Sculptor Jonathan Schipper uses mechanical and robotic means to slow down and show the inherent beauty in various destructive events. This scaled version of The Slow, Inevitable Death of American Muscle uses mechanical reduction to simulate a super-slow motion front end impact between two die cast cars.

The entire system was first modeled in SolidWorks. The frame is based off of known dimensions such as the hydraulic cylinders, standard steel I-beam sizes, and typical automobile wheelbases and lengths.

Solidworks Finite Element Analysis was used to determine if the steel carts could handle the 40 tons of force supplied by the hydraulic system. Material selection, gusset placement, and thickening high-stress areas brought the hotspots under control. This particular plot shows the stress in the moving steel carts that are fastened to each automobile.

An elecro-hydraulic reducer was designed to power the sculpture. An electric motor spins at 1000 rpm at the beginning of the gear chain reduction. After eleven sets of reductions though the Ford V8 timing chains, the output spins at one-half of a revolution per minute. The eccentric shaft output works a hydraulic pump back-and-forth to suck oil out of the reservoir and supply hydraulic pressure to the ram cylinders. The combination of mechanical and hydraulic reduction allows each car to move at a super-slow 0.005 inches per minute.

A keyway broach is used on the custom timing gear hubs in order to lock them onto the shaft. Here, the lathe is used as a press to drive the broach through the hub.

All the timing gear assemblies are put together and ready to be assembled onto the shafts and into the hydraulic pump unit. For V8 timing gears give a 2:1 ratio and use really nice looking timing chains.

The timing gears and chains are assembled into the hydraulic unit. Each chain is connected to the next assembly in line. With this reduction a 1000 RPM input speed turns into a 0.5 RPM output.

One half-length of the crash system with the hydraulic cylinder and cart installed. We used steel wheels to roll along the I-beams like a roller coaster. In later designs these wheels were switched out for plastic slides. When installing in outdoor areas and gravel, it's tough to keep stones out from the path of the rollers. The hydraulics are strong enough to crush little stones, but this could jam up the system. Sliders just push the stones out of the way.
 

The installation area was in an interior courtyard at Stuk Art Center in Leuven, Belgium. It was also one level down from the street. The cars and system were lifted into a patio area by crane, then wheeled into the courtyard through the Art Center.

A time lapse taken over a couple days of the first running of The Slow Inevitable Death of American Muscle

The results after a four-day installation at the Artopolis Chicago Art Exhibition. An unexpected "rearing up" of both cars gives quite a dramatic presentation at the end of the show.

A time-lapse video of the Chicago Car Crash.
 

BMW vs Mercedes at an indoor courtyard in The Hague.

Firebird vs Camaro in Marfa, TX

A Duster vs Camaro outside the museum in Wolfsburg, Germany.

The chairs cycle via powered gearmotors and electronic clutches....

The key to getting the desired movement was a combination of a strong and compact gearmotor with an electronic clutch release. AXN|RXN designed an electromechanical assembly that could slowly winch the chairs together, but then also release them when needed.

Footage of the Gravity Harps playing Solstice in the Biophilia tour.

The gravity harps were designed in SolidWorks. This rendering showed how four harp assemblies might look clamped to vertical beams.

One interesting challenge was to sonically isolate the harp from all the other vibrations in the system. In order to only hear the harp's strings and sound board resonate, all the other vibrations from bearings, bushings, motors and support structures had to be removed in order for them not be picked up by the acoustic microphones. Custom mounting hubs were designed to house layers of specific sound-deadening materials so that there was not a direct connection between the harp and the surrounding mechanical system.

There are a number of barbell-style weights that act as a counter weight to the pendulum. By adding more weight or placing that weight farther from the axis, you can slow down the frequency of the pendulum. We adjust that weight by having adjustment nuts on a threaded rod. This threaded rod gets assembled to the pendulum via the threaded hub shown.

Here are the finished machined aluminum parts that were made for the three-pendulumn version of the Gravity Harps. All material is 6061 aluminum CNC machined or turned on the lathe.

As the pendulum swings through it's arc, the plucker arm stays vertical via a mechanical four-bar linkage. These three assemblies are actually part of that pivoting linkage that keeps the plucking arm in line.

Testing of the four-bar plucker linkage.

Initially, the gravity harps were envisioned as a 20-ft tall ring 40-feet in diameter. Each harp only played one note. Bjork would be on a pedestal in the center as harps swung in a pattern all around her. Super Cool, but also Super Expensive and Super Complicated!

The finished Hive on photo day. Photo courtesy of Green Card Pictures.

Here is a SolidWorks rendering using PhotoView 360. You can compare this to the physical version about. While it's not exactly the same, you can get a good idea of what the final version will look like even before you order any materials.
 

Here are three tuned Stella Chalices with solenoid-based strikers. The solenoid-based striker had a bunch of advantages over other designs. A solienoid is mechanically simple and only involved a magnetic housing and a steel bolt. By throwing a current through the solenoid, the bolt shoots in a direction like a bullet out of a gun. By varying the intensity and duration of the current pulse to the solenoid, you can adjust the musical volume of the strike. An Arduino microcontroller makes this a snap, because it can all be done by just adjusting lines of code.

A nice detailed pic of a Stella Chalice / Striker assembly. Photo courtesy of Green Card Pictures.

This is an assembly of the chalice and solenoid in SolidWorks as compared to the picture of the final physical version above.

Here is a view of the internals of the Hive. Two layers of plywood were machined to create a clamping "sandwich" to hold all the chalice arms. With this construction all arms can be adjusted in and out as well as rotated in order to fit. A length of polished stainless sheet was wrapped and adhered to the plywood in order to make a "bucket" for all the electronics.
 

This is one mechanical prototype for a Chalice striker as used in the Stella Hive. The parts were designed in Solidworks and printed via AXN|RXN's Dimension 3D printer. The cam is driven via a stepper motor and driver. The Arduino code is on the monitor in the background. This version was found to be too slow in operation and had too much mechanical noise to use in the final version.

A photo of the completed Pyrophone on the day of unveiling to Stella Artois. All the robotics, electronics, and gas canisters are hidden in the maple star-shaped base.

Photo by Jenny Long

This video gives an idea of what the pyrophone sounds like. These are LOUD! Imagine twenty of these all at once. It's quite impressive.

Packing 20 burner bots into the base was a bit of a challenge. Assembling everything in Solidworks before the build makes sure that there is enough room so that the moving torches don't collide with each other.
 

The little bots that lift the torch tips up and down are made up of a small stepper motor and a 5-start Acme screw. The 5-start screw just means that it has five independent threads going along the screw. This increases speed that the lifter can go up and down versus a standard screw. An Arduino using the AccelStepper library controls the position and ramping of the screw motion.

Here is some roof top testing early on in the Pyrophone design process. As seen here, we could make quite an impressive flame using just propane and a glass tube, but no sound to save our lives.

This is a test of the initial valve-based design. We were trying to get the tube to resonate by turning a supply of gas on & off via an electrically-actuated valve. We could get a beautifully full flame with this method, but we couldn't get it to make any sound. It turned out that a simple Bernz-O-Matic torch flame worked the best.

The official Stella Artois description of the Violina.

The final version of the Violina stacks three wheel assemblies onto a vertical pole. Not shown, and one of the trickiest parts of the design, was running the wiring for all the wiring and communication through a 1.5" tube.

The critical parts of the Violina's rotating wheels are machined from solid maple. Machining wood is more of an art form than metals due its inherent grain direction. While a piece of aluminum can be set up in any orientation and produce the same results, wood may chip out or burn in one orientation, and be too weak to withstand a load in the other.

Using CNC milling equipment, six arcs of the Violina wheel could be machined from one slab of maple. This helps save time a material over cutting each independently on a band saw.
 

The "Horse Wheel" is the outermost spinning body in the Violina assembly. It is responsible for supporting the tensioned bowstrings that cause the Chalices to resonate. Made of solid maple and aluminum, it has to be accurately machined and balanced in order to produce a consistent note. Do to it's large diameter, each ring layer is actually made up from six separately machined arcs. With three layers in each hoop, twelve maple arcs plus six aluminum arcs have to be assembled together without compromising strength nor balance. By overlapping the seams of the arcs and connecting with steel dowels, the entire assembly becomes accurately round and strong enough to take the tension form the bow string.

Each playable level of the Violina is comprised of an inner and outer ring. The inner ring spins one direction and contains six tunes Stella Chalices mounted onto electronically controlled actuating arms. When the motor in the actuating arm is provided current, the motor turns a gear set which in turn pushed the chalice into the bow string. The chalice can be adjusted in any direction in order to get contact with the bow strings just right.

Photo Courtesy of Jenny Long

Inspiration came from turn of the century motorcycle designs. These designs were very lightweight and mechanically simple, borrowing much of their components from common bicycles. The Indian motorcycle shown above shared many elements with what were intending on building. It had a totally rigid suspension, a simple saddle-type frame, and still used standard bicycle components.

Oddly, AXN|RXN was also inspired by the ancient firearms seen on display at the Metropolitan Museum of Art in NYC. These rifles had long octagonal barrels and delicate scrolling designs on their fastening plates and brackets. Certain points of this old aesthetic could work well with the retro ideas in motorcycle design. One style of this type of rifle is called a Blunderbuss, of which we borrowed the name for our project.

The Blunderbuss started with a SolidWorks assembly model. We based most ergonomics off of common mountain bike dimensions. Wheels and tires were borrowed from a downhill racing bike as well as the brakes, hubs, and headset bearings. The moped was designed to be bolted together from three main components: the body, the front forks, and the frame. The body is fabricated from two formed lengths of aluminum sheet which are fastened together by rivet strips to form a tubular structure. The front forks are made from simple steel tubes with CNC machined cross braces and dropouts. The main U-shaped frame resembles the Indian frame on the previous page and is designed to be CNC bent from a single length of tubing.

The Blunderbuss Alpha Prototype was fabricated in Brooklyn. This was a quick and dirty version that would help validate our design and check out the ergonomics. Certain short cuts were made for fabrication purposes. We welded the two body halves together rather than using rivets. The U-shaped frame was welded from bent sections rather that using a single length of tube. The headlight and taillight caps were hand fabricated from aluminum rather than the vacuum formed plastic pieces in our model.

This is the maiden voyage of the Blundebuss Motorcycle concept outside the Kingsland D&E shop.
 

Another view of the Blunderbuss prototype outside the Brooklyn shop.

We've learned a number of things from the prototyping effort. The overall weight of the bike is around 120lbs, which is light for a moped, but too heavy to use bicycle components. The tires and brakes are not suitable for that much weight, so we've upgraded to minibike brakes and are sourcing off-the-shelf moped wheels and tires. The next step is to bring the moped in to a licensing station and find out what else we have to accomplish before we can legally ride it on the road. Perhaps for the Beta prototype we will focus on weight savings or go directly to an electric power train.

This image shows a mono-shock pushrod suspension in the works.

The Coincidence Clock is was designed in SolidWork using sketches provided by the artist. The square base plate is machined from half-inch thick 6061 aluminum. All dials, face plates and rotating signs are machined from brass.

The Coincidence Clock keeps track of coincidence probability rather than the standard seconds, minutes, hours. The innermost circle counts half-days as it rotates through 180-degrees. The outermost ring indexes 1/12 of a rotation (one month) for ever 30 days of the center. In addition, the Coincidence Dial indexes once every two months. At the end of 26000 years, the rectangular Season Symbols rotate to flip the season direction.

Each season symbol is machined onto a pivoting brass block. The opposite season is machined onto the other side. At the end of the clock coincidence cycle, RC servo motors are used to flip the blocks 180-degrees to reveal another season.

Three separate NEMA 17 stepper motors were used with Automation Technologies digital stepper drivers to animate the Coincidence Clock. The center Day Dial is mounted directly onto a motor shaft, the other two motors power the Month Dial and Coincidence Dial via MXL timing belts and pulleys. All motion is controlled via an Aduino Uno and the AccelStepper library.

This is a quick SolidWorks animation showing the assembly of the Coincidence Clock.
 

Inhale featured on the invitation the Bruce Pearson's solo show at Ronald Feldman Fine Arts.
 

Not the best representation of the piece, but the one that I took while at the opening.

The electro-mechanical bed is made up of two stepper motors, linear bearings, bearing rails, and bearings. In order to keep sound to a minimum, all motors are mounted on vibration deadening bushings. The assembly shown here is them clad with wood and painted white to blend in to the ceiling.

The automation electronics for Inhale are based on two NEMA 17 stepper motors, digital stepper drivers, and Arduino Microcontroller and a Stepper Controller. The calculations for the patterns are programmed into the Arduino and sent to the stepper controller. The controller takes those patterns and turns them into the electrical pulses that directly control the stepper motors.

Here is a photograph from Hamilton's 1955 installation of Man, Machine and Motion.

Reproduction frame connecting hardware was CNC milled from mild steel in order to duplicate Hamilton's original design. All hardware was nickel plated and installed with stainless steel screws.

A machined steel sample for the New Museum to verify useability before the production run.

Even the clevis mounts used to hang the artwork in the frames was custom machined. Over four hundred mounts were needed for the installation. All machining and fixture design was done in-house at Kingsland Design & Engineering Lab.

The Whirling Dervish was featured at London's Royal Orchestra. It played the part of Satan's musical instrument, which is quite fitting.

The Whirling Dervish was designed and engineered using SolidWorks. Due to limited time, the layout for the Whirling Dervish was much more focused on ease of fabrication and packaging than pleasant aesthetics. The original Whirling Dervish piece was envisioned as a field of spinning wings mounted on tall tripods. We soon found out that we were given a 12' x 16' indoor space with 12-foot ceilings. In order to accomodate this space, AXN|RXN packaged the spinning arms vertically in two rows of shelves. This design could be easily fabricated, assembled and fit within the space we were given.

All parts for the Whirling Dervish were machined in-house by AXN|RXN at Kingsland Design & Engineering Lab. The above photo shows six CNC machined aluminum hubs that are used to mount the whirling wings to the motor shaft.
 

This is the first test of the whirley arm on a NEMA 23 DC motor. The whirley arms had to be able to accelerate quickly in order to get up to speed and produce a note. This setup fit the bill.

Steel tabs were welded onto the tubular frame uprights to help support the horizontal shelves. The shelves nest onto these tabs during assembly so that one person can assembly the entire unit. They also provide structural support so that the frame doesn't sway when a number of whirley arms are spinning about.

After the frame was welded and assembled, two whirley motor assemblies were fit up in order to check clearance. It was a concern that the wings might droop or move around when they were spinning. This could cause a collision which would have disastrous consequences.

Large diameter aluminum hoops surround each rotating whirley blade for safety. These hoops were rolled using a three-wheeled tube rolling machine. You run the tube back-and-forth through the wheels while gradually increasing the bending pressure. After a few passes, the tubes are the desired diameter.
 

The best way to see if the rolled aluminum tubing is the correct diameter is to match is up with a full-scale print out. A large format printer can print a full scale drawing from SolidWorks for comparison. A little adjustment here and there and its perfect!

This is the first time the entire system had been installed and played. Fourteen independent whirley wings are controlled via a pre-programmed score. It looks great even when it's just idling.

Each piece of the set is of the exact same outside dimensions. All combined, they form an elegant machined block on your desktop.
 

Each model has exactly the some outside dimension. All are CNC milled from 6061 aluminum, hand sanded to a straight line finish and anodized black.

The Brickhaus Tape Dispenser uses a standard tape roll with a magnetic cap to keep it secured. The serrated teeth are machined right into the aluminum housing.
 

A simple bin for paper clips, stamps, or whatever is on your desk. A magnetic lid covers it all. Also great for jewelry!

A machined aluminum tray for your business cards. The back wall is angled for the cards to lean on. The bottom surface is machined with ridges so that they don't slide and fall flat.

Hold your pens, pencils, spoons, X-Acto Knives, whatever!

Hidden inside this aluminum block is a standard X-Acto razor blade. Swipe your letter through the slot and the blade shaves off the top of the envelope.

The entire bed was designed in SolidWorks beforehand. Here is a perspective view of how the platform bed should look after it is finished.

The oak leg is glued up from multiple thicknesses of oak. It is designed to be assembled over the round cross tubes. A 3/6-16 threaded insert for wood is installed into the center hole. This way the leg can be fastened tight to the C-Channel with a button head screw.

The red oak slats were cut from 3" x 5" boards. Here they are waiting to be drilled for dowel holes and then sanding and finishing.

They oak legs were glued up, tapered on the table saw via a fixture, then CNC drilled and milled in order to fit the frame. The three holes are pre-drilled then milled out with a 1/2" end mill to create the larger holes that fit over the 1-1/2" steel tube and the bottoming hole for the threaded insert.

Structural steel C-Channel was cut to length, radiused, and had square holes CNC cut along it's length. The square holes accept 1.5" square red oak slats that run the width of the entire bed. Oak legs are designed to mount onto the header and footer tubes and fasted with button head screws.

The steel channel is blackened and sealed with rub-on polyurethane. The red oak components are stained and also finished in polyurethane.

An iPhone photo of the Brickhaus bed with a queen foam mattress installed. NYC apartment are small, so please excuse quality of the picture!

The pictures above show the initial setup of the client's original system as well as a typical exercise one would perform. Detailed notes and video were taken of the existing construction and assembly methods as well as all of the exercises that could be performed. An Industrial Design colleague and I brainstormed different product concepts that met the clients exercise requirements, simplified use, and met the production budgets. These various concepts were then pooled together into three different product concept boards and presented to the client.

The first of three product concept offered was a conservative design the bears much resemblance to the original system. This concept uses a bent wood laminate as the exercise surface and cast aluminum or molded plastic end caps rather than the original all-sheet metal design. We also included an adjustable wing idea that allows the user to vary the incline of the exercise surface. The handle bar apparatus in this concept was left exactly as it was the original unit except for a quick-release connection where it met the base frame. The square tubing of the base frame was rotated 45-degrees to easily mate with a quick-release attaching concept.

Concept 2 is a significant departure from the client's original design. From the research phase we learned that it was not necessary for the user to detach and reverse the large working surface in order to conduct all the desired exercises. If both the convex and concave shapes were put on one side of the form, the user would not have to pick up the awkward form and flip it around to change exercises. This method also allows adjustment notches to be molded into the form that would fit snugly around the square base tube of the Handle Bar Apparatus. So, to change exercise positions, the form need only be tipped up a couple inches, rotated to the desire orientation, then dropped back into position. Ease-of-use is dramatically improved. This form can be manufactured by relatively inexpensive blow or rotationally molded means and reduces cost even further by totally eliminating the need for a Base Frame.
A more fluid Handle Bar Apparatus design was put forth to help eliminate the hard mechanical looks of the original design. A center pad was designed into the bars as a chest pad in case a users arms became too tired to support themselves when exercising.

Concept 3 uses a pivoting Working Surface to provide the necessary contour for exercising. The pivoting nature allows the user to easily adjust the degree of difficulty by means of a gas spring or an adjustable rigid leg that snaps into positioning holes in the Base Frame. If the user wants to switch from the convex to the concave form, they simply rotate the form around the center pivot so the opposite side faces up. This concept suggests that the Working surface could be molded from one piece, but the stacked aesthetic also lends itself well to construction from flat stock, simplifying prototyping and reducing production costs. It also allows the hiding of the mounting points for the Base Frame and adjusting legs within the form.
The Handle Apparatus in this concept makes use off-the-shelf handle and Aero bars to give the user a more bicycle-like feeling. It also much less expensive to use off-the-shelf curved handles than custom bent tube when dealing with lower quantities. A round vertical handle tube is used to further get away from the original rigid mechanical look.
My designer colleague and I liked this concept the best out of the three because of its ease of use, smaller size, and the unique stacked look of the working surface form. Because the form could be created form flat stock it gave us a lot of material choices and aesthetic flexibility. The client agreed that this concept had the most promise and gave the go-ahead for the engineering phase.

The actual product represented in Concept 3 had to undergo a series of changes in order to meet design, manufacturing and budget constraints. Aesthetic items that were expensive to prototype and manufacture (i.e. corner bumpers, curved handle bars, extrusions) had to be eliminated. The client decided that a more simple, robust Handle Bar Apparatus was needed for the increased loads seen during exercise. So, standard heavy-wall square tubing was used to fabricate the telescoping Handle Bar Apparatus. Oval tubing and a bent acrylic Aero pad were used as inexpensive accents help to soften the mechanical looking structure.
Various methods of attaching/detaching the different components were proved out both physically and digitally. Through the use of a breadboard mockup, we found that we could get away with a base frame half the size as shown in the concept board. Round handles were added to end of the Working Surface so that it could be rotated to the vertical position, detached from the Base Frame, then reattached with the opposite side facing up. Because the Working Surface now had handles and was much smaller than the original design, this procedure could be done with little effort.
Once we felt that the engineering phase was complete, a detailed drawing and BOM package was compiled. These packages went to different manufacturers to get production quotes.

The alpha prototype was built with parts both fabricated in-house and with outsourced components. Transparent blue acrylic was laser cut and bolted together with machined acrylic spacers to create the working surface. All steel and plastic parts were machined by manual mill and lathe if possible and otherwise outsourced for CNC cutting. Welding was performed with both MIG and TIG welding processes. Two complete systems as shown above were delivered to the client.
With the delivery of the alpha prototype the client also received a complete engineering release package containing native SolidWorks solid and drawing files, a detailed BOM, as well as purchased and fabricated part quotes.

Pro/Engineer was used for sheetmetal design, weldments, and surfacing the model used to create the unique vacuum-formed ABS cover.

The topper to the trailer hinges at the front for easy rear access to storage. Gas springs help assist the opening, which is especially important when bicycles or kayaks are mounted to the topper.

A number of different frame designs were analized using Finite Element Analysis. Using this method, AXN|RXN simulated 1500-lbs on different fame designs in order to maximize strength while minimizing weight and fabrication time. 

After setting up the model and applying loads to it, FEA will simulate the resulting stress and deflections in the frame. This plot shows what areas of the frame are deflecting the most.
 

The Ellipsa stroller was modeled in Pro/Engineer. Each individual part becomes it's own project when thinking about molding and other fabrication processes. Not only does the stroller have to fold, but the seat has to adjust and recline, the tray has to pivot and be removed, the canopy has to stay in the correct position. So many challenges packed into one product!

The Pro/E assembly model was created with moveable mechanical joint constraints so that it could be rotated through is entire range of motion. Any interferences during the folding process could be found and corrected in the engineering phase before any prototype parts were fabricated.
One particular challenge was to design the unit for efficient shipping box size. The client calculated what size box the unit had to fit into in order to efficiently fill a shipping container and obviously the stroller has to fit within this space. In order to balance this collapsed size with an ergonomically correct usage height, the front casters, uprights, and footrest were designed as a separate subassembly that was snapped on by the consumer after opening the package.

Another challenging addition is the introduction of all the soft goods (seat fabric, foam, covers, etc.) that cannot be modeled accurately in the computer during the preliminary design.

The Ellipsa stroller is a very complex articulating assembly using many different parts made from many different materials and processes. Casters were made from injection molded yokes with expanded foam tires and had internal rubber bushing to act as suspension. The frame was constructed from oval and round bent and welded tubing and formed sheet metal. Press fit snap features and crush ribs were designed into plastic tube ends that were to be permanently mated to the metal structure. A vacuum-formed tray liner can be removed from the tray and disposed of when it is too dirty.

One interesting area of design in the Ellipsa stroller is the Main Pivot Joint. This is the joint that locks the stroller in its open position once it is unfolded. The Locking Lever shown in orange keeps the gray Rotating Arm from moving by nesting itself into the notch. To collapse the stroller, the user pulls a pistol grip lever at the top of the handlebar assembly. This lever pulls cables that extend down the handle bar tubes and connect to cylindrical Cable Ends that slides freely within the tube. The Cable End interfaces with the Locking Lever with a pin and slot connection. When the cable is pulled, the Cable End is pulled upward through the tube and in turn rotates the Locking Lever out of the notch. The rotating arm is then free to pivot. Support Ribs were designed into the Joint Housing to help distribute the forces from the Locking Lever.

A “looks-like, works-like” Alpha Prototype was built in-house from the Pro/E models. Certain complicated features that were difficult to prototype were sacrificed due to time constraints. Most parts were created from SLA rapid prototypes or CNC machined from Ren-Shape foam. The prototype was critical in testing things that cannot be easily designed in 3D CAD software such as soft goods and ergonomic usage issues.

The surface model for the roto-molding pattern started with three orthographic views from the Brooks Stevens design staff and a breadboard mockup for reference. The mockup was used to debug the coiled hose pocket, chemical bin sizes, retractable hose clearances and for general ergonomic concerns. Modeling for roto-molded parts is easier than for injection molded parts because internal geometry is not important for tooling. The model can be left as a solid and need not be shelled or internally ribbed.

The Pro/E model was sent off to build a mahogany pattern for tooling. All shrinkage is calculated and put into the wooden pattern, then it is used in an aluminum casting process for the roto-molding tool. The tool is cut along precise parting lines so that the finished plastic part can be easily pulled from the mold. The above image shows the order that the mold is pulled from part and is colored to show where the parting lines would occur.

The key component to this system was the IV pole itself, shown above in it's rolling configuration. This pole features a large accessory hoop for attaching pumps and other equipment, an oval IV bag hanging hoop, and an all stainless and aluminum design.

One engineering challenge behind the Streamline pole was to make a stable base in the deployed position, but also to decrease it's size when it was to be attached to an ICU bed. The solution was to use a 4-bar linkage mechanism to collapse the legs, much like what is found in an umbrella or music stand. This mechanism was actuated by an internal gas spring. So, once the pole was secured to the bed, the red foot pad was depressed which in-turn released the spring and retracted the legs. To redeploy the legs, the user simply stepped down on the stainless foot pad until the legs locked back into their original position.

You can see from the picture above how the base of the IV Pole rides up the main pole until the legs are in the collapsed position. The tubular stainless positioning links were designed to give the casters enough room to swivel out of the way of the mechanism when retracting and create the smallest possible footprint.

The second part of system involves mounting the IV pole onto the bed or wheelchair. The simplest solution would normally be to bolt a rigid mount to the bed and then secure the pole to the mount. However, in practice the location of the pole as the bed is traveling in the hospital is of great importance. In many cases staff has to navigate around other beds or into tight quarters such as elevators, so the ability to move the mounted pole around the bed is essential. I traveled to a local hospital and measured a typical bed in order to build a mockup and test these situations in the shop.
The solution was to use a pivoting and telescoping arm mounted to the end of the ICU bed. The arm can be swiveled 180 degrees and an internal flat spring returns the telescoping arm into five different nesting positions around the perimeter of the bed.

The above picture shows the Streamline IV Pole when mounted to the articulating arm. In the shown position, the user can easily reach the ICU bed drive controls and the retracted legs allow foot space for walking on each side of the base.
To disconnect the IV pole from the bed, the user releases the locking handle, then pressed on the foot pad to extend the legs. The pole is then released and able to be rolled around freely.

The COBRA baggage scanner was featured in Wired as a new product in the security industry. If you travel by air frequently, you'll notice that the current generation of scanning equipment looks much more like a gym locker than high-tech equipment. The COBRA unit stands to change all that.

To access the complicated internals of the COBRA system the outer skin had to be intelligently divided up in multiple panels made from fire-rated fiberglass composite. There were many different engineering challenges to tackle. For air flow and cooling issues each panel had to seal against the next, so all panel had to have a series of overlapping joints to hold gasketing material. Also, internal bracketry had to be designed to connect each panel to the underlying sheet metal lattice structure. This structure was not always square, so adjustment features had to be designed into the bracketing system to make it look correct in the field. Certain panels were designed to be quickly removed and replaced in case a bag jammed up the system. Other panels had to be designed with key locking features or sheet metal filter housings or electronics.
Each of the panels shown above was used to create a fiberglass tool models for in-house prototyping. Molds were CNC machined from Ren Shape blocks and hand laid-up with fiberglass composite.
The height-adjustable user's workstation was necessary to accommodate all different sizes of TSA employees. An electronic actuator with push-button controls powers the unit up and down. The workstation is fabricated from formed and welded stainless sheet and uses an internal bearing block and rail system for smooth operation. The workstation provides two 20” flat screen monitors, keyed system lockouts, an emergency stop, and a writing surface.

After the alpha prototype was shown at a security convention in Chicago, we plunged into Beta prototyping efforts. Major adjustments were made to the tunnel opening size and to the bag jam access doors. The Pro/E model was updated and new fiberglass tools were cut and sent to a professional fiberglass productions house. I spent a number of weeks helping to trim finished panels and fit them to a prototype framework that we built on the production house floor. After all the panels were painted, finished, and fit, they were shipped off for Beta testing. The above photo shows the unit undergoing testing at Logan International Airport in Boston.

The elevator system was broken into two analysis projects. The first dealt with the sling, which is the L-shaped rolling carriage that lifts the elevator box to the different floors. The second dealt with the rail system that connects the sling to the load-bearing wall as well as the safety braking system that catches the elevator in case of cable breakage.

ASME code states that the sling must have a safety factor of five at its maximum rated load. The sling assembly was analyzed at maximum rated load using both Finite Element Analysis and hand-calculations to double-check the results.

The stress plot above shows that there are significant areas colored red, which indicates that the stresses are over the allowable limit. Hand calculation verified these results.

In order to bring the stresses back within the code, supports plates and gussets were added to the design. These gussets boxed in the c-channel arms of the sling and helped to distribute the forces to neighboring areas.

This plot shows how the areas over the stress limit have been diminished.

As with the sling assembly, the rail system also had to be analyzed for agreement with code. The code states that the static stresses in rails system and all of its components must stay below 15 ksi and have less than a quarter-inch of deflection when the sling is in a position to create maximum load on the rails. It also requires that during Emergency Braking (say, if the cable snaps and brakes engage) all stresses in the system must be below 27.5 ksi.
The first part of this problem was to find out exactly where on the rail system the sling had to be positioned in order to create the maximum stress. The sling moves along the rails much like a train or roller coaster moves on its track. The maximum stress in the rail depends on where the rollers on the sling are positioned relative to the wall supports along the rail. A spreadsheet was set up using continuous beam theory in order to plot the maximum stress in the rail versus the location of the sling rollers. As it turns out, that the maximum stress occurs when the rollers straddle the wall mounts. Now this positioning was known, it can be used for all the following calculations and simulations for maximum stress and deflection.

The second part of the problem was to figure out the dynamic load put on the system as the elevator cab drops and the safety brakes engage. The rate of this acceleration has an enormous effect on the stress levels in the system, so an accurate measurement of this figure is needed.
If we knew the elevator cab weight, how far the cab dropped when the cable were cut, and how long it took the cab to stop once the brakes engages, then we can calculate the average acceleration on the cab using kinematic equations. So, we designed a series of drop tests and measured the total distance the cab dropped, as well as the “skid mark” that the knurled braking rollers left on the rail as they slow the cab. Now, we had all the information we needed to calculate the additional force put on the system during emergency braking and this value could be used in the following FEA simulations and hand calculations.

As the safety brakes engage, the elevator sling rollers produces a large moment on the supporting rails. At the top of the sling, this moment causes the rollers try to pull the rails and wall brackets away from the wall, and at the bottom it pushes them into the wall. The brackets were designed well to be loaded with additional support from the wall, but lacked strength when being pulled away. The above image shows the pulling stress plot. In this situation it was found that the brackets contained stress in excess of the code. To bring this bracket into agreement, a supporting gusset had to be designed that could be installed easily in the field.

Because the wall brackets had already been produced, the reinforcing bracket had to be easily retrofit to existing systems. I designed a gusseted support bracket that could be fastened in at existing bolt locations with no drilling or cutting needed. The plot above shows how the support eliminates the high stresses in the bracket. The plot does shows high stresses around the bolt hole, but this is a normal situation that arises around fastening points when using FEA.

ASME A17.1 states that the stresses in the safety components must have a Factor of Safety of at least 3.5. The mechanism works by using a wedge-shaped steel block, a knurled roller attached to a sprung armature, and the support rail that the elevator rides along. If the lift chains lose tension, the armature drives the knurled roller into the wedge created by the safety block and the guide rail. As the weight of the elevator cab drives the safety block downward, the knurled roller wedges itself tighter and tighter and quickly stops the cab from falling.
The above stress plot shows a simulation of this situation. It seems worrying at a first glance, but the areas above the limit in red can be explained. The loading situation between the braking roller and the safety block is a line-on-surface contact situation and will always cause a stress riser in the plot. The other areas of concern are located around the bolt holes, but since these holes were used as fixture constraints, high stresses in the plot are expected. Because the sections of red are concentrated to small areas of expected high-stress, the design was deemed to fit within the limits of the code.

AXN|RXN started by dropping the designer's orthographic views into SolidWorks so I can be sure that the final surface files will match the original design intent. The above picture shows the three orthographic views and a few of my initial setup curves. Modeling continues in this manner and build a complete wireframe of the product. Then, patching in surfaces from the wireframe curves arrives at the final model. Because all of the curves are based off of the designer’s sketches, any adjustment they want to make to these sketches can be quickly updated through the model.

This three-piece injection molded ABS/Polycarbonate design includes an aesthetic chrome-plated top bezel that is heat-staked onto the top shell. The reflectivity of the plating accentuates the red "in-use" LED display.