Maintenance, Emissions, And Privacy: The OBD Story

The 90s were a pivotal time in world history, and 1996 was no different. You might have spent the year glued to the TV playing Super Mario 64, or perhaps you were busy campaigning for Bill Clinton or Bob Dole, or maybe you were so depressed that Princess Diana and Prince Charles divorced that you spent the whole year locked in your room, a prisoner of your own existential nihilism. Whatever you did, though, it’s likely that one major event passed you by without a thought: The standardization of on-board vehicle diagnostics (in the US), otherwise known as OBD-II.

In the 1970s, vehicles (in some western countries, at least) were subject to ever-increasing restrictions on emissions. Most companies began switching from carburetors to efficient fuel injection systems, but even that wouldn’t be enough for the new standards. Cars began to carry rudimentary computer systems to manage and control the influx of valves, meters, and sensors that became the new norm. And, as one would guess, every car company had their own standard for managing and monitoring these computer systems. Eventually they would settle on the OBD system that we have today.

Sea Change

First, though, we should go back in time to see what led to this sea change in the automotive industry. In the beginning, the only way to diagnose a problem with a car was to investigate unusual noises, poor performance issues, or other problems directly. This can be especially troublesome on electrical problems because of their invisible nature. It was only a matter of time before someone hacked together a way to make the troubleshooting process easier. That’s partially why the first mass-produced vehicle diagnostics systems had almost nothing to do with emissions and focused mostly on helping mechanics identify common problems.

img_1151_thumbnailThe first of these systems came online in the late 60s on some Volkswagens. (Here’s on in evidence in the author’s Beetle.) The proprietary test port that was installed in these cars gave the mechanics easy access to the electrical system via a “small” computer that could test a number of different aspects of the car. From there they could tell if lights were burned out, some information on the ignition system, and (for fuel injected cars) some information on the fuel delivery system. Even though a 1970s Volkswagen is pretty late in the game as far as old Volkswagens go (the first model year was 1938) it was still early for the industry and gave VW mechanics a leg up on troubleshooting.

The next car company to take a stab at a vehicle diagnostics system was Nissan on their 280Z sports car. Unlike Volkswagen, this system was intended to help manage the increasing demands of the fuel injection system in the post-EPA world. Not to mention the fact that the 280Z was a pioneering performance car, and any gains that could be made by using a computer to fine-tune the engine were welcomed.

After the success of Volkswagen and Nissan, other car companies saw the perks of having a diagnostics system that was easy for mechanics to plug into to identify problems with vehicles. General Motors created their own proprietary system in the 80s that helped factory workers identify problems with cars as they rolled off the assembly line. Although it’s probably not a good sign for your manufacturing process that you need to specially design diagnostic tools for brand new cars (after all, when was the last time you saw a Cadillac Cimarron on the road?), the GM diagnostic system did find some utility with mechanics for a time.


(Slide from this talk on OBD-II)
(Slide from this talk on OBD-II)

At this point there are many different car companies having equally different standards for interacting with the on-board diagnostic systems. This is where the US federal government stepped in with a mandate that all cars have a standard OBD system by a certain date. As a result, diagnostic systems really start to focus almost all of their attention to the emissions systems. The first standard was known as OBD-I and took effect gradually in the late 80s and early 90s. This system was primitive by comparison to its successor, OBD-II, which is the current standard for all cars sold in the United States since 1996. Most other first world countries adopted this (or similar) standards around this time as well.

OBDThe current OBD-II standard specifies the protocols used to monitor many of the sensors in the car and can alert the driver if any of them aren’t operating properly. Most cars built under this standard have an increased number of oxygen sensors, control over mass air flow sensors and the fuel injectors, and many other aspects of the car. All of this means that there are more efficient engines on the road that all communicate in a way that makes it easy to tell why the check engine light on your dashboard came on. It also makes it easy for a mechanic to test the emissions of the car if you live in a place the requires your vechile to pass an emissions test. All the mechanic has to do for an OBD-II compliant car is plug into the OBD port and view the output on another monitor, rather than fuss with attaching sensors to the exhaust and running the car through a battery of tests.

On the other hand, there’s a lot to be desired from a system that was designed years before anyone had heard of Super Smash Bros. There’s some talk about a new OBD-III standard that would allow cars to take advantage of the growing wireless network, essentially adding “all new cars” to the list of Internet of Things devices. It would then be possible for a car to report itself to regulatory agencies, and for the car to be shut down remotely if it violates any emissions laws. While this could be argued to be a win for the environment, there are certainly concerns about government overreach and Big Brother that will need to be addressed before anything substantial is brought to market. It would have probably made it harder for Volkswagen to have cheated their emissions test results, for example, but the benefits might not outweigh the unintended consequences of something like this. Whatever the new OBD standard will include, though, it will be a welcome improvement for the aging system we have in OBD-II.

113 thoughts on “Maintenance, Emissions, And Privacy: The OBD Story

  1. “It would then be possible for a car to report itself to regulatory agencies, and for the car to be shut down remotely if it violates any emissions laws. While this could be argued to be a win for the environment, there are certainly concerns about government overreach and Big Brother that will need to be addressed before anything substantial is brought to market.”

    Ya think?

    1. Fort McMurray May 3rd 2016 “Thousands die, burned alive, unable to flee fire, as their vehicles get shut off for emission violations due to amount of woodsmoke present in intake air….”

      expect headlines like that.

      1. Same thing could happen down south but slightly different. “Thousands perish in hurricane as vehicles shut down after topping up with lawn equipment gas and stale gas trying to make it out of path.”

        1. Periodic car “IoT car payment” verification messages could not be received due to local tower outages that proceeded the storm conditions so thousands of “cloud based” cars were erroneously disabled due to not being able to receive “car payment or even emissions verification” messages in the days leading up to the hurricane’s arrival, preventing evacuations for those individuals. Additionally, due to the traffic, some disabled vehicles were actually disabled while they were in the path of traffic, completely blocking off evacuations of everybody else.

          The legal liability for a company to do this would be enormous. Insurance companies would exclude coverage too. Your one car gets vandalized, it’s covered. Hundreds of cars get vandalized? Riot. Not covered.

          1. @iowncalculu
            That’s just one of many reasons why you should never do business with those shady operations.
            Find an honest dealer or buy from an individual and have someone knowledgeable look over the car for you if you do not know how to spot trouble yourself.

          2. Though I confess to buying a car from one of them myself… 11 years ago… they had spliced together wrecks on the front… looked round the back, find the best car on the lot, one they took as a trade, and got it for $500… drove it 9 years and 180,000 miles. :-D

      1. Which will be yet another arrow in the quiver of those who intend to eliminate private vehicle ownership. And please don’t kid yourself – this is a growing movement of not only those with environmental concerns, but a more dangerous and better leveraged group that are concerned about the rising costs of maintaining the infrastructure cars depend on. One of the many benefits of less cars on the road is less wear and tear on those same roads.

        1. Well that’s all divorced from reality. If 100% of the money raised from private motorists actually ended up spent on infrastructure we wouldn’t have any issues. Whether from gas taxes, license fees, tolls etc… it all ends up in general tax revenues and lucky if 20% gets where it was originally earmarked for…..

          On top of that, studies have shown that a commercial truck, and one would also suspect a large bus, does 10,000x as much wear/damage to a road surface per mile as a private car does, with very very little attempt to recover actual cost vs wear of those vehicles…. so remove the private motorist from the equation and those vehicles have to subsidise the entire infrastructure, and we’ll have a bus ticket across town looking like the price of an air ticket to your nearest sunshine destination today.

          1. I always hated how Ohio charges more for title and license for a truck. My little s10 weighed less than and got better gas mileage than our crappy mninvan (and therefor less damage to roads or environment) yet I had to pay more because the s10 was a truck and of course the perception that I hated the earth because I had the truck.

          2. Well I’ll challenge your contention that heavy vehicles are responsible for the bulk of road wear. While they may contribute more per unit-mile I doubt that they come close to the 10,000x value you are suggesting. At any rate that still will not stop considerable pressure from mounting on private vehicle ownership because reducing the number of cars on the road will extend the useful life of the infrastructure and reduce maintenance costs and given the state of much of said infrastructure now that will become a major potential expense and thus a major political issue.

          3. @Chad – Then it should be easy for you to post a reference. Note too that the question revolves around total wear on the system and not per unit-mile of either type of vehicle. At any rate the term ‘truck’ is rather broad and it is clear that a step van is not in the same class as an interstate hauler or dump truck full of gravel in this regard.

          4. It isn’t that trucks just inherently wear roads more, it is the way the roads are made that cause this. Our roads and bridges are intentionally built to wear out quickly because that makes jobs and justifies taxes that mostly get spent on things that have nothing to do with roads. It isn’t just trucks either…in the south it is common to design roads to wear much faster on the hottest days of the year in spite of the fact that making roads that would be unaffected by the heat would be easy. In the north it is common practice to design roads that disintegrate every time it freezes and thaws. I can say for certain that here in Wisconsin, the difference between a truck and a car means nothing…the roads come apart because they design them to hold water that freezes and jacks them apart long before they can wear. If roads were actually built for the conditions and vehicles that the engineers know they will face then they would last a lot longer…but then unemployment in the construction sector would skyrocket, not to mention lay-offs at the companies that supply the equipment and materials required…those laid-off workers would stop going out to eat, etc, etc, etc. Sure, it would make more sense to simply spend that same money on improving the roads…but selling improvements is a lot harder than selling repairs and creating special taxes to fund new roads no one needs is political suicide while making new taxes to fund repairs on roads that should have already been paid for is easy.

          1. Traditionally, vehicles are considered to wear roads in proportion to the 4th power of axle weight

            For a more detailed discussion, see

            This is probably a surprise to most car drivers. Requirements for bicycle registration are often mooted around the world by the cager majority. I’d be happy to pay bicycle rego at a cost proportional to road wear :)

        1. That’s it – the writing is on the wall for anyone with the eyes to see it. The costs of owner-driven cars will rise to the point where few can afford to operate them. Just look at what they have done to motocycles in many jurisdictions. What once was inexpensive alternative transport has become a costly hobby. The same will happen with cars.

          1. @Wretch

            Here in canada almost any sport bike over 500cc is inordinately expensive to insure.

            I have a perfect driving/riding record, no accidents or tickets on my history. 500cc bike was approx 1200/yr cdn, paid over the summer months, so you dont remove it during winter. If i went to a 600 it was 4000/yr, and a 1000cc was over 10000$/yr.

            My evo, which was faster than my ninja 500, cost me less than the bike for fully comprehensive insurance, the bike insurance above was liability only (aka required by law).

            Thats what they’ve done with motorcycling, and what will happen to cars that you drive in due time.

          2. @DV82XL
            People need to start fighting this bulls**t because that is exactly what it is.
            Even a Harley probably does less lifetime damage to the environment than a Tesla S.
            You gotta be kidding me that is ridiculous in the US you can insure an airplane for less than that.

          3. Thing is is that it is not that easy to fight this sort of thing up here in Canada when the Provinces are both the vehicle and broad health insurers and, as a consequence of the climate, motorcycling is a seasonal activity. Bikers just don’t represent a meaningful demographic in terms of their political influence.

      2. Same here and if I had to drive one because I won it or got it cheap I’d rip the stock computer out introduce it to a couple of 7.62 rounds and install an Accel DFI or mega squirt in it’s place.

      3. “The current OBD-II standard specifies the protocols used to monitor many of the sensors in the car and can alert the driver if any of them aren’t operating properly. Most cars built under this standard have an increased number of oxygen sensors, control over mass air flow sensors and the fuel injectors, and many other aspects of the car. All of this means that there are more efficient engines on the road that all communicate in a way that makes it easy to tell why the check engine light on your dashboard came on. ”

        That’s in a perfect world. In the real world, Manufacturers still implement non-standard codes and protocols through the OBD-II port, making accessing to advanced sensors impossible without expensive proprietary equipment. For example, VagComm for WVs. Manufacturers even do it worse, like Peugeot does by cryptographically pairing the injectors with the ECU, making any garage other than Peugeot unable to change the injectors.

    2. Sorry, whats to stop Volkswagen from cheating again? Again, this reporting (and lying) is (and was) being done by the firmware. Really, this will probably just usher in a whole new era of software fakery.

      1. Indeed, it is (too much) hyperbole to suggest that car would shut themselves down because of unmet emissions standards. The car is not made by the driver, so it would be unfair to punish her for the shortcomings of the manufacturer.

    3. one word: ransomware

      if the computer in a car can prohibit a vehicle from operating then a malicious entity cna reprogram that computer to shut down whenever they want… given how bad IOT security is at the given moment, it would be reckless to even consider the next OBD standard to include wireless.

  2. ” Although it’s probably not a good sign for your manufacturing process that you need to specially design diagnostic tools for brand new cars (after all, when was the last time you saw a Cadillac Cimarron on the road?), the GM diagnostic system did find some utility with mechanics for a time.”

    Sort of a JTAG for cars.

      1. Some of the OBD II cars I have encountered have had a way to make the CEL light flash the code. On my x-terra it’s like turn the key on, press the accelerator to the floor 5 times in 5 seconds, wait 7 seconds, press the accelerator to the floor for 5 seconds. And the light starts flashing what codes the computer has stored. I’ve done similar things to other cars, usually some type of “code” entered with pedals and the ignition to get the codes or change things.

        I have also caused my friends car to un-pair with his key fob by inserting the the key into the ignition and pulling it out again so many times in so many seconds. I was doing it to bleed of some nervous energy and one time the horn beeped and lights flashed, then later we found out his key fob no longer worked.

  3. There are positive and negative things with standards.

    What a standard needs to support to be useful is always a good question, and what the implications of that support in regards to ease of use and safety is always something that one can discuss for ages.

    That car manufactures will start trying to use IoT would not be a surprise.
    Though, most people don’t know the difference between safety and encryption, and most companies falls into that group as well…

      1. You are right that mobile internet services are already used in cars isn’t new.
        And yes, there are stuff like OnStar. (That someone should look into more and see how they address safety. I’m not from the USA so I don’t really care. And my car is far from that new.)

        But it is another thing all together when something goes from “application specific” design, to having a web browser that you can watch youtube on, or even check into a hotel through any of the many web booking services available. Simply stated, more support, is normally adding more risks and potential flaws. Even if the original system might have had a fair few to start with.

        In the end, I’m no car manufacturer.

      2. In the OnStar systems that used an analog cellphone module, the OnStar box has a GPS module that can simply be unplugged after opening the box. Give it power and apply a serial signal level adapter to make it RS232 compliant and you have a GPS device that works with anything which reads standard NMEA code.

        The last couple years of the analog OnStar could be upgraded to digital. Previous years are SOL. No more service for them since 2009.

  4. From a hackers perspective, OBD-II is great. With the right parts (most of the time passives only – though newer cars may necessitate a micro or two) you can make the ECU think anything you want about the state of the engine or emissions, and pass this onto any interfaced computer. In short, OBD-II plus knowledge of rudimentary circuits is enough to never fail an emissions test even if there is liquid gasoline pouring from your exhaust pipe. It helps that most people who run the emissions tests 1-Don’t Care, and 2-Will believe anything the computer tells them.

    1. And this is only needed if the ECU in question is one where the aftermarket community hasn’t figured out how to reprogram it directly. There’s a good chance that any vehicle with a strong aftermarket has at least one tuning solution with the ability to directly disable the computer from setting specified error codes.

    2. It actually is. I use my reader for gauges that didn’t come with the vehicle. Plus you combine OBD-II with other information like GPS, maps, and other information and it makes it easier to sus out those intermittent problems, and improve performance.

      1. You can also get a Bluetooth connected OBD-II plug that talks to your smartphone. Its useful for all sorts of things. Plus with a window mount your phone it becomes more or less a HUD – or as you said, provides gauges that are otherwise unavailable.

          1. Yup. $10 for the bluetooth adapter on amazon, then Torque ODB on your phone/tablet. I permanently mounted my Sony Z3 Tablet Compact in my 2000 Sierra’s dash, replacing the stock stereo and such. It’s waterproof, has a radio, can have a full screen dash showing all the engine data.

            It’s extremely handy not just to get any engine problem codes, but actually be able to see real time engine information. Coolant, oil, and transmission temperatures, air fuel mixtures, O2 sensor responses, any misfire problems per cylinder, etc.

            Not to mention adding fuel consumption display, gas mileage, etc – you get more features than you can in a 2017 vehicle. The list of data you can add gauges for is incredibly long.

            And, being an android device, you’ve got full infotainment options, gps, a dashcam, voice control, face unlock…

          2. Does the scantools adapter still use bluetooth? That’s pretty much mandatory for me, to connect with my tablet, but I’d much prefer an adapter that shuts down when my truck is off. It doesn’t draw much, and is fine for a weekend, but I’ve had my battery die after 3+ days without use, or less if it’s really cold(relevant in our -30C canadian winters)

      1. Military vehicles need to have be able to have radio silence.
        A vehicle that’s always transmitting would quickly be discovered by the enemy and find it’s self on the wrong end of a missile.

        1. Anti-radiation missiles are (very) expensive and only used on high-value targets (certainly not one APC) or SAMs (again, not an APC)…
          In a real scenario, ELINT would probably pinpoint their position and send a drone to investigate… but depending on whom the conflict is with, they would very likely just keep listening while keeping an eye on where the vehicle is, as that can yield far more tactical information.

          Especially with most recent communication hardware, the operator might have actual problems making sure he is truly keeping radio silence, since the brass want their “future soldier” systems that give you real-time battlefield data, including unit positions :P

          That being said, a BRDM or BRDM-2 (available in running condition in most ex-soviet satellites for reasonable money) is a probably a reasonable choice, as the only electronics in it will be the ignition system and headlights :D

  5. “OBDIII” would have no impact on preventing Dieselgate. If the core issue of Dieselgate was that VW was shipping every TDI with a missing Check Engine Light, then maybe. But the ECU doesn’t *know* with any certainty what the vehicle’s actual emissions are. It just knows that the sensor inputs it sees are either within allowable limits, or outside of those limits. The manufacturer sets those limits based on the vehicle and actual sensors in question. Automotive environments are also harsh enough on sensors that you can never treat an errant reading as gospel – you still need to diagnose the issue and see if the code is due to an actual mechanical problem, or if the sensor itself has failed.

    The only change would be that VW could update the programming on the cars remotely, instead of requiring them to come in for a dealer service… but that would only be a benefit if VW actually had updated programming available yet.

    1. Correct the best the OBD can do in regards to emissions is to report the the vehicle’s systems are functional and the vehicle *should* be meeting commission standards. That is what the industry wanted, rather than testing what is actually exiting the tail pipe. Having said that in all likelihood the current system is probably reducing emissions at a lower cost for the vehicle owners.

      1. Especially as no vehicle has onboard CO or NOx sensors. They have O2 sensors to regulate the fuel/air mixture so that the catalytic converter _should_ make the best out of it, if all works as intended.

  6. Attention author, What are the chances of giving the readers an unbiased article on the history and future of OBD standards? Preferably something based around facts and not random internet forum echo chamber garbage. Fiction can be fun but facts and logic are way more useful to us readers. You look like a clown to those of us in the automotive field. Step up your game son. I look forward to reading a better version of the article.

    1. Unfortunately there are people who want data-connected (IoT) cars. I prefer a separation between GPS, ECU and the cellphone modem. The latter shall only be connected to a handsfree phone solution, if it is there at all.

  7. Ah the fear mongering… With out any connection to the internet how would big brother control a vehicle vi IoT? In the USA coomerce it almost totally dependent on privately own vehicles. Forcing vehicle owners to purchase a data plan they can’t afford is going to cut into the profits of many businesses across the board. I know many who wouldn’t have a phone at all if it weren’t for prepaid no contract mobile phone plans.. I don’t fully understand how that business model supports itself, so I can’t understand how similar data plan model for motor vehicles can support itself either. Just it it possible for regulatory agencies to do something doesn’t mean they will For Christ’s sake I live in Kansas, where there’s no mandatory vehicle safety inspections, much less vehicle vehicle emissions. would seem the Insurance industry doesn’t have the power to get mandatory safety inspections. in place, and the Federal government powerless to get emission inspections in place, I’m not going down the fear mongering road. Not that I’m saying it can’t never happen, just that it’s unlikely; because it would cut into the profit’s of those that have the most power in our government of the merchants,by the merchants, for the merchant.s here in the USA.

    1. They don’t need to do it via the internet. If a car has receive-only cellular capability a signal can be sent to it to shut it off. Granted, it makes more sense to have the car also able to transmit information, but plenty of intrusion can be done read-only. I remember England was looking at using on-board GPS to literally enforce speed limits. Set the speed limit to zero on a road and nobody can use that road, except those cars without the limiting devices or those granted an exception to the speed limit, such as emergency vehicles and politicians. No transmitter necessary.

      1. Even that would have problems as cellular towers can only serve so many connections at once.
        They don’t even need to be transmitting data just the handshakes are enough to use up available slots.

      2. Normal cellular technology requires a two-way transceiver. And if the car has no transmitter, I am already happy. It can not report to anybody if the such obnoxious devices are defeated.

    2. I don’t think it would be to hard given the equipment that is on nearly vehicle and the sensor network installed at nearly every intersection… Could use IR, (many vehicles have IR sensors for auto headlight setting), or perhaps through RF ( your vehicle doesn’t have a keyfob and listen for a code from it…?) etc. many ways to skin that cat with just software/firmware updates.

  8. Could never afford a new car. Government is taxing us to death.
    Next Plan. Do a major crime. Goto jail.
    Now I have free room and board. Free medical. Free food. And free schooling.
    Thank you everyone…

    Oh one last thing lots of s## that I dont want.

  9. 1996 was the year OBDii was required for cars, trucks were made without it until the late 1990’s. Ford switched from the V8 with OBDi to V10 with OBDii, skipping a model year.

  10. “It would then be possible for a car to report itself to regulatory agencies, and for the car to be shut down remotely if it violates any emissions laws”
    All I can say about that is take all of my nope as there is no way I’d want a feature like that on my vehicle.
    One it would be inconvenient as now you can still drive your car with a few bad sensors until you get a chance to replace them ,two it would be abused by authorities,and three it will be hacked.

  11. mostly OBD II is for end user simple checks and emissions cvn/run monitor tests. all the major players use dedicated test equipment still uses proprietary protocols and hardware for proper diagnosis, some of them even require a live connection to a central server at the OEM, so nothings really changed their, some OEMs just buy in a solution from one of the dedicated scantool makers, others devise their own and the protocols/info are kept a closely guarded secret, or an expensive one

    there is no proper standard with OBD that defines how accurate;.reliable most of the information is (mostly just pass/fail), or how much latency. very very few cars have real time OBD. heck my speedo is CAN bus and it can be upto a second off. some cars are better than others, but there is no guarantees because ODB doesn’t require anything to be realtime

    1. You are correct. But for the average mechanic, getting to the problem areas with a set code in the odb ecu, is more efficient as it takes a level of troubleshooting off the technician so they can focus on repair instead of diagnostics.

      1. Agreed, OBD2 is much better than the lack thereof, if for no reason other than emissions testing. I had a Volvo 240 that would run and drive perfectly fine, but would fail emissions due to what I think was a lean condition that I was never able to fully track down. It was a pain to diagnose because it always ran the same.

        Anything OBD2, I know full well whether or not I’m going to pass before I get to the testing station. If the light’s off and I haven’t had to pull the battery for any reason recently, I’m good. If the light’s on, then I typically have only one or two systems to run diagnostics on to either identify the problem sensor, or the problem mechanical component.

      2. only for very basic troubleshooting stuff, and a lot of depends on the implementation of the obd subsystem and what it checks, like a lean code might be a maf sensor, might be an air leak, might be all sorts of things. it can even point you in the entirely wrong direction.

        obd ii codes are extremely wide ranging, of course a lot of mechanics fix things by simply swapping out parts til it goes away.

        there are specific codes (that arent always implemented) that will tell you say open sensor 02 circuit 1 or such, so you know to check the sensor, harness and ecu.

        on top of that the history thats stored for when a fault occurs is a very small amount of data, if its there at all.

        so yes it can help simple troubleshooting and a lot of the basics for failures and not so much intermittens or ghosts( which is what a lot of obd codes are) , but you’re much better off with a real diagnostic tool.

        plus the obd ii may not even have the subsystems you need to diagnose.

        1. @charliex. That’s exactly the point. Horseshoes and hand grenades type of close. But I have been in the situation where the Key On Engine Off test completely failed (the flashing check engine light had a seizure of sorts) but was tracked down to a burnt O2 harness touching to the exhaust. I started unplugging sensors to purposefully set codes. only to find that once the O2 harness was unplugged the Key On Engine Off test went all the way through.

          I have also been sent in the wrong direction with an Idle Air Control valve code just to discover that a cracked hose near the evaporative emissions solenoid identified the problem and was setting the code. It made the IAC work overtime causing it set an voltage out of bounds code.

          Every sensor should atleast provide feedback to the ecu. obdi was not always implemented well with regards to that.

          1. yes that is the point very basic troubleshooting that may or may not be valuable. that is all it is, but there is no real standard, no guarantees beyond certain emission parts, that is the point i’m trying to make, you cannot be sure since there is no standard or requirement for accuracy for most of the obd ii subsystem and what it reports, or when

        2. Codes are of limited use (but better than nothing) but even a cheap Chinese elm327 odb2 module can relay live data just fine.

          Use Torque, put relevant sensors on-screen graphed, monitor them while driving.

          Being able to see live MAF and O2 sensor data graphed trivializes diagnosing issues with them, as an example.

          And that costs basically nothing – an Android device, even the free version of TorqueODB, and a $10 adapter.

          1. yes usually better than nothing, sometimes misleading, sometimes wrong, sometimes right, sometimes off base. there is no guarantee that the data is correct or realtime, you can say graph two items simultaneously but they may have come from different samples

            it’s knowing that its flawed and not to reply on it too heavily that is important.

            i can assure that graphing a MAF/O2 sensor doesn’t trivialise anything but trivial problems, especially via OBD II its all very dependent on the implementation and the problem, if it falls in the venn diagram of whats wrong, it’ll help.

            the dealer tools are much better, and those are the ones we should concentrate on, OBD II is the POST of automotive.

    2. One second latency is pretty bad even for OBDII. I don’t think that’s an issue inherent in CAN or OBD II, likely an issue with the design of the car’s network. What kind of car do you have?

      1. it has nothing to do with the network the data is being delivered on, it is to do with the priority given to the servicing of the OBD II request by the ECU, it is basically maskable, or able to be ignored by the code, or serviced when it feels like it, there is nothing in the spec that says, you must give me this answer in this time frame, this is one of the issues with OBD is that timing isn’t well defined, so there are a lot of tools that don’t work because they don’t want long enough for a response and so on.

        if the ECU is busy doing something else, it can hold off on the request, or not answer.. OBD II is really only for federal requirements on emissions tests and so on, so it doesn’t have to be guaranteed beyond that, or what diagnostics it can do. that is why OEMs don’t use it for actual diagnostics.

        so when we do ecu design work and we’re short of run time, we can say, i don’t care about replying to that message, its ultra low priority, ask again later, or here is some cached data.

  12. I like how the conversation is mostly “Big Brother”.
    One thing I would like to add about the OBDII issue is that during the transition period, many manufacturers requested extensions for light trucks and especially vehicles with a commercial or off-road usage. I have a 1996 Ford F350 and it has an OBD1 port. I have had the inspection place refuse to inspect it because it has the “wrong” connector or so their emissions tester computer claims. They finally pass without emissions testing after I show them the vehicles build sheet which has the emissions system certification for the OBD1 system utilized.

    1. I have a ’95 F350 with a Powerstroke. It was built as a cab and chassis fleet truck and it does have the OBD2 plug. It also has the OBD2 protocol. But a consumer truck may or may not have the OBD2 protocol even as late as ’97, Which were likely factory unregistered ’96’s. It’s unbelievable how much disinformation is out there, mostly due to armchair “professionals” and poorly written pieces like this, which is even bad for the “new” HAD. Has the author ever touched a OBD interface? No mention of the many cheap scan tools for hacking OBD, or the existence of CAN tells me this is another half-hour of research and writing to satisfy the time clock/ word counter at Supplyframe. BTW: Supplyframe will never, ever get a cent of cash from me, I openly spread their destruction of this once very good website at every opportunity. The contrived pieces about editorial quality only energize me in my mission.

      1. Awesome Paul. My friend has a 99 f650 that was purchased as a chassis cab for a rollback bed. If anyone knows about Ford part numbers you know that the first letter is the decade and the number that follows is the year. So E = 8 and a E0 would be 1980. His Cab was manufactured in 1980 and used to build a 1999 chassis cab. I often look for parts on these trucks as they are interchangeable with my Fords. And usually in good shape due to the later year model.

  13. In 1994 and 1995, there are several models of cars for which none of the consumer/prosumer aftermarket OBD1 scan tools work. On GMs, those models also cannot have two pins of the ALDL shorted to blink out codes with the check engine light. Only the factory and dealer test equipment can interface with them.

    Since the dealers have mostly gotten rid of their pre-OBD2 gear, you may find an independent shop that bought it secondhand.

    Some of the professional shop diagnostic equipment, like the red and black units Snap-On sold, had modules that would support those oddballs, but nothing you can buy new does.

    I had one of those cars, a 1995 Buick Century. For some crazy reason, GM decided to use a different setup on the 1994 and 1995 Buick Century. It’s not even compatible with all the other front drive A-body models those years. No way to get blink codes and neither Autel, Innova nor any of the other companies makes an OBD1 tool that works with them. There was no sane reason to do such wildcatting when all the manufacturers knew that a universal standard was coming in 1996.

    Some 1995 cars use the OBD2 connector but are NOT OBD2. As for being “universal” only the basics have to be. The manufacturers can define their own codes and do things like restrict the data rate without a proprietary device to unlock the full speed. That makes realtime data impossible on some vehicles, which means the data can’t be used for things like a fast responding speed display, nor can it catch short duration, intermittent problems.

    Do some digging and you’ll find that OBD2 tools break down into a few groups. Ford/Mazda, VAG (Volkswagen Audi Group), SEAT, BMW, Mercedes Benz, GM, MoPar/Mercedes etc. Since Ford has gotten out of involvement with Mazda, they’ve diverged from the common codes they both used. Jaguar was also in their OBD group before Ford sold off their interest in the company. I assume that since MoPar is now owned by FIAT that there’s been some OBD consolidation there, at least in North America.

    It’s not just the code numbers and definitions that can vary. There’s several different protocols, which a scan tool must be able to query then switch to in order to receive data.

    Thus OBD2 is a “standard” only in the sense that you can (should be able to) plug any OBD2 scanner into any 1996 and later vehicle with the connector and get *some* data that’s useful. For the rest, they all speak different dialects and some are a little slow.

    But then there’s CAN. Controller Area Network. Different protocols, different codes, same connector. Earliest implementation was in some 2004 vehicles. By 2006 all vehicles *should* be using CAN. Many of those cheap Bluetooth dongles that claim they support “All OBD2 standards” actually only work with CAN and will not work on any 2003 or older vehicles, and only work with the few 2004 models that were early adopters of CAN.

    I have two of those Bluetooth dongles. Neither worked on a 1998 Chrysler Sebring Convertible, 1997 Ford Taurus, 1998 Mercury Mountaineer nor a 2004 Dodge Dakota. I haven’t had a chance to try them on any 2005 or newer vehicles.

    If the vehicle was a brand new model in 2004, it likely uses CAN. If 2004 was a model’s final year (such as the 1997-2004 Dodge Dakota, which got a major makeover for 2005) it’ll still be OBD2. If the vehicle model carried on through at least 2005 it had to be redone for CAN. (The transition to front disc brakes in the USA was done in similar fashion. All new models for 1974 were required to have front disc brakes. Models being ended in 1974 did not. Models not being discontinued until 1975 or later had to get discs in 1974. Thus the new 1974 Mustang got them while the final year of the Maverick and Comet only had front discs as an option.)

    Overlapping phasing in of new things is done to not cost the manufacturers expenses they’ll likely not recover from a single model year – but sometimes the DOT makes them do a change for one year only anyway. That’s why the 2004 Dodge Dakota requires 16″ wheels due to the larger brakes Dodge was forced to use starting that year, despite already working on an all new Dakota for 2005.

    The early 1990’s were a period of rapidly changing DOT and EPA regulations. While fixing a 1994 Oldsmobile Acheiva I discovered there were many parts under the hood exclusive to 1994. The same part for 1993 or 1995 was a different shape, in a different location and/or had different connectors. I got to where I could spot a 94 in a junkyard from a fair distance if its hood was open or off. It didn’t help that the idiot who rear-ended a Subaru with it then attempted to “fix” it threw away a lot of parts, then believed the incorrect interchange data that claimed the Chevy Corsica body and other parts would fit. He tried to weld a Corsica radiator support onto the Acheiva and tried to make the Corsica bumper support fit under the Acheiva cover. Yes, I contacted the junkyard interchange software database company to set them right.

    1. 1994-1995 GM was a Very Special Place. At least for the popular (read: LT1-powered) vehicles there is still some ability to tune them, though I think even then a lot of the effort has gone towards documentation of converting to either a later LSx ECU or Megasquirt.

    2. Galane and Iowncalculus bring up information about older cars that I never knew before. …There was a Hackaday article in September from a fellow who used a configurable computer to replace the ecu in a Datsun. he ended up by cautioning that this made his 28oz sports car for “off road use only”. One of the discussion points brought up in the comments below that article was, that even though a specific car was running cleaner than factory specs, such a car would be failed in CA because it used non-CARB approved parts. To display my vast ignorance on the subject, My Question Is, how difficult is it/ what would be necessary to make a system such as a megasquirt output through a vehicle’s OBD II port in a way to reassure an Air Testing Station sniffer that it is indeed an appropriate part?

      1. @packrat. In Texas if you have an inspection license and an ASE master cert (not to mention a Texas Dept of Licensing and Regulation Dealers license for rebuilders.) You can build the car with Megasquirt as a “One Off” creation and then get it titled as a new vehicle. These vehicles will have a exemption that do not need emissions testing in Texas. This is what shops like Gas Monkey Garage do when building rides. Its called a salvage vehicle rebuilder endorsement.

        1. That’s good info. –I would still like to be able to have an ole car that successfully passes the tailpipe test to have it’s universal computer tell the test station computer, “yes, this is a 1996 MalaiseMobile Plasticfest 88, just like you expected, and everything is going dandy, thanks for asking”. For us that don’t have a SAE endorsement, that would seem to be… easier.

          1. In 90% of states, none of this matters (they only have basic emissions checks or nothing at all) but in California tricking them would be very difficult. They check the physical equipment and model numbers to make sure it is CARB approved. So you’d have to put fake model numbers on your equipment or have dummy parts.

    3. You gotta hate gov regulations that made it where outgoing models did not have to comply with the new regulations. I guess that is why the 1996 Fseries was the last of the good trucks before the modular madness and 4 and 6 cylinder ecoboost powered trucks.

  14. I love the fear over the ODB-II port nothing more than an interface to the true car network the CAN. As far as hackability goes it’s only gotten better after 07/08 when the federal government started mandating tire pressure monitoring in the wake of the Ford/Firestone debacle. Now every car after 2008 has a wireless interface you can exploit while on the road via the RF used on the tire pressure system. Since cars tend to integrate parts made by 3rd parties (see Takata airbags) this makes for a very easy to explicit part of the system right now since it’s insecure and easy to compromise and the car companies are very unaware because they buy the system capability as a black box from someone else.

    If you have something to fear it’s someone making a small SDR powered hack that attacks and does a buffer overflow in the TPM system to access the ECU/CAN network and do bad thing. Good luck finding that as a cause of crash.

    I leave you with this:

    Defcon has more info as well

    1. Here in Europe not all cars measure tyre pressure this way. My last company car just measured the turning speed of each wheel, as a wheel will get slightly smaller with lower pressure. It was quite effective and simple.

  15. Hate to say it folks but the beging of 1996 obd2 sandards mandated tha all systems speak the same code. When rolled out there has been a RF reciver placed in every car and truck. Car manufactures and insurance companys don’t want you to know that. Kinda of a insiders secret. Nothing new. Just another away for our big brother to keep an eye on us. Remember cash for clunkers program. As he’s keeping the fly swatter near by. Ready to strike

  16. CAN-bus is largely replacing OBD-II alone in cars from a diagnostic and hacking perspective. Car/Controller Area Network links a bunch of black-epoxy-blob “computers” together. The body-control-module(s), engine controls, drivetrain controls, etc etc.

    How do you think your newish car knows to turn up the radio more when you get to 40’ish MPH with the windows down, but only if the windows are down? Or maybe you drive a luxury vehicle that automagically turns on the wipers and headlights when it starts raining.

    CAN effectively supports all of the US federally-mandated standards of OBD-ii, with add-ons. Funny enough, it’s not always “open” enough for the home hacker. Sure, you *can* speak CAN, and I think you should. It just hasn’t been around as long as the previous iteration.

    Funny enough, different manufacturers also add features for systems-monitoring. Kia built a fuel-sensor-monitoring check into their software during the early 2000s that would turn on the check-engine light if the thermistor in the bottom of the fuel tank failed. Never you mind that your gauge works, the low-fuel lamp will not, so let’s fail your emissions test for you in most areas.

    Years of spinning wrenches for a living before I moved to IT taught me some interesting things. It’s entirely up to state audits to catch “cheaters” at smog tests. As an example, the smog-check guy can plug in your pickup but put the sticker on your race car. You’re less likely to get stopped by the cops if you have a sticker, so you don’t have to haul it via trailer to/from the track. Yet, the VIN in the engine-controller is a mismatch (as is the engine configuration). Which will usually flag for the auditors.

    I’ve seen a datsun engine missing half of one cylinder. The engine block split in a most spectacular fashion while racing. We could watch the piston and connecting rod move up and down as it ran! The OBD-ii engine computer thought it had a steady misfire.

  17. …almost forgot.

    in the 80s and early 90s for Chevrolet and GMC pickups, you could dremel a car-key into a slotted two-pin jumper. That was your diagnostic tool. Short two pins, short two different pins, watch the maintenance lamp flash out it’s code-sequence.

    Some techs got creative and ordered the male connector from GM, then added a switch to toggle jumper-states.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.