Tell me about this "wideband o2 sensor" thing

I think I want to get one. I'd probably want to be able to install so permanantly in a car but also move it to other cars later. I know little about them.

The current car (Saab t5 ecu) I think has the ability that I can wire it into the ecu and the ecu can datalog it as well as possibly use it instead of the narrowband for fuel trims.

So you need the sensor, the controller, a bung, and maybe a gauge? Don't some cars come with oem widebands? Could I use a junkyard sensor?

And of course, what is the cheap way to get this all done?

My goal is mostly for being able to look at datalogs and optimize fuel at wot. I don't actually care if I can see a gauge or not.

I wouldn't use a junkyard sensor personally.  But some of the controllers do use the same sensors that were OEM in some cars, making the sensors fairly cheap.  

If your car has the ability to take in a WB sensor, it would be there, on board.

Or do you mean that there are extra input channels where you can put any voltage into the ECU, and you can record it?

And what do you really intend to do with the WOT fuel?  Right now, I'm quite sure it's what it is to protect the engine.  How do you think you can change that?  Is there  a real hack that you can turn off the component protection fuel and let it run more of a power fuel?  If that's what you want to do, but don't have that tool- I'd suggest leaving it alone, as the other ways of getting it where you think you want it will likely break something else.

WB sensors started to go directly into cars in the 2006 range, and by 2010, virtually all cars have them as the primary sensor.  You already know that how the sensors are controlled are very different than for normal O2 sensors.

In reply to alfadriver :

I'm using some free software called t5suite that allows me to connect to the stock Saab ecu and program and modify just about any table or constant value. There are others who have definitely connected widebands to the computer and also been able to log that info in t5suite (it also has datalogging capability for all engine sensors). But that is very specific and I will ask about that on the t5suite specific forums, was looking for more general info here. My goal would be to switch to e85 using the base e85 maps others have made for similar Saabs and then optimize my wot maps based on o2 sensor data.

Good to know about widebands on more current cars. Any thoughts on which controllers work with oem type widebands? I'm assuming an oem replacement part is generally cheaper than an aftermarket sensor, but could be wrong.

https://txsuite.org

If you are curious.

rslifkin said:

I wouldn't use a junkyard sensor personally.  But some of the controllers do use the same sensors that were OEM in some cars, making the sensors fairly cheap.  

Yeah, that is sort of what I meant. Though I wouldn't turn down a nearly free sensor from a smashed up late model with low miles in a jy.

Which controllers can use oem type sensors?

There is no relative cheap way to do it. some gauge/sensor/controller combos can be about 200-300 bucks, and if you want to get swank and monitor both cylinder banks by using 2 sensors its about 500. 

and if you have an rx7 they like to eat the widebands like candy, ask me how I know...

I would personally only leave it on for tuning and take it off the rest of the time. 

In reply to Robbie :

https://wbo2.com/

In reply to Robbie :

So you know, there's a right way to tune the WOT fuel, and there's a hack way.  The right way will be more robust and make more power.  The hack way might make more power, but might also break something else.

I've seen too many hackers blow their motor up to not warn you of that.  And some recent help for someone else, and the suggestions they are getting also confirms that.

alfadriver said:

In reply to Robbie :

So you know, there's a right way to tune the WOT fuel, and there's a hack way.  The right way will be more robust and make more power.  The hack way might make more power, but might also break something else.

I've seen too many hackers blow their motor up to not warn you of that.  And some recent help for someone else, and the suggestions they are getting also confirms that.

So, what is the right way? I was going to do a few runs and look at the logs, then update the tables to shoot for an Afr value or range.

Then, update the knock table to have a little more fuel in every square than the normal table. Seems simple just time consuming to me but am i missing something?

Hey Robbie - What year/computer is your Saab?  

General info about O2 sensors that I've gleamed.  Someone please correct me if I'm off the line anywhere through here.

A Narrow (single wire) get enough electricity out of the magical/exotic metals it's imbued with reacting with the oxygen in the exhaust stream that it can output a voltage range of something like .5-.8 volts, which gives it enough resolution to know get a very limited scale that reads something like "really rich, sorta rich (14.2-ish), PERFECT STOICH!, sorta lean (15-ish), really lean."   Conditions have to be perfect for this to work, since it's relying on the metallurgical differences to generate the voltage, which is why most cars with narrow bands were running various fixed tables for different conditions and only "actively tuning" when running in that perfect range.

The wideband (four wire) O2 is a much more sophisticated piece of equipment which has a heater that keeps it a the perfect temperature, and then runs a reference voltage of 5V into it, and the differences in the oxygen levels affects the voltage out which is what it reads.   Since the range is now 0-5V, under controlled conditions, that gives the wideband something like 600% more resolution than the NB.  But the downside is that you have to have a controller to deal with making sure that it's receiving 12V solid power in for the heater, 5V in for the reference line and scaling things appropriately if either those those are off (your car runs at 13.8V instead of 14.4V, etc.).   OEM computers have this built into them, standalones do not. 

Most standalones are looking to simply use the 0-5V, and they rely on WB controller to take care of all of the ancillary stuff around it. 

----------------------------

That being said, depending on your computer, it could either actively use a wideband and let you just put in a target AFR/Lambda table, and it will try to adjust fuel to hit it, or it could be 100% manual where now you can log it, but you have to manually override the values yourself.  Some of the pictures on the https://txsuite.org site you linked to show target lambda tables, but I'm not sure which series you're dealing with.

I know in Megasquirt & AEM land, once you have a wideband, you just set your AFR/Lambda table and start driving.  It'll dial in your table quite nicely over time.

In reply to WonkoTheSane :

It's a 1999 Saab 9-3. The o2 sensor is a 4 wire, but I didn't realize that was a wideband?

In t5 suite I do get an "afr" reading and it seems decently detailed, but I thought it may have been a programatic guestimate based on measured air and fuel going in, or a combination of that plus o2 sensor feedback. I was assuming this because iI thought the stock sensor was a narrowband. If it isn't, then maybe I don't have to do anything?

Also, t5suite does have an autotune feature, I will check it out how you set the targets.

I have an AEM wideband meter.  I use their O2 sensor instead of the stock one since the meter has an output that mimics the stock O2 inputs to the ECU.

There are narrow-band 4-wire sensors (signal, ground, two for heater; some don't have heat, some don't have dedicated ground, hence 1, 2, and 3 wire sensors)

Robbie said:

alfadriver said:

In reply to Robbie :

So you know, there's a right way to tune the WOT fuel, and there's a hack way.  The right way will be more robust and make more power.  The hack way might make more power, but might also break something else.

I've seen too many hackers blow their motor up to not warn you of that.  And some recent help for someone else, and the suggestions they are getting also confirms that.

So, what is the right way? I was going to do a few runs and look at the logs, then update the tables to shoot for an Afr value or range.

Then, update the knock table to have a little more fuel in every square than the normal table. Seems simple just time consuming to me but am i missing something?

For a vehicle of your vintage, I would be shocked if there was a simple table that had a target a/f in it.  By the late 90's, pretty much all enrichments were one of two things- power enrichment for WOP and component protection to keep the exhaust from degrading.  The WOP/WOT table is should just be a line of lambad/phi vs. engine speed, and that's it.  But 99% of all enrichments will be component protection.

If you are not messing with either of those, what are you messing with?

Another recommendation to pony up and buy a real wideband and controller, if for no other reason than you know it's accurate. It only takes one lean pull in boost to blow an engine.

For what it's worth, my 05 Subaru came with a narrowband sensor and the ECU "estimated" what the AFR should be based on MAF readings and fuel load. So the logs would say 12.0:1 even though the sensor had no resolution in that area.

Also, most aftermarket systems these days are a BOSCH 4.9LSU wideband sensor, and have the ability to communicate with various data logging systems.

As I understand it there is a volumetric efficiency map (rpm and map - it's turbocharged), and some correction factors for coolant and system voltage and intake air temp.

Then, there is a fuel injection map (again rpm and map) where you can adjust the fuel injection factor in each square. Finally there is an overall injection factor "constant". The fuel injection factors in the table have to be between .5 and 1.5 so if you are maxing out the injection factor you need to up the constant value. All of these things are multiplied together essentially to get the fuel. I would plan to update the fuel injection factor map in the cells where I hit for wot.

There is a copy of the same fuel injection factor table but it is used only when knock condition is present. That is the knock map (in general just make sure it adds more fuel than the normal map). There are also acceleration and deceleration enrichment/enleanment tables. I wouldn't plan to update those.

There are boost request maps for rpm and throttle position, and there are spark advance maps for rpm and map.

I guess it was a pretty advanced system in it's time. 

morello159 said:

Another recommendation to pony up and buy a real wideband and controller, if for no other reason than you know it's accurate. It only takes one lean pull in boost to blow an engine.

For what it's worth, my 05 Subaru came with a narrowband sensor and the ECU "estimated" what the AFR should be based on MAF readings and fuel load. So the logs would say 12.0:1 even though the sensor had no resolution in that area.

Also, most aftermarket systems these days are a BOSCH 4.9LSU wideband sensor, and have the ability to communicate with various data logging systems.

 

Yeah, that's a fair point about buying a good one to trust. I'm sure my software is doing the exact same thing estimating an afr (but that is t5suite, not the car's ecu).

In reply to Robbie :

What year is it again?  

Reading what you are doing- you are changing more global items, when the better change is the specific reason of who is deciding how rich to run.  What you are doing is taking a 10:1 commanded a/f and delivering 12:1, which then changes anything that uses the same load or fuel pulse width someplace else.  What you need to do is change the 10:1 command to 12:1 (which is actually easier than trying to tweak the air calculation and then the fuel calculation).  Not adjust factors and whatnot.

Ransom said:

There are narrow-band 4-wire sensors (signal, ground, two for heater; some don't have heat, some don't have dedicated ground, hence 1, 2, and 3 wire sensors)

Thanks for the reminder of that.. I forgot that they started heating narrowbands.  I sort of jumped right from the single wire to the widebands in my cars :)

Robbie - From what I can tell, you're running a narrowband with a heater.. It must have a dedicated ground if you're seeing 4 wires (Bosch 15738?), but according to Advance's site, the NTK that crosses with an "OE connector" shows only 3 wires, so it uses the exhaust itself for the ground.

I agree with alfa's assessment, you need to figure out how to edit your commanded table after you have an accurate measurement..

1999 Saab 9-3. I looked at the tuning software again today. Looks like you can wire a wideband into one of three pins on the ecu and then the tuning software (not the ecu) can log the wideband signal (like an unused egr pin, the post cat o2 signal if you aren't running a cat, or one other option).

For "afr targets" I'm pretty sure there aren't any (in the ecu) because the ecu doesn't have a wideband to measure them. The ecu uses the ve map as a starting point and can adapt fuel based on narrowband and knock sensor feedback. There are adaptation tables that store adaptation values over time. So I think to manually tune fuel you have to play with the ve maps. The ecu does have boost request maps and it measures map directly so it is also constantly using that to regulate boost. But the fuel comes from the ve map and any adaptations.

However, if you log the wideband data into the tuning software, then the software's autotune feature can be setup with an afr target map and then as you drive around it will give you suggestions to update your ve maps to get closer to your target afr. 

Sounds like I just need to go get a wideband sensor and gauge/controller and wire the output into one of the perscribed pins on the ecu.

In reply to Robbie :

About the a/f target- you don't need to have a WB to have a target a/f.  You'd be surprised how good the air measurement is so that if you are very sure what 14.6:1 is over a wide range of air flow, when you go WOT, you can ask for a specific a/f, and you will get it.  And it will be really close to being right without feedback.  All those adaptive tables you see are just for that.  And the time that the car is running from start without the O2 sensor.

In 1999, I'm sure that Bosch did that (which I suspect that your controller is).  We did.  Heck, we had been doing that since I started working in 1992.  

And being a post OBDII car, you are getting your open loop fuel request from WOT power request or component enrichment.   That's pretty much all that is legally allowed then.