|
Mechanical Forensics Engineering Services, LLC, 179 Cross Road, Rochester NH, 03867
|
|
Event Data Recorders and the Bosch Crash Data Retrieval System
|
If you want to BUY one of these systems, please don't call me! Visit the distributors at http://www.cdr-system.com/, where purchasing a system sometimes gets you free tuition to a class on how to use it, which is a REALLY good thing. Alternatively, you could go directly to Bosch's CDR website. Either way, don’t call me, because I don’t sell the kits, I'm just a humble user. Also, please note that this technology is changing almost every day, so I'll do my best to keep this page updated, but new information may come out at any time, consider yourself warned! There, with all that out of the way, we can move on to more important matters.
INTRODUCTION
The generic term for any device which records information about particular events, such as crashes or airbag deployments, is commonly called (intuitively enough) an Event Data Recorder, or EDR. There are some dedicated OEM and aftermarket EDR systems for airplanes, Railroad Engines, ships, and heavy trucks. There are some aftermarket passenger vehicle EDR-systems which are usually used for fleet applications. I'm not going to discuss any of those. No passenger vehicles currently sold in the US have a dedicated EDR system. The closest thing we've got is that some cars have some EDR functionality built into various components. This functionality is usually found in the airbag contol module (ACM), but may also be found in the engine controller or Powertrain Control Module (PCM) of some vehicles. THESE are what I'll be discussing on this webpage.
For ACMs the real job is to monitor the car's status, determine when the car is experiencing an event which meets airbag deployment criteria, and deploy the belt tensioners/airbags/etc. in a timely fashion. The module is working in a black closet with almost no information about the car and none about the environment and is trying to predict what's coming! The EDR-aspects of ACM (and PCM) functionality have been a secondary job at best. The goal of such recording was initially contemplated as a means of providing information to investigators about the actual crash environment the module was experiencing when it decided to either fire the bags or not. Better understanding of how crashes occur in the real world and what causes people to get hurt can lead to an improvement in overall safety for everyone. The diagnostic kit required to access the currently available crop of EDRs is sold by Bosch (who purchased this business unit from Vetronix a few years ago), and is called the Crash Data Retrieval System, or CDR-Tool. Because Vetronix originally sold these devices, you still sometimes hear their name come up, but they are out of this business, AFAIK.
These EDR systems do not record audio (voices) or video, direction of travel, location, date, time, who is in the car, phone operation, or any history regarding where you've been or how fast you have driven (except for the few seconds prior to a crash, as noted below). There may be some other systems which contain some of those things, like OnStar, but the "black box" of the EDR does not, and that's all I'm discussing here, and it's all the CDR kit can access. The data from these EDR modules has been used in crash investigations across the country to help determine what happened during a crash. It has also been used in legal matters to clear or convict people. See http://harristechnical.com/cdr.htm for a list of legal cases and state laws regarding EDRs. Questions of data ownership and accessibility to various types of investigators has also been a topic of much discussion. These issues seem to getting settled slowly on a state-by-state basis through legislation and case law.
Up to now, manufacturers have each approached EDRs in their own way, but in August 2006, NHTSA issued a rule governing what data an EDR must record when a car has one. In January 2008, they issued their updated Final Rule which addressed some of the petitions they got after the first Rule was announced. It now applies to cars built after 01SEP2012. It mandates a list of things that an EDR must record if the car has an EDR which is intended to be downloaded after a crash. Think about that for a moment: if a car doesn't have an EDR intended to be downloaded after a crash, then the rule does not apply. As with many laws, this 200+ page beast will require some interpretation as it gets implemented. Many cars, though, already have some form of EDR in them, and NHTSA has said that if manufacturers stop putting EDR functionality in their cars, NHTSA will revisit the issue of mandating them. The regulation is technically called "49CFR Part 563", but I'll just call it The 563 Rule. The actual text from the 14JAN08 Federal Register can be downloaded in PDF format by CLICKING HERE. you can find the original (now obsolete) rule online if you look about a bit.
In addition to specifying what data must be recorded and at what rates, the 563 Rule includes some survivability requirements (which should be a piece of cake for current EDRs, IMHO). The "must-record" data includes some pre-crash data as well as a lengthy crash pulse. The Rule also requires manufacturers make a tool or translation device to download that data available to the public within 90 days of a car’s release for sale. It does not require them to give away these decoding devices, but rather that they be commerically available. Affordability is not addressed. The Rule does not require that Bosch be the outlet, and I am sure there will soon be more players in the EDR-downloading equipment market. The Bosch kit may continue as the only device to access GM/Ford/Chrysler/some others, but there will probably be kits from one or more other companies (i.e. OTC, B&B, Drew, or the OEM themselves) to access the other dozen or so manufacturers whose products will be covered by the rule.
To the right is an example of an airbag control module from a GM vehicle (click on the thumbnail for the full picture). GM calls these units Sensing and Diagnostic Modules (SDM), while the Ford version is called a Restraint Control Module (RCM). You'll notice that it's not BLACK, despite the common terminology. They are almost all silver. The SDMs measure about 4x4 inches across, and are about 1.5 inches high, with a big yellow plug on one end and an arrow to indicate FORWARD as installed. Yellow plugs are used throughout the airbag system in all US cars, perhaps to make those connectors easy to find. MFES is one of the several hundred companies nationwide equipped with the CDR system (shown in the photo to the left), having been one of the very earliest adopters of this technology. The CDR system can download the airbag module data from some GM and Ford vehicles after an airbag deployment or in some cases, a non-deployment event. It is the same tool used by NHTSA, GM, and Ford in their investigations. The coverage, locations, what pre-impact and post-impact data gets collected, and best data retrieval procedures vary by manufacturer, so this page will discuss each separately (GM / Ford).
|
DOES THIS VEHICLE HAVE ANY EDR FUNCTIONALITY AND IS IT COVERED BY THE CDR-KIT?
| Version 3.0 shipped in January 2008, which included the addition of several Chrysler products. Version 3.5 is shipping as I write this. The vehicle coverage list for this and earlier software versions can be found on the Crash Data Group website. |
General Motors (Cadillac, Chevrolet, Buick, Olds, GMC, Pontiac, Saturn, Saab, Geo, Hummer, Isuzu)
Of all the OEMs, GM has the largest number of models with publicly accessible data. Some data exists in some GM models as early as 1994, while models newer than 2000 generally have all the pre- and post-impact data discussed below, though there are some exceptions to both rules. The Isuzu and Saab coverage is very limited. |
Ford (Ford, Lincoln, Mercury) Ford coverage on some models goes back to 2001, but is much less comprehensive than GM, though the list of covered vehicles is growing with each update, including the about to be released Ver2.9. There are a bunch of Ford RCMs not covered by either 2.8 or 2.9 that we will probably never get access to for what have been termed "internal/supplier" reasons, though some of them may get PCM coverage eventually. See below for more on PCMs.
|
Chrysler and Bosch inked a deal in 2006 or so, and coverage has been slowing coming along starting in early 2008 with Ver.3, with some models covered back to 2005. This does NOT include any Mercedes models, as they are two separate entities again. The Chrysler add-on included a LOT of new direct-to-module cables. Even though I have a software update subscription, the new hardware cost me $1300. Ouch.
The Owner's Manuals for various '06 and '07 Dodge vehicles contained this interesting blurb: "In the event of an airbag deployment, your vehicle is designed to record up to 2-seconds of specific vehicle data parameters (see the following list) in
the event data recorder prior to the moment of airbag deployment. Please note that such data are ONLY recorded if an airbag deploys, and are otherwise unavailable...Parameters that May Be Recorded: DTCs, “Time” of airbag deployment (in terms of ignition cycles and vehicle mileage), Airbag deployment level (if applicable), Seatbelt status, Brake status (service and parking brakes), Accelerator status (including vehicle speed), Engine control status (including engine speed), Cruise control status." I have heard from others who should probably know about such things that Chryslers prior to that really have nothing, and furthermore, Chryslers only record events. There is no recording of non-deployment events. To visit a page that discusses this issue with respect to JEEPS, CLICK HERE,
All Other OEMs According to NHTSA, "65-90% of model year 2004 passenger vehicles" of most manufacturers have some EDR capability built into either their ACM or other on-board systems. NHTSA also noted that "Honda, BMW and other vehicle manufacturers record small amounts of crash data." while they cited Ford, GM, and Toyota as having "what would be considered a large amount of crash data." The Bosch CDR kit is the only publicly available tool to download crash data, and the coverage list cited above should be examined for coverage. The OEM's ability to read their own data remains a hot topic of debate.
Honda: One '05 CRV download done in Japan for a Northern Hemisphere client took months to get, but provided Seatbelt status for driver and right front, Driver's seat position, Passenger seat empty or occupied, Which airbags deployed, and Delta V. The '06 MDX and Acura TL reportedly has seatbelt buckled, seat weight sensor, brake on/off, but no collision pulse. Earlier modules often just have "Parameters for deployment were met" or words to that effect. Some new-platform '06 models reportedly can be accessed by a dealer with the "new computer system" which can scroll through a number of recorded parameters if the battery has not been disconnected for too long. After a time without power, though, this data may go away.
Toyota, long rumored to be in talks with Vetronix, stepped away from the table back in 2007 or so. Downloads used to require a visit to or from factory personnel, if you could get it done at all, but this is changing as a result of the 2009-2010 sudden acceleration issue. Toyota has now supplied multiple download kits to NHTSA, and perhaps others working on their behalf, for use in their investigations. They have generally stated that they do not believe the system is reliable enough to serve any reconstruction purpose, so they won't provide the data. I don't think anyone in this business really believed them about the uncertain reliability part back then, and this position may come back to bite them, I think, as they now try to use EDR data to support the position that sudden acceleration cases are not the car's fault.
As of early 2008, there were rumors of a different “Asian” company coming onboard with Vetronix soon, but it hasn't happened as of July 2010.
An '06 Mercedes E350 involved in one case reportedly was downloaded by a road-service-technician from the dealership after some sort of side-event. The data he could access was associated with fault codes for the side curtains or side bags, and included the odometer reading at the time of the event, the current odometer reading, the speed at the time of event, abs on/off, the number of ignition cycles without the fault reoccurring, and brake status (ON/OFF). It's not clear how confident we could be that the speed a fault code being recorded was the speed at the event itself. Perhaps we'll get to do more research on that once the kits become available.
Factory download reports for '04 Volvo XC90 show delta-V, crash pulse, vehicle info, and some diagnostic stuff. I thought it was interesting. Volvo continues to be inaccessible by public tools, though.
Nissan/Infinity have data that might be accessible by the dealer, but more likely by the manufacturer directly. The modules out there so far reportedly do not have speed, but it might become an available parameter in 2007.
Some Hyundai data may be available either by way of the dealer's factory scan tool, or the OTC i-pro scan tool. The OTC website used to say that this tool has "ABS and air bag crash data for OEMs that support it" though it's not clear what that really means, and is no longer listed on their website (July 2010), and some internet chatter I found suggests that the software was not updated after 2002, so maybe not so much, eh? One wonders about their Genisys scanner, though.
Reportedly, Audi A3's could store up to 3 crashes, but newer ones get only 1 before requiring replacement.
BMW has publicly stated that they don't record data and won't do so until mandated to, because their customers don't want it. That said, I know one law enforcement guy who has seen the printout from a 1999 M3 with speed, occupant status, and some other stuff. Either the M3 was really special, or there's something there that us mortals are just not going to get access to.
If the case is big enough, and you've got a search warrant or similar pull, Kia will send someone out from Kia of America in California to download the codes in a crash-involved event. In one pedestrian case with no deployment the effort resulted in no useful data, but in one case with a 2006 Kia Spectra it included an acceleration and delta-V. It's not clear what data might be in there for deployment event for other models or years.
Many observers anticipate big changes in this arena with the advent of The 563 Rule, when most everyone will have to make the data available. If they drag their feet too much (either by making public-tools too hard to obtain or disabling their non-compliant modules), NHTSA may step in the way they did with OBDII regs, and nobody wants that, so compliance seems likely, after a fashion.
|
WHAT DATA MIGHT I GET WITH A CDR-DOWNLOAD?
General Motors Pre-crash
The data for essentially all pre-2000 GM cars contained no pre-crash data, but include driver's seatbelt status (buckled/unbuckled), some airbag system diagnostic information, and the crash pulse. The 2000 and newer cars generally include those things along with the vehicle speed, brake status (ON/OFF), throttle position, and engine RPM for about 5 seconds prior to impact. Though there are exceptions to both of these generalizations.
|
Ford Pre-Crash Currently supported Ford RCM modules do not record any significant pre-crash data, but they do have driver's seatbelt status (buckled/unbuckled), some airbag system diagnostic information, and the early part of the crash pulse. The recorded crash pulse can include a tiny portion of the pre-wake-up accelerations, but that’s it. See warnings regarding the PCM for some vehicles below.
|
General Motors Post-crash
SDMs record the calculated post-impact delta-V, and report it in 10ms intervals for up to 300ms in some modules, less in others. Seatbelt switch status (buckled/unbuckled), airbag warning lamp condition, key cycles at deployment, key cycles at investigation, deployment command times, and some other parameters may also be recorded, depending on the module type & manufacturer used in that vehicle.
This is what the Delta-V (change in velocity) data may look like when it is recorded:
What data is probably available in a vehicle of interest varies by vehicle model and year. The Vetronix website used to have a list of what you could generally expect from each vehicle, but that list seems to have gone away with the Bosch takeover. Your best bet to learn what data you might get from a specific vehicle is to contact someone who has the system and ask them to check the HELP file in their software. Also when you get a file READ THE DATA LIMITATIONS section of the report. It details the idiosyncrasies of that particular module. It is the ‘must read” part of any report.
Some of the newer Epsilon-class cars may also have information regarding steering wheel position, ABS activation, both Throttle Plate position and Throttle Pedal Position, and more.
|
Ford Post-Crash RCMs record the acceleration pulse and the calculated delta-vee in the longitudinal direction each 2ms for approximately 70ms after Wakeup, and in the case of 2001 Taurus/Sables also in the lateral direction (which I think is only accessible with the right early software version). Additionally, RCMs record deployment command times for driver and passenger first and second stages (for both frontal and side airbags), DTCs when they exist, and more.
If the second deployment stage isn’t necessary for a particular crash, it will be “disposed of” about 100 ms later in order to ensure that no unused pyrotechnics remain in the car after a deployment. The timing data will tell this tale.
An example of Ford data looks like this:

Several researchers have noted that the current Ford modules almost never get a full crash pulse due to their short post-wakeup recording length. The length is enough to get pertinent information about the vehicle environment in the part of a crash where an airbag might help, but doesn't capture any of the long-pulse stuff where airbags probably shouldn't deploy anyway. This is part of the philosophical difference between Ford and GM EDRs. These differences will go away with the advent of the 563 Rule. More detail on recorded data will be forthcoming, after I get to play, uh, experiment, some. |
Some GM SDM Nuances
- Hint: READ THE DATA LIMITITATIONS in a download report.
- SDMs can store two types of events: Deployment events (which can not be overwritten) and Non-Deployment events which can be overwritten, or will be erased after 250 (or so) ignition key-cycles are recorded.
- Years ago, Non-Deployments were called “near-deployments” which was found to convey the incorrect sense to some people that the bag was “about to deploy,” this is completely wrong, and the terminology was changed to avoid some of those problems. Some of us still inadvertently use the old term sometimes, though.
- The SPEED recorded by the module is the speedometer indicated vehicle speed, not the OVER THE GROUND SPEED. This can be an important distinction particularly during non-ABS skidding (where the speed reads ZERO regardless of how fast the car is moving) and occasions when the wheels may have been airborne or on very slippery surfaces while either the throttle or brake were applied. It also important to remember that wheel slip induced by hard ABS braking or yawing events can cause the indicated speed to be different from the actual speed.
- Many modules record the number of keycycles or ignition cycles the car has experienced at the time an event is recorded. One keycycle is recorded each time the module is powered up for a few seconds (long enough to perform its self test). See the Sunfire/Deer example below for an example of this data: The module was downloaded three ignition key cycles after the event (from 10,170 to 10,173).
- If the module is left in the car after a deployment, even though the SRS-light may remain on after repairs are made, it will probably try to fire the bags if involved in another accident.
- There are several different suppliers for these modules, and sometimes units from different manufacturers record different data.
- A very few modules have 1024 bytes of EEPROM memory instead of the more common 512 Bytes. These can actually hold up to 3 events, including the SDM A/B/E (95/96 Lumina and Olds) and S (95 Saturn) as well as the 96/97-ish Cadillac SDM-CL2. None have pre-crash, but have similar data to the SDM-R series, i.e. delta-v, seatbelt, SIR light, etc...
- Data buffering issues can sometimes leave odd-looking post-crash information in a file. This is most commonly seen in cases where the crash pulse was really long.
- Newer cars on the Epsilon platform using the CAN-bus can only be downloaded with the newer blue-colored CDR box with the newer software, and sometimes offers a bunch of additional interesting (and a lot of uninteresting) information with regard to vehicle status before and during a crash
- For several years now, modules have reported an "algorithm adjusted Delta-V" instead of simply "Delta-V". The software essentially has a running re-zero operation to account for rough road operation (such as on dirt roads) when trying to determine when airbags must fire. During really long pulses, some researchers have shown this to have a measurable effect on total reported delta-vee.
- With newer systems that have a bunch of available options, How does one determine if it is "equiped"? The answer is that generally it IS equipped with that option if that data slot contains a number or zero (anything other than N/A); If there's an N/A, then we need to check the Engineering Translation Report (ETR) for further information. Sometimes we have to actually go look at the car with a Tech-2 or MasterTech in hand in order to be sure.
- The new ACMs wake up with acceleration pulses from any direction (not just rearward as with older units), and begin to run the algorithm to determine if it's a Non-Deploy or a Deploy event.
- Malibu CLASSIC and Malibu (MAXX) are different animals: The Classic (4th vin digit=N) has long been supported while the newer GMX380 platform (4th vin digit=Z) is a CAN-bus vehicle requiring the ver 2.8 or newer hardware to access.
- Chevrolet (in partnership with Suzuki) builds Aveo, Optra, and Epica models, as well as the Pontiac Wave which are rebadged Daewoos (after Daewoo went bankrupt in 1999). All of these have Siemens brand ACMs, which are not supported by the CDR system.
- One last item: READ THE DATA LIMITATIONS that come with a report.
|
Some Ford RCM Nuances
- Hint: READ THE DATA LIMITATIONS that come with a report.
- The speed-warnings noted for GMs pretty much apply here, though the actual values have not been as well documented yet. I'm sure they will be in time.
- The PCMs in '03 and newer Fords with Electronic Throttle Control and Gas engines may have a crash-related data in a circular buffer which can be overwritten by ham-fisted treatment of the car after the fact. The over-writing may occur during power-up for scanning for engine fault codes by dealer mechanics who know naught of the potential crash data contained therein. Be Very Very Careful with these cars! The hardware to access these should be coming mid-2007-ish.More information about how to best treat the modules to preserve data will come then, too.
- Many PCMs which are not covered by the upcoming update potentially have pre-crash data stored in them (mostly with up to 20 seconds pre-crash and 5 seconds of post-crash data recorded at 5 Hz), but it is in a circular buffer which can re-write itself afterwards when the key is turned on the next time. For this reason, you may want to consider sacrificing the diagnostic codes in the PCM and unplug it at the scene to preserve the circular buffer. As more data comes available on this, I’ll add more, but for now, it looks like you can choose one or the other, but not both. SIGH.
- The traditional warning has been that one should do a Ford service-type check of any airbag faults/codes BEFORE removing it, and then download it on-vehicle. An off-vehicle download MAY result in up to 10 SRS DTCs overwriting any potentially
present at the time of the event. This may not remain true with the aforementioned upcoming PCM changes. Stay tuned.
- Ford RCMs retain one event at a time. A Non-Deployment (ND) event is volatile and is overwritten by another ND which has a delta-V greater than 4 mph.
- RCMs don't keep track of ignition cycles and don't erase based on the GM criteria. The PCMs can “lock” their circular buffer if they get a signal that the bags deployed, though the logic for how long it stays locked is not yet public information.
- There may be RCM or PCM data available regarding the "time since started" up to about a minute or so and/or "lifetime seconds awake" depending on the vehicle.
- The signal that the airbags fired doesn’t always get through to the PCM when it should, so it’s possible to have unlocked pre-crash data in a PCM even though the bags deployed.
- RCM data is designed to be locked after an event - air bag deployment as well as pretensioner deployment/firing; HOWEVER the code that locks the data CAN be cleared at a dealership if they have the right software, which then makes the data no different than an "unlocked" non-deployment, and it can be overwritten. Reportedly, a dealership can't erase the data itself but can remove the lock code so it can be overwritten up to 5 times...the Ford repair guidelines still say the RCM should be replaced after an air bag DEPLOYMENT for reliability reasons. Until that code is removed, the data is secure. The CDR System will NOT remove the code to alter the data.
- Did I mention that you should READ THE DATA LIMITATIONS?
|
VALIDATION
The literature generally concludes that SDMs and RCM do a pretty good job of recording the data they are provided, and that data generally does a pretty good job of describing the vehicle's operational environment during a recorded event.
In addition to unpublished work by the OEMs to verify their systems' accuracy, Rusty Haight at CSI has conducted independent testing to compare the SDM recorded data to an IST accelerometer installed in the test car. The results showed good correlation between the two: the GM-CDR reported delta-vees for deployment-level events were within approximately 0.3 miles per hour of those generated by integrating the IST accelerometer data, with crash-durations matching within a few thousandths of a second. For low-speed events (particularly non-deployment events) the SDMs may miss the first few miles per hour change in velocity before the system wakes up, and the algorithm is enabled. This means you gotta be careful in applying this data to this type of event. Rusty's paper with this material was titled "Automobile Event Data Recorder (EDR) Technology - Evolution, Data, and Reliability" and was presented at ITAI in Nov. 2001. It used to be online, but it appears to have been taken down as of July 2010.
The Wisconsin State Police have conducted some testing which showed SDM-recorded pre-crash speeds were right in line with the radar-measured values. This testing was reported in "Crash Data Retrieval System Validation Testing, Wisconsin State Patrol Academy, Nov.2, 2001, Tpr. Tim Austin", a copy of which can be downloaded by clicking here. Additionally, there have been a number of papers published at ESV conferences and through SAE which address the question of accuracy in various circumstances, including the following:
- SAE 2002-01-0679, The Accuracy and Sensitivity of Event Data Recorders in Low-Speed Collisions"
- SAE 2002-01-0549, Possible Errors Occurring During Accident Reconstruction Based on Car "Black Box" Records
- SAE 2003-01-0889, The Accuracy of Pre-Crash Speed Captured by Event Data Recorders
- 2004-01-1195 The efficacy of event data recorders in pedestrian-related accidents
- 2004-01-1194 Crash severity: A comparison of event data recorder measurements with accident reconstruction estimates
- 2005-01-1190 Accuracy and Sensitivity of '03 and '04 GM EDRs in Low-Speed Barrier and Vehicle Collisions
- 2006-01-1397 Timing of Pre-Crash Data Recorded in GM SDMs
- Estimating Crash Severity: Can EDRs Replace Crash Reconstruction?, H.C.Gabler, C.Hampton, and T.A.Roston, 18th ESV Conference, 2003, paper #490
- Evaluation of Event Data Recorders in Full System Crash Tests, Neihoff, Gabler, Brophy, Chidester, Hinch, Ragland, 19th ESV Conference, 2005, Paper 05-0271
|
CDR LINKS & OTHER REFERENCES
Loads of EDR-related links from NHTSA
Rick Ruth is "Mr. Ford EDR" and has some info on his web site.
EDR Durability PDF file of slides from Wade Bartlett's 2006 CDR-User's conference presentation on drowned & abused SDMs. NOTE: This is a 3Meg file! (Added 02/07/07)
The ARC Network EDR Page
A 1999 NHTSA paper by Chidester on EDRs
Automotive Black Box Data Recovery Systems By Don Gilman, Vetronix (also available at Taro-Gilman)
A bar-journal article by David King, one of the lead MEA Engineering (formerly MacInnis Eng.) guys. They do a lot of EDR validation stuff.
A 2003 Car and Driver article on EDRs in cars. (300kb)
A Washington Post article from 09/09/2003 titled "The Snoop in Your Coupe" , but requires a subscription to access on their website.
Detroit News article 11/20/2003 titled (I think) "Auto black boxes defended" is no longer online
The Manchester Union Leader published an article on October 10, 2004, featuring your's truly. You have to pay to see it online, now, though - search their archive for "black boxes", pay $2.50.
"Crash data retrieval kit recovers reconstruction data from G.M. black boxes", John Goebelbeck, AIQ, Winter 2000
Several papers have dealt with this issue in recent years:
|
CDR TRAINING
|
Starting in late 2005, the CSI-certification process has shifted to being a two-layered affair: The first layer is "Technician", which is primarily concerned with getting and preserving the data using CDR equipment. The second layer is "Analyst" which is primarily concerned with interpreting the data so obtained. Training classes have occasionally been available from Vetronix themselves, but the only training listed on their website is from CSI. Additionally, IPTM offers a class titled "Methodology and Techniques of Crash Data Retrieval" quite frequently around the country.
|
CASE STUDIES
SUV Tangles with Camaro
|
The system was used to verify the speed of a Camaro involved in an accident in which an SUV had crossed the center line sideways, and was struck on the right side by an oncoming Camaro. Any doubt about the Camaro's speed was erased when the Vetronix system showed that it had indeed been travelling at about 30 miles per hour during the five seconds prior to the crash, and that the driver had initiated braking one to two seconds prior to impact. |
 |
|
 Wade downloading the Camaro.
|
|
Car Attacks Grocery Store
| In a case of "sudden unexplained acceleration," resulting in a car striking the side of a grocery store with no airbag deployment, the Vetronix system showed that contrary to the driver's description of being stopped prior to the engine's roaring-to-life and his ineffectual brake application, he was rolling at about 4 miles per hour with no braking, and pressed the throttle wide-open one to two seconds prior to impact. (see my essay on Unexplained Accelerations at suddenaccel.html) |
 |
|
Rental Car Blues
| Whenever I rent a GM car, I scan the module to see what I find. Usually it's something at low or no speed that looks like a parking-lot incident. A Chevy Impala I rented during a recent trip, however, had a near-deployment at 35mph, with the operator's foot coming off the brake pedal for a couple seconds, but getting back on the brake with 2% throttle application for a couple seconds prior to algorithm enable. Positing potential scenarios for this set of data has been amusing. |  |
|
The Suicidal Deer and the Sunfire
| Brad Muir sent me this file which resulted from a 1996 Pontiac Sunfire striking a 200-pound deer. The buck's rear legs were right at the center-mounted license plate, he wrapped over the hood, glanced off the driver's side of the car taking the mirror with him. Both airbags deployed. Peak change in velocity was 2.63mph, but it showed 1.54 at the first data-point. Here's the car and system status at deployment: |  |
| System Status at Deployment |
| SIR Warning lamp status | OFF |
| Driver's Belt Switch Circuit Status | BUCKLED |
| Passenger Front Airbag Suppression Switch Circuit Status | ON |
| Ignition Cycles At Deployment | 10170 |
| Ignition Cycles At Investigation | 10173 |
| Time from Algorithm Enable to Deployment Command | 13.75 msec |
| . |
|
|
NHTSA RECONSTRUCTIONS WITH EDRs
|
Some US tax-dollars go to pay for investigations of severe roadway accidents and the results are public documents. Some of them are pretty neat. Here's the NHTSA Special Crash Investigation (SCI) Search page. You can tailor a search for vehicles which may have downloadable EDRs, and then hunt through the reports for ones where they report the results of such a download. Some of the EDR-equipped vehicles involved in SCI-cases include there:
Additionally, there have been a number of EDRs downloaded after NHTSA crash testing. the list of those tests can be found at
http://www-nrd.nhtsa.dot.gov/database/aspx/eventdata.aspx. The tests listed there include barrier tests, offset frontal, side impact, and car-to-car developmental tests. You can hunt down specific reports using this NHTSA search engine. The online reports for these tests generally include the entire text (usually in PDF format) of the report which is hundreds of pages long, photos before and after, and one or more videos taken at full speed or sometimes high-speed (which plays back like super slow-motion).
|
WARNINGS/INFO FOR CDR-SYSTEM USERS
December 07: The Ver 2.9 shipped in April. In June or July there was a patch released to fix some "features" on that early disk, and a couple months ago I got a 2.91 disk in the mail. Some Canadian users have gotten their 2.91 disks only quite recently, there having been some hangup in the cross-border shipment. We're expecting 3.0 soon. I hope.
14March07: Vetronix has advised a fellow user that Ver. 2.9 will be shipping "in the next couple of weeks." Also, the new Belkin USB-Serial converter, which is like a little pigtail worked fine with my CDR with XP. I'm not inclined to try Vista yet!
14Feb07: A couple months ago, Siemens released their final VERONICA report on EDRs in Europe. Not only does it have some familiar names on the list of players (Otte and Steffan jump out immediately), it looks like a great summary of what's out there, and where they think it should go. Sort of parallels some of the NHTSA stuff on this issue leading up to the recent Rule563. Siemens Final Report.
08Feb07: Version 2.9 has been released from engineering. Discs are in the process of being burned now and should be shipping between the 15th and 28th of this month.
The Feb 2007 CDR-User's Conference in Houston Texas was time and money well spent. Here's some tidbits:
-Though their website doesn't reflect it just yet (07FEB07), Vetronix is moving to an annual software subscription model for their CDR system. Buy one year ($395) or buy 3 years at a whack for $1185. There might be a discount for conference attendees. Buying 3 years at once locks in the rate, but doesn't get any other financial benefit. This only covers SOFTWARE.
-There WILL be hardware introductions soon (chief among them a new High-Speed CAN box to talk to the Ford PCMs, but also various cables which may also be purchased separately as needed.) The upcoming High-speed CAN box is DIFFERENT from the recently introduced blue Low-speed CAN box. _sigh_
-Version 2.9, which is a DLC-only software upgrade which gets access to the RCM in a bunch more Fords and more 06-07 GMs was supposed to ship on Monday 05Feb...Haven't heard if it happened, though.
-PCM access with more good stuff will be a hardware upgrade, and should be coming SOON (maybe June-ish? Maybe a little earlier if things go well?), but will not be covered by the software subscription. I'll talk more about PCMs after we get access to them in our sweaty little hands.
-About the same time, a software update should arrive with coverage for one family of Chrysler vehicles (but not Mercedes Benz). More Chrysler coverage will follow.
-The 49CFR563 (Part 563 Rule) on EDRs looks to make a lot of headaches for the OEMs, and though standardization is a nice idea, there may be unintended consequences. In part, the rule mandates that IF a vehicle has any recording, it must meet an extensive list of parameters. One unintended side-effect of the rule is that if an OEM has an EDR (maybe even a pretty good EDR) that doesn't meet all the requirements on 9/1/10, they may just disable the units they make after that (maybe by popping the EEPROM out of modules before shipping them?) to avoid breaking the law. We're from the government. We're here to help.
-There's a Ver2.81 patch to fix Silverado issues primarily (maybe only?)...It's on the conference CD, it's also reportedly going to be posted at the Vetronix website. If I've understood correctly, the software didn't recognize some supported trucks as actually being supported.
-CDR-Tool Kit prices will go up with Ver2.9. A limited number of 2.8Kits are on the shelf at the old price, and when they're gone, they're gone.
While I expect NHTSA to use the specter of an OBDii-type mandate to help goad the OEMs into incorporating Rule 563-compliant EDR functionality in their cars, I don't expect plug-standardization (ala the ubiquitous OBDii Diagnostic Link Connector) any time soon. Or ever. We're going to have dozens of little cables in the box. Glad I bought a full-sized van for carrying stuff in!
Nov 2006: Vetronix has been bought by Bosch, so some things will probably be changing in the future. Exactly _what_ is not yet clear.
Mar 2006 Updates and commentary
-The conference was a good affair. Yours truly gave a presentation on how modules usually tolerate being dunked, but sometimes don't. We did NOT get a new OEM this time around, though we're told it's coming soon.
-Ver2.8 has updates for GM and Ford, and comes with two new cables and a new interface box to handle CAN-bus cars.
-The Ver 2.8 software will work with the old interface box, but that box can't talk to CAN-bus cars.
-If you bought the Ver2.7 back in 2005 - your new cable should be there soon, mine has arrived.
-If you bought the Ford add-on, you may have a cable (3274) that needs re-working by Vetronix.
-The Chevy Aveo is really a DaeWoo Kalos with unreadable Siemens modules, so it will not be supported by CDR.
-A private analysis of 68 CDR files showed approximately 6±2.5 key cycles per day...I thought that was interesting
-The Help file has an error: for the 2006 Cadillac SRX, use cable 2003293
-Even though the 2.7/2.8 software may ACCEPT a Jaguar VIN, it can't decode Jags. There are no plans for coverage, and spoofing will not work. Thanks to Jeff Weger 02/21/2006
-When downloading via DLC in vehicles which use GMLAN (module cables 02003320, 02003321) such as the '05 Cobalt, leave the key out of the ignition. Per Jeff Weger 02/13/2006
-Ford Sport-Trac will be covered in the next software release...date unknown. Per Jeff Weger 02/03/2006
-Lately, I've been using an old Libretto 50CT for downloads: so small it fits inside the CDR case!
-Jim Harris keeps a list of CDR-related legal cases, as well as state-by-state CDR laws: http://harristechnical.com/cdr.htm
Aug 2005: The next big thing: CDR-User's Conference 2006 will be a 2-day affair, with a train-the-trainer event, as well as introducing some new manufacturer coverage / updates
December 2004: Version 2.7 is available, with more Ford coverage, and several 2005 GMs (though there are some odd ommisions due to platform sharing with other OEMs). The CD ships now, cables (for non-DLC downloads) will be coming in the next few months. If you've got Ver2.0 installed, you can get the update here: LINK LONG SINCE DEAD
June 2004: Version 2.4 is available now...FWIW, The word on the street is that more FORD coverage is coming, but nobody's talking until it's final and it's not final yet.
Dec 2003: Vetronix has started shipping a redesigned CDR interface module with new CDR kits. The new CDR module has a new blue enclosure with an updated circuit card designed to output more power, eliminating the need for a CDR power booster adapter. Therefore, you can plug the 02003002 cable directly into the new module. All other functionality is the same as the original design (black module). Ahhhh the joy of upgrades!
Oct 2003: The Preferred Customer update kit is now shipping.
June 2003: Software Version 2.1 is available at the yahoogroup, some of the cables are still waiting to ship.
YAHOOGROUPS E-List An email-group has been formed to share info and download the latest software versions at http://groups.yahoo.com/group/CDRtool/...membership requires attendance at a Vetroinx-approved class.
Harris Technical has created a list of CDR users to help people find someone close to home, click their name to go to that page.
2003 GM model years are largely covered by ver 1.68 - currently only accessible via the DLC (cables shipping with Ver 2.0).
March 2003: Vetronix has announced that "CDR Preferred Customer Program members who have paid the $395 for a year's subscription to the update service will soon be receiving the v2.0 CD and three cables (1 Ford & 2 GM). These cables support select 2001 - 2003 Ford vehicles and select 2002 - 2003 GM vehicles. In June, members will receive (at no charge) two more cables and an adapter, all GM. These cables support select 1994 - 1998 GM vehicles (02003002 vehicles). Please review the Supported GM Vehicles' list in the Help Files for a complete list of vehicles. This information is also available in the CDR Preferred Customer Program mailing (page 3) that was sent out to all customers and posted on the Yahoo web site. Please review this document for more information." Sent by: James Kerr, Vetronix Corp.
Coverage issues: It has been reported that the '97 W-bodies are NOT covered, and that the '00 GMC Savanna and Chevrolet Express use the new-and-improved SDM-G cable. (My thanks to Brad Muir for this tip.)....
More coverage Mid-december, Jeff Weger advised me that the 2003 Yukon is supported in version 1.68 via DLC only. Module cables should be available soon.
All Software Versions: Overwriting Near-Deployment File During Recovery: Brad Muir also reported that in one vehicle he examined, the deployment occurred at "n" key-cycles, the near-deployment occurred at "n+1" cycles, and the investigation occurred at "n+2" cycles. This appears to be a case of the tow-truck operator turning on the key (perhaps to unlock the front wheels), and bumping the system hard enough to overwrite whatever was in the near-deployment file before.
All Software Versions: Overwriting Near-Deployment File In the Office: If you remove the airbag module from the car, be careful to not bang it while it is connected to the CDR system (and therefore powered-up)...this could register as a near-deployment, overwriting the previous near-deployment file, which may be the one you were interested in. (from Vetronix in a special bulletin)
Ver 1.1: Overwriting whole files in your computer: If you are an icon-user, be careful that you do not inadvertently hit the SAVE icon instead of the OPEN icon in the main screen....Given the right circumstances, you could overwrite a previously opened filename with the currently-viewed file, thus wiping all the older file's real data from your computer. (My thanks to Brad Muir for this tip.)
Ver 1.1: Laptop problems Software V1.1 does not allow the user to change inputs from COM1 to LPT1, which means that laptops which have but no 9-pin serial port(COM1) won't work unless you buy a serial-to-USB converter (for about $70). CDR software version 1.2 is reportedly now in Beta testing, and may address this issue. (My thanks to Gary Johnsonfor this tip! - added 08/01/2001)
Ver 1.1 is not compatible with Windows ME but Ver1.2 (due out Feb 01) should be. (My thanks to Gary Johnsonfor this tip! - added 08/01/2001)
Software Ver 1.3 has been shipped, some newer updates are available through the yahoogroups e-list mentioned above. These worked fine with my Windows-ME, through a Belkin USB-Serial connector.
Windows XP apparently doesn't play nice with the Belkin serial adapter, particularly older ones....I bought a VAIO FXA-49 with a built in serial port running XP and it has worked flawlessly. April 2002
|
Copyright 2000-2010 by Wade Bartlett
Mechanical Forensics Engineering Services, LLC.
This page created on 08-SEP-2000, and last modified 29-JULY-2010
|