Anablog

The Tuesday tech panels at Convergence were in three tracks. The first track was a safety track, the second was about sustainability and the environment and the third track was about standards. Standards are essential to any good engineering regime and one of the complete failures of the auto business is in setting standards. When the government mandated that the cars all have the same connector for a test interface, the idiots in the auto business still managed to have three variances, so each little batch of middle-manager petty bureaucrats could keep their connector proprietary. They spend more time trying to insure you cannot take the car to an independent garage then they do making sure the transmission does not fall out of the car in 40,000 miles. The government got sick of this and mandated a connector, but it still has user-selectable pins and no, the Shrinking Three still could not agree to standardize those pins. Sorry for the harangue, I just really get incensed when I see still more reasons for why I had to give up being an automotive engineer and move to Silicon Valley. So I didn’t bother going to a single standards meeting since standards are a joke in the US auto business and spending any time talking about them is a pure waste. Maybe when all three US automakers combine into one frail bankrupt company there will be one standard, but you can bet it will still be different that European or Japanese standards.

I stared out the day with a talk from Motohiro Tanaka from Mitsubishi. It was on eco-conscious design. It was pretty much a big corporate buzzword program where they assign points or cookies and orange juice and a pat on the forehead for technologies that are considered eco friendly. It was pretty interesting and I guess it is nice Mitsubishi has taken up an interest in the environment. When it was obvious that the main eco-friendly criterion was CO2 production I got disgusted and walked out. I wish my personal ethics allowed me to drink alcohol at these conferences since I would have gotten drunk. I guess it is easier to spend all your corporate resources trying to limit plant food rather than, oh, polyphenol bichlorides or all the other really nasty things that corporations spew into the environment. The message for the new millennium is “Sorry about the cancer clusters, but look at how small our carbon footprint is.” Yippee.

The next talk I took in was from Toyohei Nakajima of Honda. He talked about advanced driver assist systems. This was the best-attended talk I saw so far. He discussed radar avoidance and other safety systems like night vision and pedestrian recognition. He noted that safety systems were the last things people are willing to pay for but rather than ask for laws he pointed out that the systems could be designed so people actually want them. I had an old-industry moment when an SAE watchdog stopped me from taking pictures—yeah, give me a free pass and then stop me from promoting the show, thanks SAE. So the one picture I did get before being busted was about lane assist technology.

Honda uses a camera and a processor to determine when the car is leaving the lane. It will apply torque to the steering wheel to keep the car in the lane and if the car exits the lane there will be a warning. It is a bit too Astro-Boy for me, but anything Honda does is worth watching. Unlike the US auto companies, Honda lets engineers design the cars and finance guys count the money, not the other way around. Not bad for an outfit founded by a garage mechanic that never went to college. Yeah Soichiro. One impressive fact in the talk was how night vision systems can promote a 60% reduction in driver reaction times. He also discussed the various semiconductors useful for radar—gallium nitride (GAN) looks promising but is expensive. Gallium arsenide is way expensive; silicon germanium is promising especially since it can be combined with CMOS for processing on the same chip, and CMOS with MEMS which was still in the research phase. Toyohei pointed out that Honda found that a heads-up-display (HUD) was the preferred method to interact with the driver, but noted that BMW and Mercedes used two different systems. He called for some standardization of the interface, which is pretty funny since, as I pointed out, the US auto industry never standardizes anything, and the Germans would never ever ever ever adopt a Honda human interface even when there are reams of data to prove it is superior.

The next talk was about environmentally friendly plastics given by Troy Robinson of Yazaki. What I liked about this was that they knew the plastics also had to be cheaper than normal if they expected the auto business to buy them. One technique was mixing natural fibers with polypropylene (PP). Their product was PP/kenaf. By mixing these plant fibers into the plastic at concentrations of up to 60% fiber, the properties of the plastic composite are actually improved. Troy pointed out that short fibers, under a mm long, are easier to process and break up further in the injection molding process. The other green plastic was a blend of polylactic acid derived from corn along with poly polypropylene. The properties of this blend are not very good but can be acceptable, and PLA is biodegradable so that is a nice plus. Troy had the winning operative comment about the auto industry when he said “Cost is everything”.

Next up was a great talk by Peter Fiske, who has more degrees from elite colleges than I have toenails. His talk was about biomimetic design. What biomimetic design does in copy the logarithmic swirls and simple structures of nature in order to improve on the linear and boxy designs that are easy to do and analyze. For doing fan design, the big advance came with cheap computational fluid dynamics (CFD) software that lets designers see how swoopy designs will work. Peter started out with the joke that auto engineers want biodegradability right after the warrantee has expired. Actually maybe the finance idiots do, but every auto engineer I have ever known wants his stuff to last a lifetime.

Peter pointed out the invention of Velcro came about when George de Mestral looked at the burrs stuck in his socks after a hunting trip. Peter also noted that Leonardo da Vinci had 140 wing designs, mostly variations of bat wings. It seems that engineers abhor turbulence but nature adores it. A human heart uses turbulent flow to close the valves and consumes 1.25 watts. Peter also gave credit to Bowles Fluidics, who made a windshield washer that used a turbulent flow cavity. Peter than gave examples of natural systems like the fuzz on owl wings making them quieter, how the tubicles on whale fins allow a greater angle of attack and how Mercedes designed a car based on the box fish. Peter did go a little commercial when he pointed out the successful fan designs his company, PAXFan,  has developed for refrigerators and cars, but these were interesting in their own right — all of them were much quieter and used way less energy or both. Peter finished up telling us about the asknature website, where these concepts are discussed further.

The last talk I saw Tuesday was about how remanufacturing electronic modules can keep support up for older vehicles. This support is essential for over the road trucks and other products that need ten and twenty year lifetimes. The talk, given by Joseph Kripli of Flight Systems Electronics was a real eye-opener. His company is not an aftermarket rebuilder, he works with the OEMs and tier-one supplies to make sure there are service parts for cars. He noted that there are 16 modules in an average vehicle as well as 25 microprocessors. He pointed out that the semiconductor companies are great at selling you new stuff but once the volumes dorm below a million (or a billion) per month, they obsolete the products forcing him to do a lifetime buy. He gave an example of a Cadillac escalade part that went obsolete 6 months after the module went out of production. I blame this on both the semiconductor companies and the engineers that uses custom parts rather than plain-vanilla design with standard parts. Joe says they take in 20,000 module part numbers and then upgrade and homologate them into 11,000 part numbers. He sited one module that had 50 software changes in one year. I knew that Adobe was incompetent but I had no idea that embedded systems designers in the auto business were also unable to finish the software before shipping the product. Joe noted he still does 10 ignition modules a month for a 1978 Chrysler Cordoba.

Joe also noted that may of the modules just have a cracked capacitor or diode and can be simply fixed and put back into service. This is environmentalism we all should support. Joe did note that it was essential to test the modules at three temperatures, -40, room, and +125 degrees C in to insure that the parts would pass warrantee, which is the same as a new module. He also pointed out that they have patented a device that uses water jets to cut out the potting in many modules.

Joe also stressed that it is essential to upgrade the software on modules. He pointed out that a 1996 Tarus ABS module will stop the car in 80 feet. A 1998 module does it in 60 feet. The only difference is software. He said if you just buy an aftermarket module you are never sure if you are getting the 80 foot part or the 60 foot part. One interesting fact was that 20% of the modules he gets in are OK, with no fault found. All in all this was a great talk and stressing a design factor that nobody seems to care about—serviceability. If we can keep older cars on the road it saves a lot more energy that a Prius since you have to melt the steel and mold the plastic to make a shiny new car as opposed to just keeping an older one on the road.