Analog vs Digital O2 Sensors

[sjrobinson]

I’ve looked into this in the past and as I analyze the data from other cars I notice something odd about mine. I’m wondering if its model and make specific or perhaps an underlying issue. I’m always wanting to keep my Accord in good condition.

If the info I read was correct, with the introduction of downstream and upstream sensors in a car, they were made with two different types of data collection: Analog and Digital.

The upstream o2 sensor, collecting data for AFR, is said to send digital signals. The downstream, reading the cat converter efficiency, is analog.

When I analyze different cars using a scan tool I see the spikes (digital) of the upstream sensor and the flat lines (analog) of the downstream sensor. However on my car I see analog data coming from both sensors. My car is a 97 Accord. Is that how the sensors were designed or is there a problem? They are aftermarket sensors, would a downstream sensor placed in the upstream spot cause this?

I’m curious since I would think that a digital signal would provide quicker and more accurate AFR readings for the computer.

Any ideas? I would love to learn more about the nuances in o2 sensor design and behavior.

Thanks

Topic

[Paul]

I am currently diagnosing an emission problem with my Honda and have been researching AFR sensors. I’m no expert on the subject.

From what I can tell, the air-fuel ratio sensor is designed to be more sensitive for measuring O2, but over a narrower range than a conventional oxygen sensor. They look the same, but they have some internal differences. As the AFR is more sensitive, the signal changes should appear more dramatic compared to the signal secondary/downstream oxygen sensor. Maybe this is what you see as spiking in the signal.

Also, I would expect both sensors operate in analog mode, and the signals are digitized at the computer/control module input because heat would be a problem for a digitizer intergrated into the sensor body.

Again, I am not an expert.

[sjrobinson]

Thanks for your input. I’ve wondered what would make one digital and the other analog right at the sensor. At the ECU would make more sense. But I should mention that I am readings these at the OBD2 port so I’m seeing them as the ECU would translate them.

In comparing what I find with my car to other cars it seems that my upstream sensor signals are not as quick or ‘narrow’ as others.

[sjrobinson]

Bump?

[Gumpy Guss]

All sensors that measure a continuous, variable quantity, such as air pressure, oil pressure, gasoline level, manifold vacuum pressure, throttle position, mass air flow, temperature, mixture, and oxygen, are at the lowest level, analog sensors. Sensors that measure discrete events, such as brake light activation, crankshaft position and revolutions, those can be either analog or digital at the source.

As electronics has advanced, it’s become possible and indeed desirable to convert these continuous analog values to digital as soon as possible, sometimes right inside the sensor, as it’s easier to transmit digital signals. Analog signals often need thicker shielded cables to avoid distorting the analog voltage values. And digital signals can be easily put onto a common bus, minimizing the number of separate wires going to the ECU.

So you will find a mixture of allegedly “analog” and “digital” sensors, some truly analog, some truly digital, some a hybrid combination of the two.

[sjrobinson]

I understand. That makes more sense.

So why would my car vary from all the other cars I’ve looked at? Even ones as old as mine? I see spikes in the upstream o2 sensor whereas mine is a bit slower to respond.

[Kevin Criswell]

[quote=”Hanneman” post=74427]I am currently diagnosing an emission problem with my Honda and have been researching AFR sensors. I’m no expert on the subject.

From what I can tell, the air-fuel ratio sensor is designed to be more sensitive for measuring O2, but over a narrower range than a conventional oxygen sensor. They look the same, but they have some internal differences. As the AFR is more sensitive, the signal changes should appear more dramatic compared to the signal secondary/downstream oxygen sensor. Maybe this is what you see as spiking in the signal.

Also, I would expect both sensors operate in analog mode, and the signals are digitized at the computer/control module input because heat would be a problem for a digitizer intergrated into the sensor body.

Again, I am not an expert.[/quote]

Exactly, both sensors are analog at heart. The A/F sensor (also called broadband or lambada) is able to give the onboard computer a more exact A/F ratio than a standard O2
sensor can. Where a standard O2 can only really say “lean” or “rich” the lambada can give the computer enough info to know exactly HOW lean or rich the exhaust is.

Here is a good link showing the operation of a lambada sensor and how to test one.
http://www.tiepie-automotive.com/en/Measurement_examples/Sensors/Oxygen_Sensor_Broadband

[sjrobinson]

Im gonna look more up on this but I always thought A/F sensor was a general term.

I know the difference between a wideband and narrowband sensor but whats the difference between a broadband sensor and the other two? There are 3 types of o2 sensors?

[Kevin Criswell]

As far as I know there are only the two styles. And yes technically both can be considered A/F sensors.

Narrowband or standard and wideband (also called lambada, broadband and sometimes digital (even though it’s not))

Like many parts, different manufactures like to name them different names.

[sjrobinson]

Ok I see. So does the ECU have a way to convert the signals to digital and would that really be of benefit?

And what makes the o2 signals respond quicker than others?

[Kevin Criswell]

Almost all sensors are at heart analog, even a hall effect sensor starts out analog and a device called a schmitt trigger (located in the sensor) converts the signal into a digital On/Off signal creating the square wave.

An standard O2 is really nothing more than a battery, it outputs a voltage based on how much oxygen is in the exhaust stream compared to outside the exhaust stream. Less exhaust in the stream (rich) will cause the sensor to output a higher voltage. The biggest confusion of O2s is that they output a sine wave, they don’t. The computer reacting to the O2 signal is what creates the familiar sine wave. When the voltage is high (rich) the computer narrows the pulse width of the injectors and the voltage goes lower (lean), than the computer sees the lower voltage and widens the pulse width causing the sensor to output high voltage again. This process repeats as long as the computer is in closed loop.

With a lambada sensor the computer can “see” just how rich or lean the exhaust is and get a better dial in on mixture. When a scope is hooked to a lambada sensor there is much less sweeping from high to low because the computer is better able to stabilize and control the pulse width. A scope hooked to a lambada sensor shows a more “flat” trace on the screen, normally that would be a sign of a problem in fuel control, but with a lambada sensor it is a sign the system is working correctly. In fact a properly working system with a lambada sensor will show very close to 450 millivolt output.

[sjrobinson]

Ok. I see how to read the two types of sensors when scanning them. Ill have to post the graphs I see to show the difference I’m seeing.

Ive seen that wideband sensors are more useful than narrowband. Is there a reason why narrowband sensors are still used?

[Kevin Criswell]

[quote=”sjrobinson” post=75863]Ok. I see how to read the two types of sensors when scanning them. Ill have to post the graphs I see to show the difference I’m seeing.

Ive seen that wideband sensors are more useful than narrowband. Is there a reason why narrowband sensors are still used?[/quote]

yup, price.

Wideband sensors can cost $200-$300 dollars easily, where a narrorband standard sensor is around 30-50 dollars.

[sjrobinson]

I shouldve guessed that one 😛

They are expensive, which Im glad Ive only had to pay for a narrowband sensor. Ill update this tomorrow or friday with screen shots of what im seeing. The spikes seem to tell me that the other cars’ ECU’s are reading their sensors much quicker and able to provide a more accurate AFR.

[Paul]

[quote=”sjrobinson” post=75854]Ok I see. So does the ECU have a way to convert the signals to digital and would that really be of benefit?

And what makes the o2 signals respond quicker than others?[/quote]

Most, if not all, modern day computers, microcontrollers, electronic control systems are completely digital. Analog computers, like the Enigma encryption machine and the device that cracked it, are long gone. All of the inputs on the controller are digital or contain an analog-to-digital converter (ADC) that measures the voltage and converts it to a binary representation of zeroes and ones. The program running on the controller may convert the binary number to decimal number, like an oxygen concentration or voltage signal, and subsequently the program uses that value to perform some action.

I just thought of two “artifacts” that may appear as spiking to you. First, the ODBII scanner might autoscale the graph, so a small change in the signal look large and the noise will appear more dramatic. Also, if the O2 sensor signal varyies between two adjacent ADC steps, the digitized signal will have a step-response over time on the graph.

An O2 sensor’s response time is probably limited by diffusion of gas over the active area of the sensor. With hot flowing exhaust, diffusion is very rapid. I’m guessing it’s not the response time of both sensors that appear to be slow. The time it takes for a portion of exhaust gas to pass between the 2 sensors by way of the catalytic converter probably should not exceed a couple of seconds. It’s possible that there is some lag between the ODBII scanner and the ECU.

[sjrobinson]

I should mention that I have both upstream and downstream sensors logging side by side. The upstream shows spikes on other cars but not the downstream. Which supported what I originally was taught- that downstream is analog and upstream is digital.

So when did cars start having digitizers in the ECUs?

[Author]

Replies