May 15-17, 2015

The project was received really well at Maker Faire. Here’s a post Faire team picture. Note the two Editor’s Choice ribbons attached to the frame:

We had a large space just to the West of Expo Hall:

The team did demonstrations every half hour, most of which drew hundreds of people at a time:

We invited people to come around to the back of the suit to see “where the magic happens”:

The control system consisting of 4 custom Arduino boards, one Arduino Uno, and Enfield proportional valves and controllers.

We were able to do lifting tests of up to 100lbs per arm, and we demonstrated how the force sensors control motion of the legs. We are still working on the software for walking, so we weren’t able to get the suit to leave the support frame (the yellow structure at the center of the image above).

There were almost continuous press interviews during the weekend.

Gabriel, Thelonious, and Cole being interviewed for the Popular Science video.

A few highlights: The Exosuit was the lead story in Re/Code’s article on Maker Faire. Business Insider also did a nice piece on the project, and Popular Science in collaboration with portico.tv featured the Exosuit in their Maker Faire report (click on the Popular Science logo to view the video; the Exosuit is introduced around 7:40).

Here’s one of our favorite pictures from the weekend. The guy in the baseball cap is Carl Bass, CEO of Autodesk, and the fellow to his left is Larry Page, CEO of Google. To have them both at the booth at the same time, with big smiles on their faces was just amazing.

A couple of other memorable stories: An MIT recruiter said she was blown away by the project. It’s the sort of thing they expect from their graduate students, not high schoolers. A quite a few people said they thought the Exosuit was the best thing they saw at the Faire, including a planetary scientist from NASA. We found out from her that exoskeletons will play a big role in planetary exploration because the space suits are so heavy that even on low gravity planets like Mars that the astronauts will need strength assist.

All in all an unforgettable experience for the team. We couldn’t have done it without you and our other sponsors (Autodesk, Enfield Technologies, Clippard, Perforce, Not Impossible)!

April 26, 2015

Our main goal today was to test walking. It was semi-successful. The first few attempts didn’t work very well, and it was difficult to determine exactly what the problem was, so we started paring down what we were testing. A number of things were identified as contributing factors: the back pack straps were mounted too high on the frame, the sensor mounts for the legs weren’t positioned properly, and the feet tended to tilt forward and get caught on the driveway. We decided to first focus on the sensor mounts, so we diconnected the feet and started to test each leg separately. We eventually got the right leg working pretty well, so that gave us some hope that the whole design is workable.

At the end of the day the left leg wasn’t working well, so we suspected it was because the sensors needed to be repositioned and the back pack straps were too high. During the week Joseph and Tony tweaked the sensors mounts and repositioned the back pack. Things are working much, much better now. We’ll try a full system walking test this coming weekend.

While the walking tests were going, Ed and Connor reinforced the 3D printed joystick mounts

and Sammy and Thelonious worked on the roll cage that is designed to project the head and shoulders of the wearer should a fall occur.

Cole and Thelonious also mounted the laptop stand on the frame, shown here on the right:

April 18, 2015

We have building up to a visit from Discovery Channel Canada today, but they got in touch with us earlier in the week saying that they’d like to postpone their visit. The plan was to do a first complete system test, including lifting and walking. We almost got to that point: we lifted but didn’t walk today.

Because the Discovery Channel wasn’t here today we worked to record the video we thought they would want. The best video camera we have is also the camera we use to shoot stills of the build sessions, so this update doesn’t have many pictures. Stay tuned for video from today.

Once everything was wired up and ready to test, we were all holding our breath, but the suit didn’t move. We’ve had this experience many, many times in the past, so we expect it. The diagnostic console, shown below, indicated that the sensors were working correctly, but for some reason the actuators weren’t responding.

After some debugging we realized that the control boards weren’t receiving the command to “Engage” the actuators. For safety reasons we designed the system so that when the suit powers up, the control software doesn’t try to control the suit, that is, it starts up in a disengaged state. After some further debugging we determined that it was a timing problem: the egage command was being interleaved with commands to update the user interface. Once we understood the problem, it was an easy software change to fix it.

The next problem was more mysterious. Two of the control boards were not able to drive one of their outputs. The control boards that are responsible for the right arm and right leg were unable to actuate the cylinders assigned to Arduino pin 6. For the left leg this meant the left hip, and for the arm, this meant the right shoulder. The software looked ok, so we focused on the hardware, the board layout in particular. The board layout driving pin 6 was very simple, so that led us to believe the problem was with the Arduino chip. Swapping out the Arduino with new chips fixed the problem, which is good. The remaining problem is that we don’t know why 2 Arduinos would have failed in the same way. It is likely that there is a lurking problem we haven’t yet identified. Scary.

The third problem is that one of the control boards seems to have failed. Cole, Thelonious, Connor, and Sammy worked to etch, drill, build, and test a new control board.

As of Monday night, April 20, we have 4 working limbs! This is a huge milestone, so it looks like we’re on track to walk and lift on Sunday April 26.

Wish us luck!

Saturday April 11 and Sunday April 12, 2015

The big goals this weekend were getting the valves and electronics mounted. We mounted the valves in the same position that we had temporarily clamped them last weekend. The electronics were mounted on a separate shelf, with the valve controllers (Enfield D1s) mounted below the shelf, and our 4 custom boards plus the Arduino for communication mounted above. Here’s a series of pictures showing the process:

Here’s the beast so far. Note the mounts at the end of the forearms where different hand attachments can be secured:

Ed and Connor also worked on the design and fabrication of a mount for the battery:

Connor and Thelonious discussing early designs for a closable hand:

We think we’re in good shape for a first walking and lifting test next weekend. Wish us luck!

April 4 and 5, 2015

We had a VERY productive weekend, meeting both Saturday and Sunday. It looks like we’re in good shape for being ready for the Discovery Channel visit on April 18.

Sammy and Thelonious mounted the leg sensors:

It was “build cables day” for Gabriel. Cable after cable after cable…

The valves (the dark boxes) and the drivers (the smaller, lighter colored boxes used to provide current to the valves) came in from Enfield:

Ed and Joseph attached the host fittings to the valves

Then they figured out how the valves would be mounted on the suit and connected the pneumatic hose that runs from the valves to the cylinders. The valves are just clamped in place until we get some threaded rod to mount them permanently:

Here’s the splitter used to divide the air from the compressor into the 8 streams needed for the cylinders:

Other views of the suit with the valves and hoses. It’s starting to look AWESOME!

The team T-shirt design was also finalized. We just have to order them now.

Sammy and Thelonious did a run to OSH to buy some bolts and and a new blade for the metal cross-cut saw. We’ve cut so much steel we completely wore down the original blade. Here’s a comparison between the new blade and the used one:

On Sunday Connor finished the joystick mounts and cables. They’re all ready to connect to the control boards:

Cole finished building and testing the four control boards:

Thelonious and Tony attached the drivers to the bottom of a shelf that will also hold the control boards. Taping the drivers in place on a partially transparent sheet of acrylic so that the exact position of the mounting holes can be determined:

Screwing the drivers to the acrylic:

After attaching the drivers. The next step is to bolt the acrylic to the metal shelf and then mount the control boards to the top of the shelf:

Next week we’ll mount the valves, control boards, and drivers to the suit, and we’ll test various subsystems in preparation for the full system test on the 18th.

March 29, 2015 Today was the second of a two-build-session weekend. Thankfully the energy and productivity of the team was much higher than the last time we had a double build. After today we’ll have four more build sessions before Discovery Channel Canada visits on April 18, and we’ll need them all! There was a marathon that closed the bridge this morning, so instead of our usual meeting time of 10-3 we got together from 12-5. So far the suit weighs about 50 pounds, so it takes a couple of people to man handle it: Joseph spent a good deal of time editing the video update from yesterday’s build session while Cole, Joesph and Ed worked with a guest to debug a problem PC: Connor finished 3D printing the first of two joystick mounts, and began building the cables needed to connect the joystick to the control boards: Cole debugged the second of four of the control boards (which were masterfully laid out by Gabriel) needed for the final suit: One nice aspect of this version of the control boards is that they have what are known as ISP (In System Programming) headers, meaning that we don’t have to remove the Arduino chips from the board to reprogram them. The problem with removing the chips is that each time you do there is a chance that you will bend the pins to the point that they break. The ISP headers will therefore make it much faster and easier to continually refine the software to get better and better responsiveness out of the suit. Next weekend we’ll work on mounting and connecting the Enfield proportional valves, the pneumatic hose, and the valve drivers.

March 28, 2015

Our goal for today (which we reached) was to mount and assemble the 8 pneumatic cylinders on the final armature!

Here’s the assembly:

Leading up to that, here are 5 of the 8 3 inch diameter cylinders:

Sammy cut foot plates and drilled the mounting holes:

Joseph and Thelonious used clamps to determine exact mounting points, and then welded the foot plates into position:

Ian and Sammy worked on the black detail on the orange stand:

March 15, 2015

Yesterday was a special “pi” day since the date was 3/14/15, a good approximation to the digits of pi (3.1415926…).

We did a test last week with the Enfield valves and controllers and got good speed and control on the smaller cylinders, but speed was about of what we need when using the big 3″ cylinders. Our belief was that we were flow limited because we were using narrow (1/4″ pneumatic hose). We ordered some 3/8″ hose and fittings, and Ed did a test that verified that the speed is now plenty fast. That’s a big deal. We really didn’t have a plan B.

Meanwhile, Thelonious, Sam, and Sammy continued work on the torso:

including testing out the mounting of the shoulder and elbow:

Gabriel and Cole designed and laid out the PC board for the final suit. This is likely the last custom board we’ll need for the project.

Ed and Connor continued their work on the joystick mount to control the arms:

And what session would be complete without a Cyber Sammy?!

March 8, 2015

Today was the second build session this weekend. Whereas LOTS got done yesterday, the team was pretty tuckered out today. It was also the start of daylight savings time, so lack of sleep played a factor too. Frustrations were running pretty high, and a number of mistakes were made, probably because concentration wasn’t as sharp as usual.

Connor and Ed worked on a 3D printed mount for the joystick that will be used to control the arms:

Sammy, Ian, and Thelonious continued fabricating parts for the final suit:

Cole got the boards working to control the prototype legs:

We had planned to have a third build session of the week (since it is break week), but given the low energy of the team we cancelled it. We’ll back back rarin’ to go next Sunday though.

March 7, 2015

Yikes! It’s March already! We’ve been working on the project for 9 months. There are some periods where progress seems to either proceed slowly or in reverse, and other periods where things seem to happen quickly. Today was, thankfully, one of those quick periods.

After several weeks of work, with Gabriel doing the layout, Cole got the control board for the prototype legs working:

This is an important milestone because the board for the final suit will be very similar to this one. This board contains an ATMega328 that controls two cylinders and is capable of communicating with a “master controller” (another Arduino) that gathers data necessary for display on the diagnostic console.

Joseph completed the CAD design of the final legs, hips, and spine. Here is consults with Thelonious on one of the parts:

Thelonious, Ian, and Sammy rapidly measured, cut, welded, and drilled the various parts that will soon be assembled into a second suit:

Each piece sits atop a printout of its specification, which is good way to keep track of which parts do what.

Tony contacted the nice folks at Enfield Technologies this week, the makers of the proportional valves we plan to use to get smooth speed control. The president, Ed Howe, has been very supportive, and Roy Niemann, one of their engineers was super helpful in figuring out exactly what we need and how to use it. (You would think that speed control of pneumatic cylinders would be a standard thing, but for some reason it’s not. Enfield seems to be one of the only ones doing this kind of control, and their products seem really solid and easy to use.)

Connor and Ed took that information and wired up several tests, first on a table top, and then on the prototype elbow:

Another highlight of the day was a delicious lunch from Sol Food, a very popular Puerto Rican restaurant in San Rafael, courtesy of Scott (Cole’s dad):

Finally, the Kickstarter campaign crossed the $5,000 mark today, raised from over 100 backers. Way above our goal, so we should have plenty of funding to complete the project. The promo video has been viewed more than 7,500 times, so word is starting to get out.

Progress continues tomorrow when we meet for another build session. Let’s hope it’s as productive as today!