EPA caught VW cheating - how does the car know it's being tested?

[Ewald B?hm]

You have a good point. I need to recheck my facts.

[Mitch Kaufmann]

The software code allows all of the car's emissions systems to work when the cars are taken in for clean-air testing. But as soon as the emissions tests are complete, the system reverts to spewing pollutants. The cars emitted nitrogen oxide at a level of up to 40 times the standard level, the EPA alleges.

[Sofa Slug]

According to the LA Times:

"Rather than meet the standards, the EPA says VW sneaked in the defeat device software to detect when the car is hooked up to a dynamometer, a machine that measures emissions. When emissions are being measured, the defeat device tells the car to operate at "dyno calibration," or full emission control levels, to meet the standards."

"At all other times, however, the software sets the engine to run on "road calibration," allowing the excessive emissions. How can the program tell the difference? By noting the position of the steering wheel, variations in speed and other data that suggest no one is driving the car, and thus it is likely being tested."

[sms]

Well in California they definitely check via the OBD-II port. I had replaced a battery and there were no dashboard lights indicating anything. The first thing they did was to do a scan for codes.

The number of pending codes that is allowable varies by year of manufacture. A good shop will tell you the drive sequence to clear the pending codes for each model. A bad shop won't even know this information.

[sms]

In California, one "smog check factory" in L.A. got caught because the state checked registered addresses of the vehicles and wondered why so many vehicles were being smogged at this one particular shop when their registered address was so far away. Few people will drive 25 miles in L.A. to get a smog check at a particular shop.

My brother-in-law regularly had inspectors come into his shop with test vehicles to be smogged. They would reveal who they were after the test. He did really well. He got one demerit for not telling the "customer" that they had the option of getting the vehicle repaired at his shop or any shop, even though he did ask if they wanted it to be repaired. But he still passed the inspection.

[clare]

Actually it is not pending codes that are the issue. It is the readiness monitors.. Can't remember how many readiness monitors there are - but there's a catalyst monitor, a O2 sensor monitor, and EGR monitors, and O2sensor heater and cat heater monitor on some vehicles. These are the intermittent monitors that need to be "set" .

Setting the monitor just means they have been through one or more test sequences and have aquired valid data..

The rest of the monitors are contimuous monitors - misfire, component, and fuel system, nonitors.

The evap monitor, for instance, is only "valid" in a fixed temperature range, and with the tank between something like 1/4 and 3/4 full (not

100% sure of the actual numbrs). If you reset the codes or replace the battery on a vehicle with the tank full or almost empty you can NOT set the readiness monitor for the evap system - so virtually ALL OBD2 based emission test facilities will allow at least one monitor to be un-set or not ready.

If you know what code is coming up, and want to "cheat" the system, if you can avoid setting that particular monitor, while setting all the others, you can sometimes get a vehicle to pass. You need to understand the drive cycle and what can cause the monitor you want dissabled to fail to set. (and it needs to be an intermittent or non-continuous monitor. The usual culprits are Cat, evap, or EGR.

[trader_4]

I was thinking that too, ie that it's the monitors, not pending codes. Pending codes would be fault conditions that have been detected but haven't occurred enough to put activate the check engine light. The emissions monitors are flags that are set on the critical emission related systems. If you clear the computer, those flags get cleared. When the car is driven, they get set again over time as the system accumulates info that shows they are working. There are about 10, I think. Most get set within minutes. The fuel evap system takes the longest, probably takes multiple drive cycles to set.

Here in NJ with older cars you can pass inspection with two not set. Newer cars, no more than one can be not set. Maybe some states are concerned with actual pending codes, but I doubt it, because they don't necessarily mean anything is wrong and would just create a lot of drama for nothing.

[Ewald B?hm]

It's usually documented as the FTP.

[Ewald B?hm]

In California, they can refuse the test if the readiness monitors aren't fully set, even if you'd still pass otherwise.

They get dinged on demerits for every readiness monitor not set, even if it's within the limits of the law.

Such is bureaucracy.

[Ewald B?hm]

That answers what. But it doesn't answer HOW.

[Ewald B?hm]

Finally!

Someone who both understood the question, and who posited an answer!

Of all the posters, you're the ONLY one who understood the question!

[Vincent Cheng Hoi Chuen]

Sofa Slug wrote in mtn6qu$2in$ [e-mail]:

What I'm surprised at is that each state can have a *different* procedure.

In California, they use the dyno, but in many less technical states, they still use the dumb procedures.

This explains how they noticed there was testing going on.

But that only works for the intelligent states. How did they also fool the low-tech states like NJ, Kentucky & Kansas?

[Ewald B?hm]

The best answer to the question seems to be here, as noted by Sofa Slug:

"How can the program tell the difference? By noting the position of the steering wheel, variations in speed and other data that suggest no one is driving the car, and thus it is likely being tested."

Apparently VW lied at first, & apparently they can no longer sell the cars:

"The cheating came to light when the California Air Resources Board and the EPA pressed Volkswagen for an explanation for disparities found between lab tests and road tests of its vehicle emissions. The agencies didn't find the technical reasons offered by VW to be convincing and said they would not issue certificates allowing 2016 models to be sold until the automaker offered an adequate explanation. "Only then did VW admit it had designed and installed a defeat device in these vehicles," the EPA said. VW said it was cooperating with the investigation but otherwise had no comment."

It's interesting that VW didn't fess up until they were forced to.

[trader_4]

Show us your references and data that show using the OBD method is "dumb". I guess the federal EPA must be dumb, because using the OBD is acceptable to them. The car is constantly and closely monitoring itself as the car is actually drive, so why exactly do we need a dyno? We had the dynos here in NJ for a few years, the system cost hundreds of millions, was a complete fiasco, there were huge lines waiting for inspection because of the time it took, etc. In just a few years they went to the scrap heap.

I'd say the real fools are guys like you, who think you need a dyno. And before you try to smear some states, realize that using OBD is acceptable to the Feds now, while they required the dyno in years gone by. And provide us a list of the states that still do a dyno test. I think you'll find that it's not many.

Just the facts.

[Dean Hoffman]

That's pretty much human nature going back to the Garden of Eden. The next trick is to blame someone else. A TV show from long ago had comedian Flip Wilson on. His line was "The devil made me do it".

-- =

Using Opera's mail client:

[Vincent Cheng Hoi Chuen]

trader_4 wrote in [e-mail]:

OK. Just the facts Danno:

[makolber]

this proves the EPA is not dumb.

They got their buddies selling dynos to be rich.

and now they are doing the same thing by "defining" CO2 as "pollution"

Dumb like a fox. Follow the money.

Mark

[sms]

Yes, sorry, that's what I was referring to.

Can't remember how many readiness monitors there

By the way, there's an excellent Android app for OBD-II called Torque Pro. The app is $4.95, and a Bluetooth ELM327 OBD-II adapter is less than $10 (I am using this one ). The app does a lot more than just read or clear codes. It will display electronic gauges based on the sensor readings (especially useful for vehicles without temperature gauges or tachoometers). You can set alarms for things like over-temperature. It's also a very accurate speedometer (via the GPS), and it'll measure things like 0-60.

There's no iOS version because Apple forgot to include the necessary Bluetooth profile (SPP) in its devices. There are similar apps for iOS but not nearly as good. This one is one of them but you need to get a Wi-Fi ELM327 dongle, not a Bluetooth one.

I like having a 7" tablet with TorquePro and CoPilot (GPS). I made a holder for the tablet using one of the Panavise mounting brackets . Just be sure the tablet has a GPS chip, since very low-end Android tablets don't have one, nor do Wi-Fi only iPads. You can buy a decent Asus 7" tablet with a GPS for $50 just use a virtual credit card with a $1 limit and a one month expiration date when you sign up with McAffee (required to get the rebate). Intel, which owns McAffee, is trying to promote devices with their processor inside, hence the large rebate.

[trader_4]

Thanks for providing the proof that you're wrong. You posted:

"In California, they use the dyno, but in many less technical states, they still use the dumb procedures. "

And from the executive summary in the CA document you just provided:

"Available data and information indicate that Smog Check tailpipe testing o f OBD II equipped vehicles significantly increases testing costs and inconv enience to California motorists, but provides only minimal emission benefit s that are above and beyond those that can be realized through OBD II-based inspections. The procedure for co nducting an OBD-based inspection can be completed in 5 minutes or less, com pared to 20 minutes for a tailpipe test, and the equipment required for the inspection can be purchased for as little as 10% of the cost for the analy zer and dynamometer needed for tailpipe testing."

Even CA agrees that there is little benefit to dyno testing, that it takes

4 times as long and that the eqpt costs 10x. If CA is still doing it, then it's CA that's dumb, not the other states that used OBD.

[Lucia Gallo]

Staying on topic, this article says the cheat only worked when there was a DYNO involved!

Here are the contiguous quotes: (begin quote) VW used secret software ? an algorithm that detects when cars are being tested on treadmill-like devices called dynamometers, and stealthily switches the engines to a cleaner mode.

Because *smog tests are almost always done on dynamometers*, VW got away with the scheme for seven years, until the "clean transportation" advocates went to West Virginia University, which tests emissions using equipment that fits in car trunks. (end quote)