Testing the Honeycomb for the Deck

Hello all, today was a highly productive day for Olin Sailbot. The mechanical, coding and electrical teams were in full swing. On the mechanical side, a great deal of progress was made with the deck, including tests on the honeycomb material that the deck will be fabricated from.

Mast test

Test : A piece of the mast material was inserted into a sample piece of the honeycomb material as we anticipate attaching it when the sails arrive. The piece of the mast was pulled with 10-12 lbs at 9 inches, with a resultant force of 90-108 in. lbs.

Observations : At approximately 7 lbs at 9 inches ( 63 in* lbs.), the epoxy around the mast began to fail. However, the honeycomb was flexible enough to withstand approximately 30 degrees of bending without failing. As we have shrouds which prevent this excessive motion, our mast should never be moving this much and therefore the honeycomb will withstand the forces on it.

result of

This is the carbon fiber after we applied 12 lbs to it.

Eyebolt tests

Test : We anticipate that the shrouds will be attached to the outer edges of the deck using eyebolts. In place of eyebolts, we used normal bolts and pulled on them at an angle of approximately 10 degrees with as much force as possible. There were three different situations tested: a bolt with no nut, a bolt with a nut, and a bolt with a washer and a nut.

Observations : Even when pulled at the greatest possible force (to the point at which I could not measure the force and I could not physically pull any harder), the honeycomb showed no signs of failure. In fact, the bolt itself began to bend before the honeycomb did. This was on the bolt with no nut or washer, and so was the ‘weakest’ of all the situations. This result reinforces what we had hoped before, namely that our shrouds will be strong enough to hold the mast stable and therefore we do not have to worry about the honeycomb around the mast step failing.

pulling on nut

Testing the bolts. From left to right: no nut, nut, nut and washer.

Carbon Fiber test

Test: To test the strength of the epoxy bond between the carbon fiber (which the hull is made from) and the honeycomb (which the deck is made from) we attached a clamp to the excess carbon fiber on a test piece and pulled from there. There were two different attachments of the carbon fiber to the honeycomb. One strip of carbon fiber was epoxied to honeycomb

set-up for pulling carbon away from honeycomb stright up

Pulling the carbon fiber parallel to the honeycomb.

Pulling the carbon fiber normal to the honeycomb.

which had been scratched, while the fiber strip was epoxied to the honeycomb with its surface left as it was. We then ran two tests, one pulling normal to the honeycomb and one parallel to the honeycomb.

Observations pulling normal to honeycomb: When the carbon strip without scratches was pulled, it immediately broke. While this did not show the strength of the connection, it did show us that the carbon fiber has a failure mode when bent at a sharp angle along the grain of the fibers. When we tested the fiber attached on scratches, I pulled until the scale maxed out at 40 lbs and I couldn’t pull anymore.

Observations pulling parallel to honeycomb : When the carbon strip connected with scratches was tested, I again went beyond the range of the scale and finally broke the string. I concluded from this that we do not have to worry about the carbon fiber detaching from the honeycomb as long as we scratch the honeycomb first.

The result of all this was that we know our material’s failure points, and we have decided we are working far within these. Good things were found out! We are well on our way to a finished boat, and with the end in sight the process is getting even more exciting.

Tagged , , , ,

Fabrication Speeds Up

Hello, all! The mechanical team for Sailbot is starting to speed up construction efforts as we move towards Spring Break testing. The plan for this year is to finish an initial boat by Spring Break and send students teams out onto the water with it to test and improve systems.

Current mechanical milestones

  1. We recently did a vacuum bag layup of the inner material of the hull. Our hull is an experimental (for us) composite-foam-composite sandwich, and the inner layer of composite material has given great rigidity to the hull.  The inside vacuum job had a problem with pressure leakage, we think, which left it strong but mottled.

    Empty Chaos Theory hull

    Empty Chaos Theory hull

  2. We’re cutting deck material to fit the hull, after measuring the actual boat. A jigsaw cuts the composite-honeycomb-composite sandwich smoothly.
    Measuring and capturing the lip of the hull

    Eric captures the lip of the hull before putting it into CAD

    IMG_0737

    Dante and Ben (l. to r.) cut a sheet of homemade deck material into the deck shape, after poster printing the shape we desired

  3. We’re in the process of fabricating the transmission for the rudder and the keel block. These will mostly be made out of wood that is epoxied into place in the boat.
  4. A sub-group is racing to finish the design of the sail transmission system, then they will fabricate it. Again, the goal is to get everything assembled by Spring Break.
  5. We’re working with several generous RC boat enthusiasts to create our sails, because our knowledge on sail design is rapidly increasing but our fabrication resources are still limited.

The hull floats! We packed weight into it to see how it settled, then did some brief testing to check that our chosen keel position was roughly above the center of buoyancy, and called it a day.

We put several bricks and a short keel into our new hull to see how it floats

We put several bricks and a short keel into our new hull to see how it floats

The loaded hull does list a bit to the side, which  reflects so inconsistency in the fabrication process. As this was our first time through this fab. process, it will be really important to reflect on how it went, test and retest the elements that worry us, and come back with a process we’re happy with.

Software Subteam Surfaces

Hello everyone! The Olin Robotic Sailing Software sub-team is off to a running start this semester. Last semester, a significant portion of our efforts were spent training a new crew of landlubber coders how to navigate through the fair seas of LabVIEW. This semester, we’re using our newly acquired LabVIEW legs to tackle an ambitious set of projects.

First, the old software architecture has been encrusted with the barnacles of pre-competition modifications/bug fixed and in-competition hacks. We’re givin’ ‘er a nice shine of clean architecture that should hopefully go more than skin deep, creating a scalable, modular platform for the addition of yet-to-be-imagined features on the horizon, such as boat scripting for testing!

Next, our user interfaces could stand to benefit from the receding tides of feature-creep. At present, our operator control unit is incredibly complex and no longer fits on our large, school-issued-laptop-screens. That’s a problem. The simulator interfaces is also not well integrated with the operator control unit, making simulation cumbersome. With a streamlined user interface, the software team will be better able to test and control the boat, both in simulation and in reality.

Lastly, the boat needs to be smarter. Our present sailing behavior is determined by an arbiter trying to navigate to a GPS coordinate while balancing other priorities… like not crashing. Unfortunately, the boat is not able to plan ahead or follow a defined trajectory. We are working on ways to implement true path-planning and following for the boat that is compatible with an arbiter… a difficult task for a sailboat which is not able to move in any direction it wishes!

There are certainly many more storms for the software team to weather. But with each passing day, our command of the seas becomes stronger.

Godspeed.

Tagged

Prepping for hull layup

Hi guys, the Sailbot team is gearing up to start the semester again. We’ve got an ambitious goal – try and finish our boat before Spring Break, so that we can test our systems and start improving the boat.

For the hull layup, we’ve been wanting to try composite sandwiches since the beginning of the semester. We made some very strong test pieces out of fiberglass-plastic honeycomb-fiberglass, curving the plastic honeycomb into a semi-circle before the layup, but the bend radius on the plastic honeycomb wasn’t sharp enough for the nose of the boat.

Instead, 3A Composites was generous enough to donate a sample of Airex Foam, which we will be using as a sandwich in our hull. The purpose behind making a sandwich of composites and foam is to make the hull much more rigid, and making internal ribs unnecessary. Ribs are a pain to make, by the way. Last year we laid strips of foam horizontally across the inside of our boat, then laid fiberglass over those strips, but it’s hard to get rid of air pockets and make the edges clean.

2011-2012 Boat, Blackbody, internal ribs

Here you can see the foam and fiberglass ribs on Blackbody Radiation

We’ve recently tested the thermoformability of the new foam, as you can see here. The foam fits well in the nose of the boat, and will form  to the hull shape well.

Foam core material in the boat nose

Foam core material in the boat nose

We’re excited to see our boat come together with these materials, which should be very strong for their weight. More later!

Fair winds.

Alex + foam core

Alex Crease, who’s on the hull construction team, excited about the new development

Electrical Update and Connectors

Finishing the PCB design is taking a lot longer than anticipated; we wanted to have it fully designed before Thanksgiving break, but we are still chugging along. Much of the delay was from honing down on the exact parts we were going to use to create the appropriate footprints on the PCB. Once we had our schematic created, we needed to associate the different components with patters that would define how the components would be attached to the board (e.g. holes of appropriate sizes and positions for through-hole soldering). The last step for finishing the PCB design is to lay out the traces for the board. The traces are lines of conductive material that connects the different components electrically. We should have the board done in the next week and ready to be ordered! The PCB will be smaller than the sbRIO, so our new waterproof housing can be much smaller than the old one.

Wanting to improve from last year, we’ve been researching other wire connection methods. Between asking around and searching the internet, we’ve found lots of new connectors we want to try. We ordered a sample of a wide variety of different wire-to-wire and wire-to-board connectors. We found some very good wire-to-board connectors from a company called Wago. They are great because they have a strong mechanical connection and require no soldering: switching out wires will be virtually effortless. We also found some tiny fuse holders from Littelfuse that connect directly to the PCB and perfectly match our needed range of values. We looked into some higher-grade through-housing connectors, but we didn’t find anything good enough within our price range, and ultimately decided that we would pass the wires straight through the walls and glue them in instead. This will make one less place for failure. However, if for any reason we need to redo the wire connections, it would be a very difficult and extensive task. For connectors outside the housing, we’re are pretty big fans of the MX150L series of waterproof connectors. We’ve found them to be much easier to use than other waterproof connector without a proper (and expensive) crimping tool.

Our test connector of the MX150L series.

Making a PCB: Diptrace Review

One of the electrical teams goals for the semester is to turn our wiring inside the electronics housing into a printed circuit board (PCB). To do this, you need* EDA (electronics design automation) software. There are a number of free options out there, but there is a big trade off between easy to use and full featured. ExpressPCB, for example, is easy to learn, but you can only use it to order PCB’s from their company, and it lacks some really useful features. EAGLE, which has a free version for non-commercial projects, falls at the other end of the spectrum: it can do pretty much anything you could want, but from previous experience, we’re convinced it was invented before the idea of usability. Because we wanted to get this PCB done in under a month, it was important that the EDA software have a forgiving difficulty curve.

DipTrace is the Goldilocks of EDA software in the hobbyist/low volume realm. It can do pretty much anything EAGLE can, but it also has nice tutorials, and an intuitive user interface. The free version is limited to two layer boards, but DipTrace was kind enough to sponsor our efforts by giving us a full version of the program.

They have quite the slick logo too.

We’ve finished the schematic (in DipTrace) and are just starting to design the board itself, but so far we’ve been extremely happy with DiptTrace. We’ve been able to teach a group of new members how to use it in under 3 weeks, and it’s done everything we need it to so far. We’ve just finalized the schematic, so hopefully within the next two weeks or so, we should be able to show you some awesome PCB designs.

 

*Ok, you don’t need it, but it will save you a lot of time.

Electrical Goals: Fall 2012

It’s been a little quiet on the blog for the past couple months. We’ve been working hard to get our new members up to speed and our work flow for the year planned (See figure 1). We’ve hit steady state, as much as that’s possible, so hopefully you’ll see more regular updates.

Well, ok, some of them are all on top of each other.

Figure 1: All of our ducks in a row. Mostly.

The Electrical subteam was in an odd position at the beginning of this year. We had an electrical package. It worked. It was wired. We almost didn’t know what to do with ourselves. Luckily for us, it also looked like this:

It's messy but we love it.

The electronics rat’s nest.

 

It works, or at least it did over the ~14 hours of running in Vancouver. But it’s disorganized, and that has a number of dangers: it’s more likely something could short, and if something does go wrong, it makes it much harder to debug. The electrical team’s goals for this semester are thus primarily involved with making the electronics package, sleek, clean, and beautiful. We’re doing this in a lot of different ways

  1. We’re turning the rat’s nest of wires that connect everything into a nice, custom-made printed circuit board. Once the PCB is fabricated, we can plug out components into it, and having everything just work. No more poking at connectors or digging through wires to see which one feel out. This will also shave valuable weight off the package as a whole: wire mass is not insignificant. Of course, the PCB does have some draw backs. Most significantly, it makes the package much less flexible. While we are working hard to include extra power ports and data lines on the pcb, adding a new component still won’t be as simple as throwing a few solder joints together.
  2. Connectors are the first year algebra of the robotics world: critically important, often overlooked, and liable to bite you. Last year we had problems with the ubiquitous hobby servo connector. Difficult to make properly and finicky even then, we’ve come to realize that they have no place on a robot of this scale. The electrical team has pored dozens of man hours into fixing bugs we ultimately traced back to a faulty connector. We’ve been playing with a selection of alternatives, and are quite happy with some. There will almost certainly be a post on this soon.

This is not to say that all we’re doing is house cleaning. We’re also working with the mechanical team to develop a novel mast actuation system, researching methods of power regeneration, improving our power draw models, learning to solder like NASA, and spec’ing numerous other small parts.

We’ve built an electrical system. Now we get to make it beautiful.

Making Chaos Theory – from tool to mold to boat

As you may or may not know, Olin Robotic Sailing’s plan for the first semester of this year is to produce a new boat for the Sailbot competition!

This is a series of ‘first time’ steps – this is our first time making sails, our first time taking a boat from CAD to reality, and our first time designing a mast. To produce a boat from the CAD, we first laser-cut a series of profiles 2 inches apart out of MDF (medium density fiberboard), seen in CAD here:

CAD profiles of the boat, which were cut out of MDF

Then we used the profiles to sandwich sheets of pink foam, before sanding down the pink foam to the wooden profiles. This was done by hand. In the nomenclature of composites (as far as we know it) the first plug is the ‘tool’, the mold for the final part is the ‘mold’ and the final part is, of course, the ‘part’. You can see the sanded tool here:

Our sanded tool inside of the Large Project Building

This was a bit of a difficult process, and we’re not sure whether it is worth replicating. Other ideas for taking a model from CAD to reality would be helpful!

Sanding the tool down got the boat pretty close to the shape we need, but the softness of the foam made errors and gouges in the tool inevitable. The next step involves spreading spackle on the tool, letting it dry, and sanding it to the correct shape.

Students working to spread spackle on the tool

After that we will create a mold from the tool, and a boat from the mold. More details on this as it happens!

Testing Blackbody Once Again

Last weekend a small group of students assembled to take our competition boat, Blackbody Radiation, out to Lake Waban for her first on-the-water test this semester.

The testing crew loading chase boats onto the boat trailer.

We wanted to ensure that Blackbody was still functional after traveling to across the country (and back) this summer, and we wanted to get some of our new team members out on the water with her. When we got to the lake, the returning members got to teach the new members how to assemble the boat. The only problem we encountered was a mislabeled electrical connection that required some finagling before it allowed the code to have rudder control, but fortunately our electrical subteam lead was there to figure out what was wrong and fix it.

Two new team members put up Blackbody’s sails.

Towing Blackbody to the center of the lake, because the shore is in a wind shadow.

Once enough wind came up that Blackbody could move, we confirmed that she was still able to perform all of her missions well, and we taught some of the the new team members how to input missions on the OCU.

Blackbody Radiation back in her element after spending the summer in a shipping crate.

Overall, testing was very successful, and we cannot wait to do it again with this year’s competition boat after she’s built.

Kicking Off Olin Robotic Sailing 2012-13

We’ve been back at Olin for about three weeks now, and the Olin Robotic Sailing Team is really hitting the ground running this year.

Olin club fair happened about a week after we arrived back on campus, so that all of the students would have time to get settled before being barraged by information about all of the clubs and student organizations on campus. We of course brought all of our boats from last year outside to show off, and there was a lot of interest in the team by freshmen and upperclassmen alike.

Here Jared the Sailbot coding subteam lead poses in front of last year’s competition boat Blackbody Radiation at club fair.

During the week following club fair, we held a general meeting for all of the new members, so that we could tell them more about the team’s goals for the year and how the team is organized.

The team’s founder James spoke about all of the opportunities we have this year and about the team’s goals for the future.

Following this, new members got to talk to the subteam leads individually and ask subteam-specific questions.

The next step in getting the team up to speed was to host this year’s “Sailbot Bootcamp.” In bootcamp we taught all of our new members how to solder properly, helped install LabVIEW on their computers, and taught them some common sailing terminology and theory that they’re likely to hear and use frequently while working on this project.

Electrical team lead Jeff teaches our new members how to solder properly.

James teaches our new members some basic sailing terminology and rules.

Now we’re on to regularly scheduled subteam meetings. Check back for more updates about our progress later this week!

%d bloggers like this: