Anybody know how an o2 sensor works?
07-05-2005, 06:31 PM
#1
Racer
Thread Starter
Member Since: Jun 2003
Location: Lititz PA
Posts: 368
Likes: 0
Received 0 Likes
on
0 Posts
Anybody know how an o2 sensor works?
I am in the process of putting a dual exhaust on my 79. I am contemplating putting an o2 sensor on each pipe.
From basic electrical knowledge, I presume that one would feed 12VDC to the sensor and the output would be variable relative to the oxygen in the exhaust gas, ( milliamp or millvolt output would vary ).
Does anybody know what the industry standard is for the relation of oxygen to milliamps/volts? If so, what then is the correlation between oxygen content in the exhaust gas and the initial air/fuel ratio??
Hope all of the above makes sense,
Thanks
Andy
From basic electrical knowledge, I presume that one would feed 12VDC to the sensor and the output would be variable relative to the oxygen in the exhaust gas, ( milliamp or millvolt output would vary ).
Does anybody know what the industry standard is for the relation of oxygen to milliamps/volts? If so, what then is the correlation between oxygen content in the exhaust gas and the initial air/fuel ratio??
Hope all of the above makes sense,
Thanks
Andy
07-05-2005, 07:04 PM
#2
Race Director
That's not quite how it works. The classic narrow band oxygen sensor is much like a galvanic element. It uses the difference between two elements to produce a voltage. In this case the oxygen content of the exhaust compared to the oxygen content outside the exhaust.
This function takes quite a lot of heat to happen. Exhaust temps are usually enough to kick off the reaction, but it gets iffy at idle and takes a while after a cold start so newer sensors have a built in heater. That's what the 12V are for.
As for voltages, the only really accurate correlation is 0.45V=stochiometeric mixture. Stoich is 14.7:1. Leaner than 14.7:1 quickly goes to 0V and richer quickly goes towards 1V. Exactly what the ramp looks like varies between brand of sensor, the actual sensor as well as sensor temp, barometric pressure, exhaust temp etc. You can't really say that 0.6V=14.0:1. you can say 0.6V is richer than 14.7:1 but not as rich as, say, 12:1.
If you want accuracy away from 14.7:1 you need a wideband oxygen sensor. The WB requires a special electronics box to run it. You're looking at over $300 for the kit. They're a great tuning aid though and I recommend them. Most of them have built in loggers that let you export the data to a PC for review.
This function takes quite a lot of heat to happen. Exhaust temps are usually enough to kick off the reaction, but it gets iffy at idle and takes a while after a cold start so newer sensors have a built in heater. That's what the 12V are for.
As for voltages, the only really accurate correlation is 0.45V=stochiometeric mixture. Stoich is 14.7:1. Leaner than 14.7:1 quickly goes to 0V and richer quickly goes towards 1V. Exactly what the ramp looks like varies between brand of sensor, the actual sensor as well as sensor temp, barometric pressure, exhaust temp etc. You can't really say that 0.6V=14.0:1. you can say 0.6V is richer than 14.7:1 but not as rich as, say, 12:1.
If you want accuracy away from 14.7:1 you need a wideband oxygen sensor. The WB requires a special electronics box to run it. You're looking at over $300 for the kit. They're a great tuning aid though and I recommend them. Most of them have built in loggers that let you export the data to a PC for review.
07-06-2005, 08:00 PM
#3
Racer
Thread Starter
Member Since: Jun 2003
Location: Lititz PA
Posts: 368
Likes: 0
Received 0 Likes
on
0 Posts
I have a multimeter that can supply up to 24 VDC and monitor output on another channel in mA or mV. Will this work to read a wideband o2 sensor?
Cheers
Andy
Cheers
Andy
07-06-2005, 08:11 PM
#4
Very simply put it works like this:
A narrow band O2 works on the prinicple of a potential difference (concentration difference) over a membrane, the membrane being the zirconia element of the O2 sensor, at operating temp it's permeable for oxygen (lets O2 through) and as a basic natural occurrence different concentrations when in contact want to even out (same principle as osmosis) so the higher O2 content in outside air wants to penetrate the exhaust system through the zirconia element to even out the concentrations in and outside, this passing of the O2 molecules through the zirconium develops the voltage (think of it as a resistance creating heat) and this voltage is dependant on the condition of the exhaust gas (O2 concentration), the bigger the difference the larger the driving force, the higher the voltage output (the sensor creates it's own voltage)
The voltage signal on these puppies is not linear, it's only linear for a very small section of the entire resolution, close to Lambda =1 (14,7:1 A/F) and it drops off from that rather quickly (logarithmic), that's why it's called a narrow band and the voltage is not in a direct linear relationship to the mixture, it can be improved a bit however by using a led bar that has the entire resolution sectiond in areas w/ different scaling, this allows you to scale it close to a linear graph. Someone I know was building just such an AF meter, will have to check if he still has the printed circuits and the components to build one.
A narrow band O2 works on the prinicple of a potential difference (concentration difference) over a membrane, the membrane being the zirconia element of the O2 sensor, at operating temp it's permeable for oxygen (lets O2 through) and as a basic natural occurrence different concentrations when in contact want to even out (same principle as osmosis) so the higher O2 content in outside air wants to penetrate the exhaust system through the zirconia element to even out the concentrations in and outside, this passing of the O2 molecules through the zirconium develops the voltage (think of it as a resistance creating heat) and this voltage is dependant on the condition of the exhaust gas (O2 concentration), the bigger the difference the larger the driving force, the higher the voltage output (the sensor creates it's own voltage)
The voltage signal on these puppies is not linear, it's only linear for a very small section of the entire resolution, close to Lambda =1 (14,7:1 A/F) and it drops off from that rather quickly (logarithmic), that's why it's called a narrow band and the voltage is not in a direct linear relationship to the mixture, it can be improved a bit however by using a led bar that has the entire resolution sectiond in areas w/ different scaling, this allows you to scale it close to a linear graph. Someone I know was building just such an AF meter, will have to check if he still has the printed circuits and the components to build one.