Testing A/F sensors using Techstream (2 Viewers)

[white_lx]

(This applies to '06 and '07 models only. Maybe also a few late year '05.)

Air-Fuel Sensors can fail over time because of failed heaters. These failures are pretty obvious as the ECM will throw a P0051/P0052. However, they can also fail without it being immediately obvious (no CEL). Or sometimes they will report a CEL, but at first glance the fuel trims might be very reasonable.

Below is a procedure for testing A/F sensors using the "control injection volume for A/F sensor" Active Test in TechStream. In this case it identified that a P0174 was a result of an A/F sensor, even though during testing the fuel trims readings were low.

First a little background. It is difficult to interpret the A/F sensor readings as they are in arbitrary voltage units. Luckily TechStream also presents the interpreted lambda values. Lambda is the ratio of current oxygen to fuel ratio vs an ideal stoichiometric (balanced) mixture of 14.7. So for an engine that runs perfect, lambda = (current O/F) / (desired O/F) = 14.7 / 14.7 = 1.00.

TechStream allows increasing or decreasing the fuel volume using an active test by +12.5% or -12.5% respectively. For those cases, the expected lambda value is:

+12.5% fuel (rich): lambda = (14.70/(1.00+0.125)) / 14.7 = 0.85
- 12.5% fuel (lean): lambda = (14.70/(1.00-0.125)) / 14.7 = 1.14

The calculated rich and lean lambda's are well within the linearity range of a typical A/F sensor of lambda = 0.7 to 1.3. So we would expect a proper functioning A/F sensor to reasonably read the theoretical lambda values if we increase or decrease the fuel volume by +/- 12.5%.

Below are some screenshots of an actual test. First the baseline situation (0% injector volume change):

AF Lambda Bank 1 and Bank 2 are both around 1.00. This is the lambda ratio the ECM tries to maintain. Note however that this does not mean that the actual lambda in both banks is 1.00. It is just the AF sensors reading a lambda of 1.00.

At -12.5% the lambda values are 1.149 and 1.118 for Bank1/2. The expected value is 1.143. Hmmmm, it seems that the sensor on bank 2 is off. But no error codes, and is the difference relevant enough? Next test at +12.5%.

Now it is obvious that Sensor 2 is not reading correctly. At +12.5% fuel volume it is only registering a lambda value of 0.959 vs 0.891 for other bank. The expected value lambda value for the introduced rich condition is 0.85.

Note that if the A/F sensor is further deteriorated, it will often be immediately obvious when increasing or decreasing the fuel trims by +/- 12.5% as the secondary oxygen sensors will fail to report the introduced lean or rich conditions. In this case the secondary oxygen sensors were following the lean / rich condition making the diagnosis a little trickier.

 

[white_lx]

A/F sensor showing rich combustion while registering a lambda~1.0, indicating a failing ('lying') A/F sensor.

Replaced the B2 sensor. L&R banks now read similar lambda values when injection volumes are changed.

 

[jerryb]

great post.
if we were in a circle, drinking beer. I'd say, and do you think those three pictures explain why toyota trucks get s***ty gas mileage and last longer than average?

I always get F-ed up when people talk lambda. I'm watching voltage parity while driving. I don't know why you say it's arbitrary.

I would be much happier if an LC idled and had a learner freeway cruise, much like -12%.

This clip is a great indicator for a- I need to check something moment.

 

[white_lx]

great post.

Thank you. I thought it might help out someone one day.

if we were in a circle, drinking beer. I'd say, and do you think those three pictures explain why toyota trucks get s***ty gas mileage and last longer than average?

This only applies to A/F sensors, not to regular oxygen sensors. A/F sensors were only introduced around 2004 in Toyota products, so you could argue we don't have enough data yet.

I always get F-ed up when people talk lambda. I'm watching voltage parity while driving. I don't know why you say it's arbitrary.

The A/F sensor does not measure voltage. The sensor output is a very small current (micro amps). If you were to probe the sensor there would be no 3.3V to be measured anywhere. It is a 'fictitious' (arbitrary) voltage such that the scan tool has a value to read. It might as well have been chosen to be 2.0V or 4.0V.

The problem with the voltage is, that it doesn't really mean anything. For example, if the A/F sensor is reading 2.5V. We only know it is rich. But how rich? 1% too much fuel, 50%? We have no idea. However, from the scan tool we know lambda = 0.88, i.e. 14% too much fuel.

The data above does provide guideline voltages, if you wanted to ignore the lambda values:
Normal conditions: 3.3 V
+12.5% fuel volume: 2.5 V
-12.5% fuel volume: 3.7 V

This clip is a great indicator for a- I need to check something moment.

Your B1S1=2.5V data is indeed way off. What is strange, is that the ECM is not driving it to 3.3V. Is the fuel trim maxed out?

 

[2001LC]

Thank you for posting this info. It is and area I need to study up on.

I notice B2S2 voltage, before and after R&R of B2S1. Is always lower than B1S2V. Even where B2S1 Lambda was not always lower. Does this indicate anything?

 

[white_lx]

I notice B2S2 voltage, before and after R&R of B2S1. Is always lower than B1S2V. Even where B2S1 Lambda was not always lower. Does this indicate anything?

The secondary oxygen sensors are regular oxygen sensors. Their actual voltages do not really have a lot of meaning other than above 0.45V is rich and less than 0.45V is lean. This is where they differ from the wide-band oxygen sensors (AF sensors) as they are actually able to measure how lean and how rich. Since the oxygen sensors typically switch, there is no conclusion to be drawn from just a single reading. However, graphing them will tell if the exhaust is continuously rich or lean which can then be compared to the AF sensor readings.

Note that the way the regular oxygen sensors report is reverse from the AF sensors:
- Voltage of an AF sensor above 3.3V is a lean condition, but for a regular oxygen sensor a voltage below 0.45V is a lean condition;
- Voltage of an AF sensor below 3.3V is a rich condition, but for a regular oxygen sensor a voltage above 0.45V is a rich condition.

The fuel injection volume test mode is also very useful for determining the condition of a catalytic converter (the oxygen storage test). When switching from +12.5% to -12.5% fuel volume, it will take a certain amount of time before the secondary oxygen sensors switch to lean because the initial rush of excess oxygen will be absorbed by the catalytic converter. If the switch-over is very quick (< 2 seconds or so), the catalytic converter is depleted. If it takes the secondary oxygen sensors longer to go lean, the catalytic converter is in much better condition (~ 5 seconds).

 

[jerryb]

This only applies to A/F sensors, not to regular oxygen sensors. A/F sensors were only introduced around 2004 in Toyota products, so you could argue we don't have enough data yet.



The A/F sensor does not measure voltage. The sensor output is a very small current (micro amps). If you were to probe the sensor there would be no 3.3V to be measured anywhere

The problem with the voltage is, that it doesn't really mean anything.



Your B1S1=2.5V data is indeed way off. What is strange, is that the ECM is not driving it to 3.3V. Is the fuel trim maxed out?

Another reason toyota tunes for reliability and longevity vs mileage. Toyota might have been the last maker to use WB. Toyotas have always run rich.

I have no clue how anyone would say,
the problem with voltage is that it doesn't mean anything.

Techstream outputs 2 paired values by default. Given lambda, or voltage, I'm going to teach voltage all day long.
Whether it's linear or not the displayed voltage corelates to lambda and AFR.
Indeed that if voltage didn't matter techstream would not display it.

I get null output, I know the reference V. I know these densors won't read over 18/1 and below 10/1.

That clip I posted was from your pictures. I think you thought it was my LC or something
Your B1S1=2.5V data is indeed way off. What is strange, is that the ECM is not driving it to 3.3V. Is the fuel trim maxed out?

There is usefulness in the +12 -12% tests. Relevant at idle, in your driveway.
I propose graphing sensor voltage response with trim under load and highest achievable vacuum is a better way.

 

[abuck99]

(This applies to '06 and '07 models only. Maybe also a few late year '05.)

Air-Fuel Sensors can fail over time because of failed heaters. These failures are pretty obvious as the ECM will throw a P0051/P0052. However, they can also fail without it being immediately obvious (no CEL). Or sometimes they will report a CEL, but at first glance the fuel trims might be very reasonable.

Below is a procedure for testing A/F sensors using the "control injection volume for A/F sensor" Active Test in TechStream. In this case it identified that a P0174 was a result of an A/F sensor, even though during testing the fuel trims readings were low.

First a little background. It is difficult to interpret the A/F sensor readings as they are in arbitrary voltage units. Luckily TechStream also presents the interpreted lambda values. Lambda is the ratio of current oxygen to fuel ratio vs an ideal stoichiometric (balanced) mixture of 14.7. So for an engine that runs perfect, lambda = (current O/F) / (desired O/F) = 14.7 / 14.7 = 1.00.

TechStream allows increasing or decreasing the fuel volume using an active test by +12.5% or -12.5% respectively. For those cases, the expected lambda value is:

+12.5% fuel (rich): lambda = (14.70/(1.00+0.125)) / 14.7 = 0.85
- 12.5% fuel (lean): lambda = (14.70/(1.00-0.125)) / 14.7 = 1.14

The calculated rich and lean lambda's are well within the linearity range of a typical A/F sensor of lambda = 0.7 to 1.3. So we would expect a proper functioning A/F sensor to reasonably read the theoretical lambda values if we increase or decrease the fuel volume by +/- 12.5%.

Below are some screenshots of an actual test. First the baseline situation (0% injector volume change):

View attachment 3361221

AF Lambda Bank 1 and Bank 2 are both around 1.00. This is the lambda ratio the ECM tries to maintain. Note however that this does not mean that the actual lambda in both banks is 1.00. It is just the AF sensors reading a lambda of 1.00.

View attachment 3361222

At -12.5% the lambda values are 1.149 and 1.118 for Bank1/2. The expected value is 1.143. Hmmmm, it seems that the sensor on bank 2 is off. But no error codes, and is the difference relevant enough? Next test at +12.5%.

View attachment 3361223

Now it is obvious that Sensor 2 is not reading correctly. At +12.5% fuel volume it is only registering a lambda value of 0.959 vs 0.891 for other bank. The expected value lambda value for the introduced rich condition is 0.85.

Note that if the A/F sensor is further deteriorated, it will often be immediately obvious when increasing or decreasing the fuel trims by +/- 12.5% as the secondary oxygen sensors will fail to report the introduced lean or rich conditions. In this case the secondary oxygen sensors were following the lean / rich condition making the diagnosis a little trickier.

Well written & great technical discussion. Thanks for posting.

 

[white_lx]

That clip I posted was from your pictures. I think you thought it was my LC or something
Your B1S1=2.5V data is indeed way off. What is strange, is that the ECM is not driving it to 3.3V. Is the fuel trim maxed out?

Ah, my bad, I didn't memorize the voltages. That makes sense then. In test mode the fuel control is OL and that's why the B1S1 was 2.5V. Sorry for the confusion.

For reference, see the the data for Toyota A/F sensors in the table below. When not using TechStream, the alternate normalized voltage readings may be shown when using a scanner in OBDII mode.

 

[white_lx]

Straight from the horse's mouth. Very good presentation, specific to Toyota A/F sensors: