Vehicular emissions sensors ?

I am thinking of a developing project to detect vehicle emissions, CO2, CO, NOX, HC, smoke, etc. I was able to locate some gas sensors compatible with Arduino; but, most seem to be meant for indoor air quality. I am not sure, if these sesnors would be applicable for vehicular emissions. Firstly, the exhaust will be hotter and secondly, the concentration may be higher than domestic uses.

Has anyone tried such a project ? Please guide me how to develop this project.

Note, I am not trying the OBD II route because the vehicle in question is quite old (~20 years !).

First you should check the datasheets of the sensors

  • for the temperature range,
  • if they match the expected PPM range

You can cool the exhaust gasses to acceptable levels

  • ribs
  • water bubbler.
  • mix with nitrogen (e.g. 1:10) and adjust the math

Every UK garage that does MOT tests (required by law, every year, on vehicles over 3 years old) has a machine for measuring exhaust emissions. It measures the CO, Hydrocarbons and the Lambda (whatever that is?).
A narrow metal tube is placed inside the exhaust tail pipe and the exhaust gasses are gently sucked through a 20 ft flexible tube to the machine. I suspect that the gasses, by the time they reach the machine, are sufficiently cooled so as not to be a problem.
If you use a length of exposed metal tubing to take the gasses from the tail pipe, any excess heat will be dissipated by the time the gasses reached your sensors. Your system will have to allow for any condensation that may form. I suggest using a peristaltic pump to suck the gasses.

Edit: Exhaust gasses, once they have passed through the silencer, possibly the catalytic converter and the rest of the exhaust system are not very hot, usually not above about 50oC.

Thank you, for the responses !

Exhaust gasses, once they have passed through the silencer, possibly the catalytic converter and the rest of the exhaust system are not very hot, usually not above about 50°C.

That seems to be my concern. For example, this datasheet for CO sensor - MQ-7 - reports the operating temperature limit as -20°C to 50°C. Wouldn't that be cutting too close ?

I would probably have to run experiments to find out; but, I was just curious if sensors existed that were 'automobile grade'.

cogitoergosum:
I was just curious if sensors existed that were 'automobile grade'.

Automotive test equipment from the likes of Sun, Allen, MZ have been able to measure various emissions for decades, so yep such sensors must exist. Not sure what their voltage requirements might be though, since workshop equipment will all be mains driven and could be at any voltage the makers liked to step down to internally.

Maybe try to find spares for those machines, or cannibalize scrapped ones?

cogitoergosum:
I am thinking of a developing project to detect vehicle emissions, CO2, CO, NOX, HC, smoke, etc. I was able to locate some gas sensors compatible with Arduino; but, most seem to be meant for indoor air quality.

I know that most of those sensors used in parking garages or industrial locations need to be calibrated every year. Once seen someone do it and they have a tendency to “run” around ±10% every year.

Not sure if that is applicable in your sensors though…

Not sure what their voltage requirements might be though, since workshop equipment will all be mains driven and could be at any voltage the makers liked to step down to internally.

I was thinking whether sensors existed that could be hooked upon a Arduino.

BTW, I came across Oxygen sensor (also referred to as lambda sensor). The focus seems to be more on fuel efficiency and the impact on emissions when the efficiency reduces.

Now browsing this topic - Oxygen Sensor question.

Hi, I spent 15years travelling over the countryside servicing and repairing automotive service equipment, engine analysers, electronic wheel balancers, wheel aligners and exhaust gas analysers.

The brands mentioned, Sun, Allentest, Blackhawk, Vane etc used infrared gas analysis.
CO,CO2, and HC (unburnt hyrocarbons) used this method, lambda (O2) used a separated detector that was basically a zinc/air battery.

The O2 unit had a life of about 2 years, even if you did not use it, I used to do calibrations as well.

The gas is analysed by detecting the absorption of IR by the gas sample, CO, CO2 and HC absorbed different wavelengths.

Some of the analysers were very simple units that needed constant zero adjustment, others very stable setups with feedback and temp controlled chassis.

I'm not sure how accurate or what resolution the standard heater block type of detector is, they usually are only used to detect the gas, not necessarily measure its concentration.

Tom..... :slight_smile:

cogitoergosum:
Thank you, for the responses !

Exhaust gasses, once they have passed through the silencer, possibly the catalytic converter and the rest of the exhaust system are not very hot, usually not above about 50°C.

That seems to be my concern. For example, this datasheet for CO sensor - MQ-7 - reports the operating temperature limit as -20°C to 50°C. Wouldn't that be cutting too close ?

Not if you do as I suggested above: "use a length of exposed (to cool air) metal tubing to take the gasses from the tail pipe, any excess heat will be dissipated by the time the gasses reach your sensors. "
I think that about 6 ft (2 Mtrs) would be sufficient, depending on the airflow and temperature around it. The tubing doesn't have to be straight!

All emissions equipment uses a sniffer and a pump to sample the exhaust.

Cool exhaust is sucked into the machine, i have never met an emissions station that reads Lambda {air fuel ratio}

Lamda correlates to a cars performance, but not to its emissions, specifically.

If you know the NOX the lambda is irrelevant.

The sensors you are missing are very expensive, and have a limited life span. The number of gases you analyze will affect your all up cost.

Most modern machines are 4 gas analyzers. Your best bet would likely to take an entire probe off an old machine, and build your own control scheme for it.

Old Dyno Cells will generally have what you need, and NO ONE ever uses them!

If you take a temp reading at the tip of your tail pipe, you will see the exhaust is not very hot at all. The spray paint test along the exhaust run is an easy one to perform, and illustrates where the heat "stops"

Kawgomoo:
All emissions equipment uses a sniffer and a pump to sample the exhaust.

Cool exhaust is sucked into the machine, i have never met an emissions station that reads Lambda {air fuel ratio}

All the ones used for MOT testing in the UK measure Lambda. They have to, by law.

Lamda correlates to a cars performance, but not to its emissions, specifically.
If you know the NOX the lambda is irrelevant.

And vice versa. UK exhaust testers do not measure NOX, only CO, Hydrocarbons and Lambda.

I've been around Sun testers since c1969: the ones my family use/d all have air/'fuel meters. Admittedly back in the day the mechanic had more access to adjustments (screws on carbs, jet sizes, float height, all sorts of stuff) to keep the ratio at the stochiometrically correct 14.7- not much to do nowadays except maybe shake the shit out of the aircleaner element and change the plugs.

JimboZA:
I've been around Sun testers since c1969: the ones my family use/d all have air/'fuel meters. Admittedly back in the day the mechanic had more access to adjustments (screws on carbs, jet sizes, float height, all sorts of stuff) to keep the ratio at the stochiometrically correct 14.7- not much to do nowadays except maybe shake the shit out of the aircleaner element and change the plugs.

So the lambda reading I got in July, for my old car, of 1.015 (on a Sun DGA 2500 MOT) means it was 1.5% above that 14.7 ratio? 14.7 * 1.015 = 14.9205. Not bad for a 15 year old car.

Hi, I have supplied some links to indicate the complexity of getting a stable and accurate gas analyser.
The SUN unit uses a SIEMENS IR gas bench, and when I was servicing ALLEN benches I was told that they were originally designed for blood and fluid analysis.
Lambda is a calculated value, not measured directly from the gas.

http://www1.snapon.com/display/3791/UserManuals/DGA_2500_OM_GB_EEC_Version_Rev_C.pdf

http://www.bridgeanalyzers.com/EGA/Automotive/mediaRepos/productDocs/White%20Paper%209.pdf

Hope this helps.

Tom...... :slight_smile:

TomGeorge:
Hi, I have supplied some links to indicate the complexity of getting a stable and accurate gas analyser.
The SUN unit uses a SIEMENS IR gas bench, and when I was servicing ALLEN benches I was told that they were originally designed for blood and fluid analysis.
Lambda is a calculated value, not measured directly from the gas.

http://www1.snapon.com/display/3791/UserManuals/DGA_2500_OM_GB_EEC_Version_Rev_C.pdf

http://www.bridgeanalyzers.com/EGA/Automotive/mediaRepos/productDocs/White%20Paper%209.pdf

Hope this helps.

Tom...... :slight_smile:

Thanks, Tom. Wow! That Lambda formula is mighty complicated.