Boom Time

On Tuesday and Wednesday, Kris and I traveled to Buffalo, NY to visit the HYGE Crash Test Sled Facility at Calspan. The company very generously offered to test Mechanek in a Frontal Crash Test. 

Tuesday was setup day. We designed the test to be as close to SFI 38.1 as possible - the standard developed for the HANS device. 

                                                                                      Crash Test Dummy no. 366 awaiting his fate

We conducted two tests - one with an off-the-shelf HANS device and another with the Mechanek device. In both tests, the gas piston was charged to 2600PSI, resulting in a sled acceleration of around 40g's, reaching ~35mph in less than 50ms. 

The Anthropomorphic Test Device (ATD) was fully instrumented with a 3-axis head accelerometer, a 3-axis thorax accelerometer, and neck load cells, all collecting data at 2kHz. Calspan trained four IDT high-speed cameras on the sled for both crashes, each shooting at 1000fps. 

The three videos above show our first test with the HANS device. 

Results from the tests were overwhelmingly positive. We had hoped to prove similar performance between the HANS and Mechanek devices. We ended up showing that Mechanek, in proper usage and operating conditions, can show a ~30g decrease in the resultant head acceleration for the specific crash pulse tested. 

                                                      Test 1 - HANS Device : Notice the ATD Head Resultant Acceleration Max = 122.70 [g's]

                                                      Test 1 - HANS Device : Notice the ATD Head Resultant Acceleration Max = 122.70 [g's]

                                                    Test 2 - Mechanek Device : ATD Head Resultant Acceleration decreased to 94.58 [g's]

                                                    Test 2 - Mechanek Device : ATD Head Resultant Acceleration decreased to 94.58 [g's]

Thermoforming PETG

Today we thermoformed a PETG glide plate for the back of the Mechanek Device. The mold was made from sheets of laser cut MDF coated with XTC-3D to create a smooth surface to help the formed plastic release from the mold. 

More machining + fabrication

We spent more time on the Haas MiniMill finishing up a crucial component of our device. 

After a lot of sanding + painting, we also ended up with the part seen below: 

This part was printed on the Dimension Elite FDM printer in ABS. Repairs were made with the XTC-3D compound, then the entire part was coated with sand-able primer and wet sanded to 2000grit. It was then painted with Montana Black and given a final coat of matte varnish. 

Directions for finishing 3D printed parts in a similar fashion can be found here


Buddy in the Physics Machine Shop at Penn was willing to let us use their sandblaster to finish many of the parts we've machined. Most importantly, we needed to sandblast the quick locks we machined yesterday, to soften the edges and make sure that the polypropylene tethers won't be cut by them. 

Before sandblasting: 

During Sandblasting: 

And post sandblasting: 

Quick Lock Machining

Kris developed some 3-axis CAM for the Quick Locks that will interface with post-anchors on the helmet. We needed to manufacture our own quick locks instead of repurposing a HANS quick lock because we're using 1" width tethers instead of the 3/4" tethers used by the HANS. The HANS also uses aluminum quick locks, where we decided to machine these from 304SS. 

Post processing will include sandblasting to soften all the edges before we send them out to be stitched in to our tethers. 

Electronics Progress

Today we began testing several electronics subsystems. For crash sensing, we're using two Sparkfun accelerometer breakouts. The first uses the ADXL377 chip, giving us a +/- 200g sensor that communicates with the Edison with low level ADC. The second chip is the ADXL345 +/- 16g sensor, which is communicating with the Edison over I2C protocols. The nice thing about this chip is it allows us to trigger an interrupt when a "rising" signal (aka a crash) is detected with a magnitude above a preset threshold. This has the advantage of pulling the data sensing out of a while() loop in the code. 

                                                            ADXL377 +/- 200 g Accelerometer (left) ADXL345 +/- 16 Accelerometer (right)

We've also received our new HANS device

CNC Turning

Today was spent making a few test parts on the Haas TL-1 CNC Lathe.

We also post-processed the ratchets we got from SDP/SI on the mill, adding mounting holes and increasing the size of the center bore.