second. This is roughly the speed of a toddler learning to walk. But even that is far faster than scale speed. Curiosity has a maximum speed of four centimeters per second (0.144 km/hr or 0.089 MPH.) Sawppy is roughly quarter scale, so matching scale means rolling with a maximum speed of one centimeter per second, just barely faster than a snail. On Mars there’s not much of an incentive to hurry and the robot explorers can take their time. Here on Earth we have no patience for that kind of speed and hyperactive kids would quickly lose interest.

The target speed is also reflected in the suspension geometry. A Mars rover would never go fast enough to catch some air, and thus would not need the ability to absorb impact of landing. My Sawppy going 40 cm/s is not a lot of stress, but anyone hoping to build Sawppy into a high speed dune buggy will be sorely disappointed. Bob Krause of Inventor Studios coaches a FIRST Robotics team, and they investigated adapting Sawppy to be their competition robot. Given the drastically different design constraints between Mars exploration and FIRST Robotics, they concluded this suspension design is the wrong choice for a competition where speed is essential, and I agree.

Because for all practical purposes, Sawppy’s “40X scale speed” is still too darned slow. The speed is fine when I’m doing something like casually cruising a Maker Faire looking at all the exhibit booths. But if I am on a schedule and have somewhere to be, I end up picking up Sawppy to carry by hand.

Sawppy Wheel Traction Has Downsides And Upsides

Sawppy copied the rocker-bogie suspension system from Mars rovers Curiosity and Perseverance, who are destined to roam Martian terrain. But Sawppy on Earth mostly roamed flat ground instead of uneven terrain. This made the suspension largely superfluous except for contrived demonstrations, adding complexity and weight that is not strictly necessary. But I love it anyway, as a tribute to our Martian robotic explorers.

Another tradeoff Sawppy inherited from the big rovers are the wheels. Sawppy’s wheels surfaces are designed to mimic that of Perseverance rover. Grousers (raised ribs on the surface) designed to scrabble over sand and rock struggle to find grip on asphalt or concrete. This is not a problem for Curiosity and Perseverance as there is no asphalt or concrete (or carpet, or tile…) on Mars. But Sawppy struggles on made-for-human interiors.

Most wheeled vehicles on Earth use rubber tires of some type for traction. Many different varieties are available, optimized for different surfaces. Mars rovers do not use rubber tires because rubber (both natural and synthetic) would quickly break down in the Martian atmosphere. They are also quite heavy, and weight is a constant enemy for anything launched into space. Which is why Martian rover wheels are lightweight thin-shelled metal constructs designed just tough enough to handle all known Martian terrain. Unfortunately, Curiosity’s wheels have been torn up by some unknown Martian terrain. Lessons learned from engineering tests led to redesigned wheels on Perseverance that should better handle its voyage.

But none of that are of concern for Sawppy rovers here on Earth, so various rover builders have explored improving wheel traction. Chris Bond replaced the wheels with RC monster truck wheels, similar to those on JPL Open Source Rover. Steve’s Tenacity rover got some rubbery overshoe to fit over standard Sawppy wheels.

But increasing traction could also magnify other problems making them worse. With my wheels, a Sawppy rover can be tolerant of minor steering angle misalignment, the wheel will just slip sideways a bit as it rolls. With high-traction wheels, steering angle misalignment would start pulling the wheel towards (or away from) the body, twisting the suspension geometry. This has proven to be a problem with JPL Open Source Rovers and their high traction rubber tires. This is not very noticeable when driving short distances, but minor steering alignment eventually become a big problem for longer drives.

Sawppy Issue: Terrain Mismatch

Backing off from specific design issues like steering trim adjustment, I want to note a general problem with public demonstrations. Sawppy is an approximately 1/4 scale model of Mars rovers Curiosity and Perseverance, faithfully replicating the geometry and articulation of their rocker-bogie suspension systems designed for traversing Martian terrain. Down here on planet Earth, Sawppy has historically not spent much time rolling on Mars-analogue terrain. Most of the time, when I take Sawppy out for show-and-tell, my rover is rolling on surfaces like asphalt, concrete, carpet, hardwood, or tile.

Mars doesn’t have any of those surfaces. With such luxuriously flat and smooth environments, there’s not much opportunity for Sawppy to show off how nifty a rocker-bogie suspension is. I found myself manufacturing challenges out of what’s on hand, which is why instead of sandy dunes or rocky fields, Sawppy finds itself running over feet…

… and backpacks.

Fundamentally, Sawppy is out of its element on smooth terrain. Robots optimized for flat ground would not need the capabilities of a rocker-bogie suspension system, where it is far more complex and heavy than required for the circumstances. But every once in a while, a Sawppy rover gets to visit a someplace resembling its natural habitat, such as when Aussie Sawppy took a trip to the beach.

This is the kind of terrain the rocker-bogie suspension — and the wheels — are designed for.

Sawppy Issue: Steering Angle Adjustment

I haven’t heard much feedback on the next topic, this is mostly my experience with my own Sawppy. The root cause here are the output shaft angle sensors used in inexpensive servo motors such as the LewanSoul/HiWonder LX-16A serial bus servos used on my rover. For the most part, they are just inexpensive potentiometers, and they are not terribly consistent. In practice, this means steering angles on each of Sawppy’s four corner steering servos are a little different for every Sawppy driving session. I don’t know all the variables involved, there’s probably some combination of temperature, humidity, or maybe even the phase of the moon. Who knows?

One solution is to buy better components with precision angle sensors. But since an overriding goal of the project is to keep the rover affordable and accessible, I’m going to resist going down that path as much as I can.

Sawppy was designed with a way to adjust steering center angle. What I had in mind were the set screws and where they dug into the steering shaft, which would have been a relatively unmoving thing. So Sawppy steering trim is adjusted by the rover chassis configuration file that is loaded by the software at bootup. This would have been fine if it was something a rover builder only had to do once when building the rover. I did not anticipate steering would fall out of trim on a recurring basis, and requiring this file be edited for every driving session is a huge hassle.

SGVHAK rover also had a trim adjustment issue. That rover’s steering motors had relative encoders but had no homing reference point so it didn’t know where “straight ahead” was upon bootup. Since I expected it to have to be adjusted on every startup, I added a menu in software to perform steering trim adjustment. This is friendlier than a configuration text file, but it was a multi-step process that takes at least thirty seconds per corner. I still consider that too much work.

For a future rover revision, I intend to add some mechanical mean to compensate for drifting potentiometers. I want to loosen something (using screwdriver, or wrench, or even a tool-less contraption if I’m clever) to release the corner steering mechanism, steer the wheel by hand to desired alignment, and tighten things back down. Ideally I want to adjust the steering angle in under 10 seconds per corner.

The JPL Open Source Rover uses high quality (and expensive!) steering angle sensors. The sensor angles have no such drift to speak of, but the steering shafts do sometimes slip inside their steering couplers resulting again in misaligned wheels. Fortunately, these sensors are easily accessible and with a crescent wrench I could turn the sensor inside their mounted position to adjust trim. This is the kind of adjustment I want.

I have a very crude equivalent for Sawppy right now. Sometimes I have an emergency and need to adjust steering trim in a hurry. I can loosen the set screw in the steering servo coupler, steer the wheel manually, and tighten that set screw back down. This only worked because the set screws I bought from McMaster-Carr are much harder than steering shaft material, so tightening it would dig a new hole at the desired position. But I could only do this a limited number of times before the shaft is too chewed up to hold any angle and I need to cut a new shaft.

But it taught me that a mechanical steering trim adjustment mechanism is fast and easy to use. I just need to find a way to preserve that user experience in a more well-thought and more enduring design. Even if Sawppy’s suspension system is rather underutilized most of the time.

Sawppy Issue: Serial Bus Servo

I was surprised when I learned Misumi aluminum extrusions weren’t as easily available as I had originally thought. I was less surprised to learn that serial bus servos were a problem as well. When I started the project I knew they were on the rare side but thought it was worth a shot. I’m still a fan and stand by everything in my Hackaday overview, but I admit spotty worldwide availability is only the first of many problems.

Even when serial bus servos are available, the selection of sizes and capabilities is limited relative to more common motors. If someone wanted to make a small Sawppy like Dean, or a big Sawppy like Quinn, it’s hard to find serial bus servos of the appropriate size. And even if serial bus servos can be found in the right size, the communication protocol is not standardized. Dean has yet to figure out the protocol for the tiny servos he found. And then there’s the software side of the problem, where each protocol needs corresponding code written for Sawppy’s brain.

Another problem is that their position feedback has been disappointingly limited, at least in the LewanSoul/Hiwonder LX-16A servos on my own Sawppy. They could turn a full 360 degrees, but only return reliable position data for ~240 degrees within that range. That makes it difficult to calculate wheel odometry. Rhys Mainwaring tackled this problem with the Curio rover software stack, but no matter how good the extrapolation code works it’s not as good as actual data. I think my effort to do things on the cheap did not pan out and, if a robot wants wheel odometry, we have to go to real wheel encoders separate from the drive motor. Marco Walther (mw46d)’s Sawppy wheel modification sets a precedent on how this might be done.

Thanks to feedback from fellow rover fans around the world, I have since learned serial bus servos are not the best choice for a project that I intended for people to build around the world. I should fall back to more common components like RC servos and DC motors. But whatever is ultimately used as the steering actuator, I want to improve how steering angle is adjusted.

Sawppy Issue: Misumi Aluminum Extrusions

Sawppy has hundreds of fasteners to hold 3D-printed parts on aluminum extrusion beams. Managing all those fasteners has turned out to be a chore and it is something I want to improve in a future revision. While I’m reconsidering those fasteners, the extrusion beams are worth a bit of thought as well.

I designed Sawppy around extrusion beams with a 15mm square profile, chosen because it felt like roughly the right size to stay proportional to the real Mars rovers. Designing in metric was an intentional choice to make it friendlier worldwide. This was in response to chatter on the JPL Open Source Rover forums, where its non-metric parts caused headaches for builders outside of the United States.

The aluminum beams in my own Sawppy was purchased from Misumi USA. I chose their 15mm HFS3 series because the slots were cut in a way that permitted the option to use commodity M3 nuts instead of their specialty HFS3 nuts. Another reason for choosing Misumi was that they are a global company with distributors around the world, making HFS3 available worldwide.

After I published Sawppy instructions online, I started receiving feedback from builders outside of the United States and learned of a catch: while Misumi USA is happy to sell to anyone with a credit card, other Misumi distributors around the world are not so hobbyist friendly. Misumi Europe, for example, deals only with businesses and is not interested in earning income from hobbyists.

Fortunately resourceful hackers in Europe quickly found that the MakerBeam XL is a direct substitute and there are hobbyist-friendly distributors in Europe. It is more expensive, but at least it is available, and several European Sawppy rovers were built with it.

Aussie Sawppy, on the other hand, had none of those options domestically. So 15mm extrusions were ordered from China. That rover had to wait a long time for parts to arrive, but at least it is an option. There are a few countries where prohibitive import duties make even that impractical. I’ve heard from two aspiring rover builders in Argentina who will try to build a Sawppy using rectangular profile metal tubes.

I think aluminum extrusion beams are still the way to go for a Sawppy-sized project. And thanks to the availability of MakerBeam XL and AliExpress, staying with existing 15mm profile is still a good option. However, I am contemplating a move up to 20mm square profile. They’ll make the rover look a little bulkier, but 20mm is far more common worldwide than 15mm. If nothing else, 20mm extrusion beams are popular in maker circles due to a large number of popular 3D printers built with them like the Creality Ender 3. If I move up to 20mm there is a long term possibility for a rover built from chassis beams of retired 3D printers. Based on feedback I’ve received, moving away from 15mm extrusion beams and serial bus servos for Sawppy V2 would help mitigate two of the biggest problems with worldwide availability of Sawppy parts.

Sawppy Issue: Fastener Overload

I wanted keep the barrier of entry for Sawppy low. Make it accessible to rover fans worldwide. In hindsight a few of the tradeoffs I made for this goal did not pan out. Some of these were a consequence of wanting to make sure Sawppy could be built with only handheld tools. It turns ensuring proper shaft alignment requires a drill press. A drill press is also potentially useful for greatly simplifying Sawppy construction.

There are a lot of M3 fasteners used all over Sawppy. Literally hundreds of M3 bolts, nuts, and washers. The original idea was to have a system that allows for easy experimentation. M3 nuts can slide all along these extrusions, allowing rover builders to experiment with different suspension geometries and attach modifications. But in my own experience, having so many fasteners made every modification a huge hassle. Even changing one little thing might involve tens of fasteners. The tedium of tending to those fasteners discouraged me from doing as much experimentation as I originally thought I would, defeating the original intent.

There is another approach to building with extrusion beams, and that is to drill holes for larger fasteners to pass through the hole to a nut or something on the other side. There are two downsides to this approach. The first is loss of flexibility as the hole position is permanent. The second is that drilling a hole through the center of an extrusion beam requires holding the drill bit steady even as it catches on the edges of extrusion rails. I have never successfully done this with a handheld drill, I could only do it with the help of a vise holding the extrusion beam in place under a drill press.

So I avoided this construction method because I didn’t want to require people to have access to a drill press and vise. In light of my experience to date with Sawppy, however, I have changed my mind on this topic. While imposing a drill press + vise requirement would push Sawppy out of reach of some hypothetical aspiring rover builders, I noticed that almost every Sawppy builder I’ve corresponded with have such access and it wouldn’t have been a barrier for them. As for the people that don’t have access to such equipment, they still have the option of building Sawppy V1.

Given that I’m reconsidering how Sawppy components will fasten to extrusion beams, it is also a good time to think about the beams themselves.

Sawppy Issue: 8mm Hole Precision

The old mechanical adage is that anything that moves is a potential point of failure, and Sawppy is not immune. Every point of rotational motion uses 8mm steel shafts and they have proven to be challenging for myself and other rover builders. They are kept in location by E-clips, and cutting precise slots for them are hard without a metalworking lathe. For pieces that need to grip on those shafts, my designed used M3 set screws that pushes against plastic with the help of heat-set inserts. And getting them right can also be frustrating.

For pieces of plastic that don’t need to grip on the shaft with set screws, they should still have a snug fit. But 3D printing a precise diameter hole is hard to accomplish given variances of 3D printers. Not just variation from printer to printer, but that a printer may have difficulty hitting the same exact XY coordinates on every print layer to line up a precise hole.

Fortunately this is a problem with an existing solution from the world of machining. When a precise diameter is required, rotational cutting tools called reamers are used. Sawppy’s instructions included use of an 8mm reamer to make precise holes that will fit tightly onto 8mm shafts, and a reamer built for metal had no problems chewing through thermoplastic.

But I found that wasn’t the end of the story, because a reamer only helps make the diameter precise. The location is only as precise as the drill turning the reamer. I had naively thought the reamer would be guided to the correct location and alignment by the existing 3D-printed hole. It did to some extent, but not enough to ensure precision. If the drill is positioned off-center or tilted off-axis, my reamer happily reamed out an off-center, off-axis hole of very precise diameter.

What happens when the hole position or tilt is wrong? In case of suspension components, the articulation will change by a few degrees. In case of wheels, it means the center of rotation would be slightly offset from wheel center. When such wheels spin up, they will visibly wobble. Actually neither would be a functional blocker for Sawppy, as the rocker-bogie suspension can tolerate these variations and keep all six wheels on the ground. But it does look a little silly and drive perfectionists mad.

I don’t think there’s any way to solve the off-tilt problem for handheld drills. Every solution I thought of to hold the drill in alignment ends up looking like a drill press. Why reinvent this wheel? Unfortunately that’s a piece of shop equipment not every aspiring rover builder owns.

The off-center problem is harder to solve. Right now my train of thought is heading towards some kind of centering jig. The challenge is to design something that can be 3D-printed yet help deliver higher precision than what 3D-printing can deliver on its own. Can it be done? I don’t know, but I’m thinking about it.

Back on the topic of drill press: there is a second reason why I might want to make them part of the Sawppy builder’s toolkit: a drill press would allow me to drastically reduce the number of fasteners required to build Sawppy.

Sawppy Issue: Getting E-Clip Slots Right Is Hard

Closely related to the challenging heat-set inserts are the shafts their set screws bite into. I designed Sawppy to use 8mm metal shafts everywhere there is rotational motion: for wheel rolling, wheel steering, and rocker-bogie suspension articulation. Generally speaking there are three steps to fabricate Sawppy shafts:

  1. Cut to the proper length.
  2. Cut slots for E-rings.
  3. Cut flats for set screws. (For some of them.)

And it turns out the E-clips make everything trickier than it really needed to be. Unlike my ignorance with heat-set inserts, I knew E-clips are not very standardized, especially in their thickness. So my documentation couldn’t really say what the proper lengths would be, I could only give the functional dimensions and tell people to add width of their E-clip slots.

For cutting those E-clip slots themselves, I documented a “Poor Man’s Lathe” technique using two motorized tools, but this is not very precise. Fortunately, high precision is not required for a 3D-printed rover, the worst thing that happens is a rover whose suspension is a little bit wobblier than it would otherwise be.

Being a perfectionist, I was not happy with the hand-built results and arranged to cut a second set of Sawppy shafts with a manually operated metalworking lathe. And I knew imprecision would bother other Sawppy builders out there as well and not everyone could get access to a real lathe.

In the meantime, people have devised other workarounds. Ameer used M8 bolts and nuts instead. Using threaded fasteners instead of a solid shaft would put stress on the threads, but probably not terribly much for a little 3D printed rover. And it is definitely easier to work with.

The challenge with this approach is in tightening these bolts. If they are too loose, rover suspension will wobble and we’ve gained nothing. But if they are too tight, it puts too much axial load on the bearings and they seize up instead of turning freely as they should. I think there is a better way down the fastener path, and have a few ideas to test. If successful, a future Sawppy would use fasteners instead of fabricating shafts and also not unduly stress its threads or be finicky about amount of tightening torque.

I don’t like it when some part of Sawppy is finicky, but I’ve come to accept that a tradeoff has to be made between finicky precision and a design that can be built by the widest range of rover enthusiasts. This became the most apparent when dealing with holes for these 8mm shafts.

Sawppy Issue: Heat Set Insert Shaft Coupling

Looking at what other Sawppy builders have done to customize their own rovers, I learned that ground clearance isn’t as valued as I thought it would be. At first I was disappointed, because giving Sawppy great ground clearance took a lot of hard work. But then I decided to look on the positive side, because this feedback also meant I could trade off some clearance for other features and not have it be considered an instant fail.

Another common theme of feedback on Sawppy wheels concern shaft coupling and heat-set inserts. It has been a perpetual problem for me, to the point where I started preparing field replacement units of shaft couplers. I also heard from other Sawppy builders that it has been problematic for them as well.

The biggest problem was caused by my ignorance: I thought all M3 inserts would be the same size, but I was wrong. Inside, yes, but not outside. Furthermore, optimal heat set inserts usage require hole diameter within a pretty narrow range. And that range appears to be narrower than the range of variation between 3D printers. All this meant my Sawppy design printed as-is would frequently not work for another rover builder and require modification. Either their insert is a different diameter, or their printer would print differently, or some combination of both.

Some builders applied their creativity to this problem and came up with alternate approaches. Chris Bond went with the direct approach: drilling and tapping a hole through the diameter of the shaft. This is a good option if a vise and drill press is available, but drilling on a round shape is very challenging with handheld tools.

TeamSG for Aussie Sawppy explored another route: instead of round shafts, use hex profile shafts. I knew precision shaped profile beams of metal are expensive and difficult to get, so I didn’t think it was a good idea. But TeamSG thought of something I did not: Hex profile wrenches are plentiful, precise, and made of strong metals. Cutting up a few is a brilliant approach to build some rover hex drives!

Laura McKeegan’s CJ rover offered yet another approach: use captive hex nuts which worked when heat-set inserts did not. Unlike heat-set inserts, M3 hex nuts are standardized with an external diameter of 5.5mm from flat to flat.

My own experiments with captive nuts have not been successful, because if I put too much force my nuts will start spinning inside their slot. To fix this, I would heat it up with my soldering iron and turn it into a crude heat-set insert. So I thought I’d just go straight to heat-set inserts. But given feedback from builders and knowing what I know now about non-uniformity of M3 heat set inserts, I think captive nuts might be worth another look.

And while I’m working on solutions for Sawppy’s shaft coupler problems, it’s worth noting that the shaft themselves have been problematic as well.

The Less Famous Rovers Marie Curie, Dusty, and Maggie

While I’m on the subject of rover suspension and ground clearance, I want to take a detour to recognize the less famous rovers that didn’t make the trip to Mars. My rover fandom has mostly been focused on siblings Curiosity and Perseverance, but those rovers would not exist if it weren’t for the twin rovers Spirit and Opportunity. Formally the Mars Exploration Rover program, they followed Sojourner the rover technology demonstrator. Each rover generation was larger and more sophisticated than the last, but one thing all those rovers had in common was the rocker-bogie suspension design that I replicated with my own Sawppy rover using references like a rover family portrait published by NASA.

But the real rovers are on Mars. Who are the rovers in this picture? The little one is named Marie Curie, and is a “Flight Spare” for Sojourner (Truth). Meaning it is fully equipped and qualified to be sent to Mars as an alternate if necessary, but there was no need so Marie Curie stayed on Earth. I assume it played a role during the mission as a testbed as the other two rovers did.

Representing Spirit and Opportunity is the MER Surface System Test Bed. This rover duplicates many of the mechanicals of Spirit and Opportunity, but is not fully equipped to go to Mars. For one thing, it runs on a power tether and have no functioning solar panels, just passive stand-ins. This machine stayed at JPL’s Mars Yard to help scientists and engineers on Earth test ideas and solutions before commands are sent to Mars.

MER SSTB is a mouthful, and I remember seeing JPL people on Twitter calling the test rover “Dusty” but I found no official confirmation of this name. After the Mars Exploration Rover program ended, Dusty(?) was given to the Smithsonian where it will continue to represent the MER program when put on display.

The Mars Science Laboratory (MSL) mission included the Earthbound Vehicle System Test Bed counterpart to Curiosity shown in the family photo, again with some differences such as powered by a tether instead of faithfully duplicating an onboard radio-thermal nuclear reactor. According to this article, the MSL VSTB counterpart to Curiosity is “Maggie”, and the Mars 2020 test bed counterpart to Perseverance is “Optimism”.

And since these are engineering tools, both names are officially engineering acronyms. Or more likely, “backronyms” where the acronym was chosen first, then words were chosen to fit it. That’s why we have convoluted names like “Mars Automated Giant Gizmo for Integrated Engineering” and “Operational Perseverance Twin for Integration of Mechanisms and Instruments Sent to Mars.”

These rovers won’t get nearly as much fame as their Mars-bound counterparts, but they all play vital roles contributing to mission success.

[Image credit: NASA/JPL-Caltech]

Sawppy Rover Ground Clearance

When it comes time for Ingenuity to demonstrate Mars helicopter technology, it will be dropped to the surface then wait for Perseverance to move a small distance away to prepare for the experiment. This arrangement is possible because the rover’s rocker-bogie suspension gave it more than enough ground clearance for a hitchhiking helicopter to ride along, holding on to the belly of the rover.

This ample ground clearance is also visible in the Mars 2020 Mission Identifier. The first I saw of this logo was on the side of the rocket payload aerodynamic fairing protecting the rover against the atmosphere during launch. I love this blocky minimalist design of the rover’s front view and believe this little logo hasn’t been used nearly enough.

Anyway, back to ground clearance: it is something obviously useful for a wheeled vehicle traversing off-road, which is desirable given the lack of roads on Mars. But raising the height of rover’s main body is only part of the equation, because it’s not the only thing that might collide with obstacles on the ground. We also have to worry about suspension components. This is something I noticed with JPL’s Open Source Rover design: if a ground obstacle is offset from a wheel, there’s a chance it will collide with aluminum structure. This picture uses a Roomba virtual wall marker to represent a rock on the ground, which is about to collide with suspension structure:

When I designed Sawppy, I thought I could improve upon this specific issue by taking advantage of the flexibility of 3D printing. I also chose Sawppy wheel motors with the requirement they must fit within the wheel with no protrusions. If Sawppy should encounter a Roomba virtual wall marker slightly offset from a wheel, there is no risk of collision:

I this was a pretty good Sawppy feature. However, after looking at how other Sawppy builders have customized their rovers, it appears this feature isn’t as valued as I thought it would be. There have been many Sawppy modifications that did not preserve this clearance.

For example, Marco Walther [mw46d] replaced wheel motors with much longer units that risks collision. He is aware of the risk, because he designed shielding to protect the motor encoders from damage. It’s a different approach, shielding vs. avoidance. Steve [jetdillo] adopted Marco’s design to good effect.

Chris Bond swapped out the wheels for RC monster truck wheels, similar to those used by JPL Open Source Rover. However, my wheel mounting arm design does not fit within the wheel. To compensate for this, Chris decided to push the wheels out. This leaves almost the entire wheel mounting arm out where it could collide with ground obstacles. It also means the steering axis is no longer aligned with the wheel, but Chris seems happy with his design so I guess it works well enough.

My lesson is that ground clearance isn’t as important as I thought it was. Or at least, not seen as important enough relative to features other builders wanted for their own rovers. This is valuable feedback for future iterations.

On the topic of rover iterations, I want to take a quick detour from my own rover to recognize some lesser-known counterparts to our robotic Mars explorers before returning to the topic of Sawppy builder feedback.

Ingenuity the Mars Helicopter Technology Demonstrator

One of the most exciting part of the Mars 2020 mission is not visible on the Perseverance rover interactive 3D web page. It is Ingenuity, officially named the Mars Helicopter Technology Demonstrator. (It has its own online interactive 3D model.) For people like myself who want to really dig into the technical details, NASA JPL published many papers on the project including this one for the 2018 AIAA Atmospheric Flight Mechanics Conference. (DOI: 10.2514/6.2018-0023)

Ingenuity is the latest in a long line of technology demonstrator projects from JPL, where ideas are tested at a small scale in a noncritical capacity. Once proven, later missions can make more extensive use of the technology. Perseverance rover itself is part of such a line, tracing back to Sojourner which was the technology demonstrator for the concept of a Mars rover. Reflected in its official name, the Microrover Flight Experiment.

Most of the popular press has covered Ingenuity’s rotors and how they had to be designed for Mars. It has the advantage that it only had to lift against Martian gravity, which is much weaker than Earth gravity. But that advantage is more than balanced out by the disadvantage of having to work in Martian atmosphere, which is much much thinner than Earth air. Designing and testing them on Earth was a pretty significant challenge.

Mechanically, the part I find the most interesting were motor and control system for the coaxial helicopter. It has been simplified relative to coaxial helicopters flying on Earth, but still far more complex than the category of multirotor aircraft commonly called drones. Most multirotor aircraft have no mechanical control linkages at all, their propellers rigidly attached to a motor and control is strictly electronic via motor power. The paper describes the challenges of implementing a coaxial helicopter control system for Mars, but it didn’t explain why the design was chosen in the first place. I’m sure someone worked through the tradeoffs. Since mechanical simplicity (and hence reliability) is highly valued in planetary missions, I am very curious what factors outweighed it. Perhaps that information was published in another paper?

Electronically, the most exciting thing is Ingenuity’s brain, which is from the Snapdragon line of processors better known as the brains for cell phones and tablets here on Earth. If it works on Mars, it would offer a huge increase in computing power for planetary missions. Perseverance itself runs on a RAD750 computer, which has proven its reliability through many successful spacecraft but is roughly equivalent to a 20-year old PowerPC desktop. Having more powerful CPUs for future missions will allow our robotic explorers to be more autonomous instead of being dependent on brains back here on Earth to tell them what to do.

Perseverance Rover Interactive 3D Model

NASA released a 3D printable static display model of Mars rover Perseverance, which seems to have some improvements over the earlier Curiosity model. But that’s not the only 3D resource for the rover currently on its way to Mars. There is also a version designed for on-screen display rather than 3D printing.

Both the 3D printable (in STL file format) and 3D render (GLB file format) models were listed on the Mars 2020 rover page, which as of this writing has curiosity disappeared from the index page of NASA 3D resources. I’m not sure what’s going on there, but hopefully it’ll be fixed shortly.

When I listed 3D resources for Curiosity there was also a model suitable for 3D rendering. Available as download files for Blender the open source 3D graphics tool and as files embedded in a web page for interactive viewing. The latter is again available: The Mars 2020 mission page has a 3D model of Perseverance that we can interact within a web browser.

This browser interactive model is the most easily accessible version, there’s no need to install Blender or any other piece of software. It serves as an index page to many other pieces of information talking about the rover. While it has a lot of detail missing from the 3D printable model, it still has a few minor flaws. One of them I noticed only because I’ve been a fanatic of the rover: the online interactive rover’s right side wheels are reversed from the actual rover.

Perseverance, like Curiosity before it, has wheel spokes that are curved to absorb impact. I simplified the idea and translated it into a 3D-printable shape for Sawppy’s wheel. For both rovers, the direction of curvature for wheel spokes are the same for all six wheels, clearly visible in rover test footage. Shot in JPL’s vehicle assembly bay, we can see that the wheel spoke curvature is “clockwise” on all six wheels of Curiosity and Perseverance.

On the online interactive 3D model, its left side wheels match the real rover but its right side wheels had been flipped so the spokes point counter-clockwise.

It’s a tiny detail that would only be noticed by the most particular of rover fans, which I certainly am. Surprisingly, I’m not the only one! Because I’ve received questions about whether Sawppy’s wheels should be printed in mirrored orientation. Some Sawppy builders choose as I did, to have six identical wheels matching the real rover. Others chose to mirror three of the wheels as the web page interactive model did.

But one of the most exciting parts of the Mars 2020 mission is not visible at all here. Riding along with Perseverance rover is the first aircraft built by humans to fly in the atmosphere of another planet.

Window Shopping: NASA Perseverance Rover 3D Print Static Model

After finishing a model of Mars rover Curiosity, the obvious question is: did NASA release a 3D print static model of its successor Perseverance? And the answer is yes! Curiously, the web site page points to a 3D file suitable for graphics rendering, not for 3D printing. But poking around the GitHub repository revealed there’s a “M2020” 3D print model. (Before receiving its name, Perseverance was called Mars 2020.)

Armed with my recent experience, I looked over the files for this printable rover. First the good news: the rocker-bogie suspension is represented in much higher fidelity on this model. The full rocker bogie geometry is represented, including a differential bar that is designed for some bent paperclips to serve as critical linkages. The wheel tracks appear to be correct with the center pair of wheels having wider tracks than the front and rear pairs. And the geometry is more accurate, no weird right angle bends as concession to ease of printing.

The lack of concession to ease of printing is also the bad news. Unlike the previous model, none of the geometry has been modified to fit flat on a print bed. Printing this model will require, at a minimum, printing with supports which is always problematic. Dual-material printer with dissolvable supports should make such designs easy to print, but I don’t have one of those.

Another change from the previous model is that this one doesn’t use snap-together construction. Parts are designed to be glued together instead. It also means these files assume a much higher printing precision, since superglue requires a much tighter fit than snap-together construction.

If I saw these traits on a Thingiverse item, I would be skeptical that the model is even buildable. That site is littered with too many things that are obviously impossible, merely the dream someone created in CAD and never test printed.

But this one appears to be real, since among the STL files is a picture of this rover design that has been built. Looking over the print quality of its parts, it was obviously printed at high detail quality on a good printer. Likely better than mine! I think I’ll hold off printing this rover design for the moment. Maybe later, when I have a well dialed-in printer that I can trust to meet precise tolerance requirements. In the meantime, I can admire the Perseverance 3D Model released by NASA that lives strictly in the digital realm.

Built NASA’s Curiosity Rover 3D Printed Static Model

I’ve completed assembly of a 3D-printed static display model, released by NASA, of Mars rover Curiosity. It had a lot of details that were demanding when printed in PETG. In hindsight, I should have printed with PLA for fewer printing problems like stringing and overhangs. It is only a display model, it’ll just sit on a shelf and not stand out in the sun as Sawppy has done (and suffered for it.) Better dimensional accuracy with cleaner printing PLA would also help make the snap-together construction more effective. PETG is more ductile and so there wasn’t a “click” to announce successful assembly.

The demanding details were fitting for a static display model. Unlike its smaller sibling, this one is even poseable with corner wheels that steer and a robot arm that can articulate through the same degrees of freedom as the robot arm of the real thing.

With its emphasis on appearance, I was disappointed at the representation of my favorite feature of NASA JPL’s Mars Rovers: their rocker-bogie suspension. The first complaint is cosmetic: this model placed all three pair of wheels with the same track (distance between left and right wheels.) Curiosity’s front and rear wheel pairs actually have a narrower track than the middle pair, which I speculated was done that way so the suspension can fold up for flight. While a static model does not need to fold up for flight, it should at least accurately represent the layout.

The next complaint is a combination of cosmetic and functional: the suspension rockers do not articulate. Their angle is fixed relative to the body. On Curiosity, the left and right rockers are connected via the differential bar which keeps the two rockers in sync with complementary movement: if one moves up, the other moves down the same amount. But on this model, the differential is a surface feature and not a functional one, without connection to the suspension rocker.

On the upside, at least this model has articulation for suspension bogies. This was also missing from its smaller sibling. With articulating bogies, this rover model can at least pretend to handle rough terrain capability even if it lacks full rocker-bogie capability. In this picture, the middle wheel is raised by a piece of 3D-printed plastic I had on hand.

And finally, the suspension arms leading up to corner steering wheels have right-angle bends that are not an accurate representation of Curiosity’s suspension. I suspect this was done as a compromise to make these parts 3D-printable without supports, but it further reduces fidelity of this model.

There are several additional print problems with this first draft. If I were excited about this model I would reprint in PLA to see if it improves as expected. But given my lack of enthusiasm about representation of rocker-bogie suspension, I am content to stop here and look around for the next project.

NASA’s Curiosity Rover Model Print Cleanup and Assembly

NASA published a 3D printable static display model for Curiosity rover, and one of the things they offered to make printing easier are STL files that have already laid out many parts so they can be printed all at once. The upside is a lot less work on setup and less time tending to the printer. The downside is that if one part fails, it dooms the entire print.

The rover suspension parts are all in a single large multipart print. The real Curiosity rover suspension structure is cylindrical, and this model tries to maintain that shape, meaning there’s very little surface area contacting the print bed at the bottom of the cylinder. In the first few failed attempts, one of the suspension parts (and never the same one twice) would pop free from the print bed and wreck havoc.

To work around this, I told MatterControl to add a brim on all parts to increase surface contact area. It allowed the print to complete, but now I have to cut all those brims off before I could proceed to assembly.

I started by cleaning up the wheel hubs and pressing them into wheels.

Following my tradition of rover building, I proceeded to build a rover wheel on a stick.

Which quickly led to a rocker-bogie assembly for one side of the rover.

Unfortunately, the rocker does not articulate on this model. Its angle relative to the body is fixed. So this particular portion of the model is no more functional than the smaller version. However, the bogie does articulate, and all four of the corner wheels can steer.

Having built one side, it was easy to build the mirror side and put everything together. I noticed I had two extra steering brackets left over. Reviewing the large multipart print, I now notice there are six steering brackets even though only four rover wheels could steer. I shrug and move on.

Assembly of the robot arm was straightforward following the directions, leaving rover head installation as the final step. The static model is complete and I can admire it in its entirety.

3D Printing NASA’s Curiosity Rover Model

I decided to build the 3D printed Curiosity rover model released by NASA, and ran into some problems with print bed adhesion. Whoever designed this model had a 3D printer with better print bed adhesion than mine. My first few printed parts would lift from my print bed.

Some of this is unavoidable, the natural orientation of some parts dictate minimal surface area. The wheels, for example, have to sit with their narrow side edges on the bed because that is the only flat side. Fortunately wheels are round and produced minimal stress.

In contrast, the body of the rover is a large rectangular solid with sharp corners. This is a recipe for lifts and they released the STL files with some pre-generated brims to help the corners stick. Unfortunately that was not enough for me, because some of the corners still lifted off the print surface. Fortunately this was only a minor cosmetic issue, since the bottom does not need to be absolutely flat to mesh with any other part.

Another cosmetic issue is the radiothermal generator at the back, which ramped up more aggressively than my Pulse XE revision D printer could handle with PETG. Fortunately this is a bottom-facing surface and shouldn’t be too much of a detraction.

The wheel spokes were the most problematic with their fine detail requiring a lot of filament retraction as the print head moves from one tiny feature to another. In my experience, retraction-heavy prints work much better in PLA than PETG, in hindsight that’s what I should have used.

An interesting nod to convenience is that, in addition to publishing STL for individual parts, the creator of this project also included STL files with many parts laid out to be printed all at once. The upside is that there’s a lot less overhead. The downside is that failures can be troublesome.

NASA’s 3D Printable Curiosity Rover

When I take Sawppy out for some publicity, people frequently ask about the 3D printable Curiosity rover static model released by NASA. Some mistakenly thought Sawppy was the NASA-released design, others wanted to know how the rovers compared. I couldn’t answer the latter because I never printed the NASA rover, to the surprise of some, so I thought I should do it at least once.

NASA’s 3D printing resources page for a printable Curiosity points to a GitHub directory that actually has two printable models. I’ve seen the smaller one at a MatterHackers event, printed by another attendee who left her little rover on Sawppy’s table to keep my rover company.

The small model has limited articulation. All six wheels can roll, but cannot steer and it could only sit on a flat surface because its rocker-bogie suspension joints are fixed. I also noticed the robot arm joint articulation doesn’t match that of the real rover’s. Still, it is undeniably a representation of Curiosity and a cute little model.

Since I’ve seen the little one, I decided to skip it and try building the larger one. “Large” is relative, of course, it would still be much smaller than Sawppy. Another important difference is that it is an unmotorized static display model, which is actually the main reason I had not tried to build it. I wanted a rover that moved!

But I’m glad I’ve built it, because it was a good study into the different compromises this model made for the sake of being 3D printing friendly.