Wideband Sensor Life Discussion

Discuss MegaSquirt, VEMS and other non-free hardware and software here.
fuzzysig
QFP80 - Contributor
Posts: 31
Joined: Thu Mar 03, 2016 3:06 am

Wideband Sensor Life Discussion

Post by fuzzysig »

ok so why does one differ from other. in terms of sensor control?
doesn't company test it under less than ideal conditions before releasing it on the market?
I mean innovate seems to need near ideal conditions for sensor to last any reasonable time. they even say that replacing sensors every year is normal with mtx-L I called them twice and got that response from them that its normal to have to replace sensors every year its a "wear item"

is it the chip used or power supply or software that makes the difference?


can the circuit be damaged by poor installation for example if my car has corroded battery terminal and bad chasis ground. (not that it does, but I had that happen before after washing the engine bay)

how does a 4.2 lsu last so long in a vw bettle for example(which has far less than ideal conditions) but cant last for more than few years with aftermarket wideband controller?
User avatar
DelSolid
QFP80 - Contributor
Posts: 59
Joined: Wed Jan 27, 2016 10:41 pm
Location: Los Angeles

Re: Widebands That You Would or Wouldn't Buy

Post by DelSolid »

fuzzysig wrote:ok so why does one differ from other. in terms of sensor control?
doesn't company test it under less than ideal conditions before releasing it on the market?
I mean innovate seems to need near ideal conditions for sensor to last any reasonable time. they even say that replacing sensors every year is normal with mtx-L I called them twice and got that response from them that its normal to have to replace sensors every year its a "wear item"

is it the chip used or power supply or software that makes the difference?

can the circuit be damaged by poor installation for example if my car has corroded battery terminal and bad chasis ground. (not that it does, but I had that happen before after washing the engine bay)

how does a 4.2 lsu last so long in a vw bettle for example(which has far less than ideal conditions) but cant last for more than few years with aftermarket wideband controller?
This is a really big question and I wont attempt to answer more than a little bit of it here. But there are quite a few things that affect the life of an O2 sensor, regardless of the controller. I'll list a few of them in no particular order

Lead content in the fuel. Bosch derates the life of an O2 sensor by as much as 80% when using leaded fuels, depending on the concentration.

Running more than ~1 AFR point richer than stoich The O2 sensor can become contaminated with oil and carbon just like your spark plugs. A good way to think of this is the factory spark plugs in a new VW Beetle easily go 100,000 miles, as do the O2 sensors. But if your car eats spark plugs every 10,000 miles it's going to eat O2 sensors almost as fast, and for many of the same reasons. This is why the O2 sensor is considered a service item by many people.

Sensor Warm Up time. The sensors are very susceptible to thermal shock and trying to slam on the sensor and get it running faster than recommended will kill your sensor much faster.

Condensation in the exhaust. If the tip of the sensor or the exhaust upstream of the sensor has condensation, the heater must either be kept off or run in a reduced power mode until it is gone. OEM's literally model the exhaust condensation and have tables that drive howe long tjhis process takes under the current startup condition. They can do this because they know the exact characteristics of the engine/vehicle/exhaust system it is being used in. Aftermarket units have to estimate the time for this phase or in the worst case, ignore it altogether.

Excessive EGT - General. The OEM's know what the exhaust temp is, either by measuring it (old way) or modeling it (the new way) and they cut the heater power in response to very high heat operation or even in anticipation of an upcoming high heat condition. This can be done by an ECU based AFR controller because it is the unit controlling the thing making the heat so it knows what is going to happen in a few seconds and can act to prevent it, but not a generic add on aftermarket controller. Those can only respond to the current heat at the tip.

Poor sensor location causing excessively high EGT. The sensor doesn't have a cooler, just a heater. So mounting it in a location where the EGT will exceed the maximum allowable temp is going to reduce the life of the sensor and there is nothing the controller can do about it. The OEM's model and then validate the EGT's at the sensor location on every car they design so they have this under control. BUT, when Joe hot-rodder welds in his own bung, how hot is it going to get there? I don't know, and in all likelihood neither does Joe. And if you put a sensor in the OEM location you can only be assured of the EGT being in the allowable range if you don't increase the power output of the engine. And who really does that? I sure don't. I boost the shit out of my car (35 PSI, GT47).

These are just a few off the top of my head. I'm sure you guys can throw up a few more.
1969 Plymouth Satellite Wagon with a 440 & TF727
1929 Ford Roadster with a 2JZ and a T400, GT47, 1,100WHP, 240+ MPH
1930 Ford Roadster with a 42 Merc Flathead with triple Holley 94's. Major work in progress
I work for AEM but am not here schilling for them. Nothing I say is official.
User avatar
DelSolid
QFP80 - Contributor
Posts: 59
Joined: Wed Jan 27, 2016 10:41 pm
Location: Los Angeles

Re: Widebands That You Would or Wouldn't Buy

Post by DelSolid »

DelSolid wrote:Bosch derates the life of an O2 sensor by as much as 80% when using leaded fuels, depending on the concentration.
More info:

It's widely known that the service life of an O2 sensor will be shortened based on the lead content of the fuel but the Bosch datasheet for the LSU 4.2 gives values of the de-rating and they are interesting.

Bosch considers the Engineering Design Life of the LSU 4.2 to be 100,000 Miles (160,000 KM) and they consider the service life to be 50,000 Miles (80,000 KM) and they de-rate the sensors based on the lead content of the fuel as follows;
0.0 Grams of lead/gallon = 50,000 Miles
0.6 Grams of lead/gallon = 37,000 Miles (26% less)
1.5 Grams of lead/gallon = 19,000 Miles (62% Less)
2.3 Grams of lead/gallon = 12,000 Miles (76% less)

But that is assuming everything else is optimum, no oil burning, always running close to lambda 1.0, within temp spec, no silicone in the exhaust, etc...

Now, here are some leaded racing fuels and their lead content* (grams/gallon)
3.0 grams for CITGO 110
4.0 grams for PHILLIPS B32
4.0 grams for PHILLIPS B33
4.0 grams for SHELL SRS 110
4.0 grams for SHELL SRS 114
4.0 grams for TRICK 112
4.0 grams for TRICK TURBO 119
4.1 grams for UNOCAL-76 110 STOCK
4.2 grams for VP RACING C-14
4.2 grams for VP RACING C-15
4.2 grams for VP RACING RED
4.2 grams for VP RACING C-11
4.2 grams for VP RACING C-12
4.2 grams for VP RACING C-18
5.0 grams for SUNOCO N.O.S.
6.0 grams for PHILLIPS B37
6.0 grams for SHELL SRS 118
6.0 grams for SUNOCO MAXIMAL #5
6.0 grams for VP RACING C14+
6.0 grams for VP RACING C-16
6.0 grams for VP RACING C-19
6.0 grams for VP RACING C-21
6.0 grams for VP RACING C-23
8.0 grams for VP RACING C-25

* I got the lead figures from: http://www.smithtex.com/racing/fuelcomp.html

Literally every leaded racing fuel listed is off-the-charts on Bosch's service life de-rating curve. :lol2:
1969 Plymouth Satellite Wagon with a 440 & TF727
1929 Ford Roadster with a 2JZ and a T400, GT47, 1,100WHP, 240+ MPH
1930 Ford Roadster with a 42 Merc Flathead with triple Holley 94's. Major work in progress
I work for AEM but am not here schilling for them. Nothing I say is official.
fuzzysig
QFP80 - Contributor
Posts: 31
Joined: Thu Mar 03, 2016 3:06 am

Re: Widebands That You Would or Wouldn't Buy

Post by fuzzysig »

so its a good idea to get a heatsink for the sensor regardless of location to stabilize the temps better?

does aem failsafe controller use rpm and TPS feedback to also control the sensor or they are only used for failsafe feature?

and also the new PLX dm6 says it has a soft start feature for cold starts and it shows sensor health and reaction time(I'm guessing the reaction time would show if the sensor is getting contaminated right?)

just a thought but maybe it would be a good idea to make link between a wideband controller and standalone or RTP so the wideband controller gets a feedback from RTP or standalone.
since even hondata and Neptune have outputs that are programmable. that might help keep the wideband in better shape than not having any feedback at all

so the lsu 4.9 has a built in temp sensor?

and one last question I hope its on topic
if the sensor does get contaminated by oil is it better to lean it out slightly above 14.7 to burn it off or to richen the mixture slightly to clean off the sensor while its heated?
I mean how do dyno shops keep their sensors clean after being used in so many different cars with unknown basemaps
they see more abuse than mine no question.
you got Hondas that spit oil in vtec all the time
turbo cars that run very rich
and untuned cars on cold start down to 10afr...

I asked around the local dyno shops and a lot of them don't use mtx-l and lc2 they only sell it to customers
but use plx and aem and the LM2 units
User avatar
DelSolid
QFP80 - Contributor
Posts: 59
Joined: Wed Jan 27, 2016 10:41 pm
Location: Los Angeles

Re: Widebands That You Would or Wouldn't Buy

Post by DelSolid »

fuzzysig wrote:so its a good idea to get a heatsink for the sensor regardless of location to stabilize the temps better?
It wont hurt. IMO you are much more likely to damage a sensor due to excessive heat (common) than not enough (rare) so may as well move in that direction.
fuzzysig wrote:does aem failsafe controller use rpm and TPS feedback to also control the sensor or they are only used for failsafe feature?
Only failsafe since they are not required to be connected to run the gauge. So the gauge would not know if they are not hooked up or if the engine is not running.
fuzzysig wrote:the new PLX dm6 says it has a soft start feature for cold starts and it shows sensor health and reaction time(I'm guessing the reaction time would show if the sensor is getting contaminated right?)
Don't know what "soft start technology" means or if that applies to the heater. Either you follow the Bosch spec for the heater cold start or you don't. I don't know how they define the measured reaction time either.
fuzzysig wrote:just a thought but maybe it would be a good idea to make link between a wideband controller and standalone or RTP so the wideband controller gets a feedback from RTP or standalone. since even hondata and Neptune have outputs that are programmable. that might help keep the wideband in better shape than not having any feedback at all
so the lsu 4.9 has a built in temp sensor?
Sort of. The temperature of the sensor element is determined by measuring the internal resistance of the sensor’s Nernst cell and that has a temp target you need to maintain. The EGT dramatically influences this temp and one of the controllers jobs is the measure the internal temp frequently and then alter the heater power to get it back to target. Knowing the EGT, especially an upcoming change in EGT or exhaust flow allows the controller to do a much better job of maintaining the optimum temp since you dont have to wait for an error to occur, you can act preemptively. This is the type of thing an ECU based controller can do but a stand alone unit cant. They have to just use a PID system (normally) and react to errors as they present themselves.
fuzzysig wrote:if the sensor does get contaminated by oil is it better to lean it out slightly above 14.7 to burn it off or to richen the mixture slightly to clean off the sensor while its heated? I mean how do dyno shops keep their sensors clean after being used in so many different cars with unknown basemaps. they see more abuse than mine no question. you got Hondas that spit oil in vtec all the time turbo cars that run very rich and untuned cars on cold start down to 10afr...
You really are not going to have much success burning it off. The sensor temp will pretty much always be the same due to the internal heating element and if you go above that to try to burn it off you will likely just do more damage. I have seen some people try to clean them with an aggressive application of brake clean but good results are rare and not long lasting. Most dyno sensors don't last longer, they just seem to since they are used for such a small amount of time. 5000 miles is like 500 hours of drive time and most dynos wont see that in an entire year year.
1969 Plymouth Satellite Wagon with a 440 & TF727
1929 Ford Roadster with a 2JZ and a T400, GT47, 1,100WHP, 240+ MPH
1930 Ford Roadster with a 42 Merc Flathead with triple Holley 94's. Major work in progress
I work for AEM but am not here schilling for them. Nothing I say is official.
HelmutVonAutobahn
LQFP112 - Up with the play
Posts: 109
Joined: Fri Nov 29, 2013 12:10 am

Re: Wideband Sensor Life Discussion

Post by HelmutVonAutobahn »

It wont hurt. IMO you are much more likely to damage a sensor due to excessive heat (common) than not enough (rare) so may as well move in that direction.
Strangely enough, colder is probably worse. The sensor is supposed t be regulated to 780C. It won't take any permanent damage until above 1000C, at constant temp. But, rapid changes in temperature will kill it quickly. So, if you have EGT spikes that force the sensor up to 900C, you get swings of +/-120C ,if you are running the sensor at 780C ( 900-780=120) . But... if you run the sensor at say 650C, the swings will be more than twice as large, at +/-250C ( 900-650=250).

If you ran the sensor at 900C, in theory, it would not swing at all ( assuming perfect control ). But, it would not be accurate. And, may age a bit more quickly.

The heat sinks tend to "round off" the EGT spikes, thus, reducing the temperature swings and the resultant thermal stresses. Which is very helpful, if you happen to be running the sensor a "few" degC colder than spec. Or, not keeping up with the thermal regulation update rate.
User avatar
DelSolid
QFP80 - Contributor
Posts: 59
Joined: Wed Jan 27, 2016 10:41 pm
Location: Los Angeles

Re: Wideband Sensor Life Discussion

Post by DelSolid »

True, But I am assuming that the sensors heater is working properly and the target temp is being maintained by the controller. If that's not true then all sorts of problems are going to crop up.

My statement that it is better to err on the cooling side is based in the fact that the controller can add heat to the sensor but it cant add "cold" to the sensor, It can only reduce the heat added. But your point is correct and taken.
1969 Plymouth Satellite Wagon with a 440 & TF727
1929 Ford Roadster with a 2JZ and a T400, GT47, 1,100WHP, 240+ MPH
1930 Ford Roadster with a 42 Merc Flathead with triple Holley 94's. Major work in progress
I work for AEM but am not here schilling for them. Nothing I say is official.
fuzzysig
QFP80 - Contributor
Posts: 31
Joined: Thu Mar 03, 2016 3:06 am

Re: Wideband Sensor Life Discussion

Post by fuzzysig »

How much temp change are we talking about in typical dyno scenarios

For example an untuned honda with a sensor in stock location in downpipe around 36 inches from the exhaust port.

Cold start untuned can be anywhere from 10-15afr depending how crapy the basemap is

During first runs it can fluctuate anywhere between 11-16 or more afr
And then gets adjusted

Thats a pretty wide swing of temps On every pull until the whole fuel table is mapped

How can a controller keep the sensor temp if its reactinf to sensor temp?
And how fast does an element heat up anyway?
Im also assuming that a bad ground and low voltage slows down the heater speed making the temp control react slower correct?

Im just trying to understand how it works so i can figure out why so many sensors went out with mtx-l
HelmutVonAutobahn
LQFP112 - Up with the play
Posts: 109
Joined: Fri Nov 29, 2013 12:10 am

Re: Wideband Sensor Life Discussion

Post by HelmutVonAutobahn »

Yes. The best it can do is add zero heat. Assuming it is at least smart enough to do that, when the sensor is over 780C, you are still better off erring on the hot side. So, if EGT pushes the sensor to 900C , while the heater is completely off, you are better off catching it, on the way down at 850C than 650C , as the temperature sweep will be 50C vs 250C.
HelmutVonAutobahn
LQFP112 - Up with the play
Posts: 109
Joined: Fri Nov 29, 2013 12:10 am

Re: Wideband Sensor Life Discussion

Post by HelmutVonAutobahn »

fuzzysig: You seem to be confusing AFR and EGT.

AFR ( lambda ) is the ratio of air to fuel mass.
EGT is the exhaust gas temperature.

AFR can affect EGT. But, other things affect it more.

The sensor temperature is a function of EGT and the mass of the exhaust flow.

Yes, the EGT and flow swing LOTS during a pull. A closed loop PID function is used to control the sensor's heater, to keep the sensor temperature as close to spec as possible.

Some controllers do this better than others. When it is done poorly, sensors are more likely to be damaged from thermal shock. Which is a rapid temperature change to a brittle object ( the sensor element is made of ceramic ). Something like throwing cold water on hot glass.
Post Reply