Stock ECU Native Wideband Review

[->PrOjEcTGS<-]

I ran this several years ago with a LC-1 and XD-16 gauge. My gauge was always spot on but it's also digital not analog. I first tried out the NB simulator and hated it termensily. I was using a Ostrich so I just changed my output on the LC-1 and the code on the eprom and restarted the car. Was pretty decent OTB but I did change my sweep points. You don't want them to narrow. If it's to narrow it will constantly overcorrect and make your afr swings worse. You can't expect a rock solid afr at idle. Cruise should be within a .8 AFR swing. I also had my LC1 located just under the firewall and mounted the controller on the firewall.

And to the OP. Don't try something find it doesn't work and make a off the hip judgement call and then write a "review" saying it shouldn't be used for whatever you are tring to use it for. The EPROM ECU in the 1G happens to use the same processor at the MS1's do. It's pretty robust and can handle a decent amount of instructions before really slowing down. Oh I also was running double width maps, stutter box, NLTS and a few other things that I did on my own.

Also something to note. Your large AFR swings can also be caused by having your injector deadtime off by a decent amount. The larger the injector the more you'll notice this. And my stock NB o2 had a afr swing of like 13-16.5 at idle. I got it to 14.2-15.1 at idle running 1000cc injectors with .8ms of ipw. I couldn't get it any better cause any lower the injector just wouldn't be squirting any fuel. You generally want at least 1.5ms of ipw at idle for a decent idle fyi which you will not see with large injectors.

The position of the O2 really isn't going to make a difference that you could even feel and unless it's past the firewall you probably wont even see it on a log.

I understand what you're trying to say here. That's why I'm giving it a second shot with the LC-1 and a little bit more programming in the code.

We'll see what it does. Right now with the LC-1 Simulating a NB it idles at around 14.5 - 15.0. That's acceptable to me. If I can't get the code to work for the WB to hold it at least the same as the NB, I would consider it unacceptable. I'm just here giving my findings and my impressions on the setup so far. If I get it to work I will edit my original post and add a paragraph as to what it took to make it function properly.

Large AFR swings

 

[Maglin]

btw it's been a few years since I've messed with the EPROM ecu but I think you actually need to widen the set points for closed loop. I think it's actually over correcting now if you don't change them. You are going from a 0-1v signal to a 0-5v singal and thus would also want to move the "swing" points but not proportinally. Just something to think about. I might have an old bin laying around. If you have issues I'll see if I can't find it and send it your way. You'll have to see what I did cause I don't remember everything. I just remember I did have to do some tweaking that wasn't mentioned in the dsm-ecu group.

 

[->PrOjEcTGS<-]

btw it's been a few years since I've messed with the EPROM ecu but I think you actually need to widen the set points for closed loop. I think it's actually over correcting now if you don't change them. You are going from a 0-1v signal to a 0-5v singal and thus would also want to move the "swing" points but not proportinally. Just something to think about. I might have an old bin laying around. If you have issues I'll see if I can't find it and send it your way. You'll have to see what I did cause I don't remember everything. I just remember I did have to do some tweaking that wasn't mentioned in the dsm-ecu group.

I know what you mean by not mentioned on the DSM-ECU group. That's a great resource for information and all. But the search feature sucks and just some basic information I had to scrounge around to find. Leaves a lot left to be desired. But hey, it's free information that has helped my car's driveability out tremendously.

I burned a chip last night and plugged it in. I used the original code that I had previously with the UEGO. It doesn't cycle near as much as the UEGO did. But it holds a richer mixture at around 14.0-14.7 average of about 14.5. I think the reason it doesn't cycle as much has to do with how much faster the LC-1 is than the AEM UEGO. It can respond quicker and relay that info back to the ECU for it to make the adjustments.

I do need some help on something. Since the LC-1 is programmable. I went to test the differences in output voltages by inputting 2.5v in both fields of the LM programmer so that both outputs would give me 2.5v all the time. The gauge read 14.6 while my logger read 2.6v. Now, to get the ECU to read 2.5v when the gauge gets 2.5v how should I change the LC-1's Output programming?

I'm guessing this is why my car runs richer since it's getting a higher voltage making the ECU think it's leaner that it really is.

 

[->PrOjEcTGS<-]

Ok so this is what I did. Since the ECU reads .10 higher voltage than the gauge. I found this out by putting 2.5v in both fields for analog 2 and putting 2.4 in both fields for analog 1. And when I logged it I got 2.50v at the ECU.


Using this formula: V * 3.008 + 7.35= AFR
I came up with this: .10v * 3.008 + 7.35 = 7.6508 afr

7.65 needed to be my new 0.00v and 4.90v at 22.39afr

I changed the output of the LC-1 to do 0.00v @ 7.65AFR and 4.9v @ 22.39AFR. That dropped the whole scale down by .10v. Now Lamda is at 2.35v for the analog output to the ECU and 2.45v for the gauge. But now they should both read the same. Woot!!

When I go home tonight from work I'll see how that helps. Maybe that will lean it out a little bit. Then I'll start messing with the ECU's code.

 

[bling5tatus]

aaaaaaa-men to that bro!
According to Innovate tech support, the only reason for the short wire from LC1 to sensor is fear of electrical interference. I think I'll call them back and ask them if they can do a custom one for me with a long wire with better shielding on the wire or ferrite beads or something.

Where did you put the sensor through the firewall? or floor? Did you make a new hole?

Gary

People may disagree with this, but I went right through the steering boot. I am FWD so my down pipe was almost directly below that anyway... if you're awd you may want to do it a little differently.

 

[->PrOjEcTGS<-]

Well, I ended up changing the switchpoints from 122 and 128 to 127 and 129. At idle it really didn't make much of a difference with the cycling. But during cruise it hovers around 14.5 - 15.0 which is acceptable to me. I first tried to widen out the switch points with 118 and 132 that seemed to take the cycling away but the AFR would be anywhere from 15.5 - 13.9. It would start out lean at 15.5 then richen up past 14.7 then come back to 14.7 and hold it steady but it took a long time for that to happen. So, I don't know which one I like better.

 

[Maglin]

you can't put the LC-1 in the stock location. After a full WOT run you will get an overheat error. Most controllers don't have an overheat function and will just give you the bad data coming from the overheated O2 sensor. That is why you are instructed to mount it further down the exhaust pipe.

 

[toofast82]

i have heard some people say they have had success in the stock location. but since the heat is a concern that close, has anybody tried one of these?
eBay Motors: WideBand Air Fuel Ratio oxygen sensor bung heatsink (item 330306879882 end time Feb-16-09 13:31:17 PST)

 

[->PrOjEcTGS<-]

Well, I just got done installing my Devils Own Stage 1 kit. So, I'll be tuning that out tonight and tomorrow. I also plan on messing around with the switch points some more trying out a few different settings and seeing how I like it.

 

[We're on Boost]

has anybody tried one of these?

Innovate sells those. They look to me like they would make a lot of difference. In addition to throwing heat off they also pull the sensor tip out of the pipe a bit, like a 1" long bung would do.

Gary

 

[PIMking]

Well, I just got done installing my Devils Own Stage 1 kit. So, I'll be tuning that out tonight and tomorrow. I also plan on messing around with the switch points some more trying out a few different settings and seeing how I like it.

So how many widebands do you have?

Im kinda confused because im getting a wideband but im not sure what I should do.

Put the LC1 in the stock location but what is going to send the data to the ecu?

put it in a different location kinda like in your picture and just have the gauge in the gauge pod.

Im just confused....

 

[toofast82]

that stock narrowband position right next to the turbo. Put the wideband there.

I think hes confused about how the ecu is going to act when you replace it with the wideband without changing anything. The answer is in the first post where he gave the hex editing to set the parameters for wideband in the ecu. He maybe doesnt do his own eprom editing and might not understand what was said. Maybe he uses dsmlink

 

[architechnik]

I've been running a native wideband setup for a few days now. I wanted to test it out as my Ostrich is on it's way and the NB simulation is killing me. I was having sweeps on the simulated output from around 13.3 - 16.3.

Since I did the change I now have sweeps from 14.3-14.8 during idle and cruise. Now, I did my own adjustments for setting the points, and set the store points a little further to keep it from trying to adjust the trims. My sensor is at the bottom of the bend in the DP and gets pretty consistent readings.

I hadn't noticed this thread until today, but I did want to add some information about the range of various gauges and sensors. It's important to note that the Bosch LSU 4.2, the standard sensor used for most systems has a range from just normal air to 11:1. If a wideband controller uses any of the standard sensors and provides a display beyond 11:1, then it is only mimicking that ratio with no real accuracy. The technology used in a single sensor does not have the ability to provide such a large range to detect chemical changes in such a hostile environment. The greater the range provided, the less accurate the device becomes, as anyone who has done lab work can attest.

Likewise, the readout provided by a wideband oxygen sensor and associated controller is dependent on the calibration done from the manufacturing or from an installed calibration. Some typical causes of changes in calibration include pressure, temperature, wear, poor fuel, different fuel, moisture, and electrical current (both interference and supply voltages). A plug and play setup such as AEM unit will always be limited. The most accurate option is to use a setup that allows periodic manual calibration that can be compared with a high precision counterpart in the primary operating range. Of course this is more work and more cost, and not always the best option.

Here's some additional info you guys can check:
http://www.boschautoparts.com/Resources/FAQS/OxygenSensors.htm
http://wbo2.com/lsu/oxygen13-17.pdf
Bosch LSU 4 Information

 

[->PrOjEcTGS<-]

So how many widebands do you have?

Im kinda confused because im getting a wideband but im not sure what I should do.

Put the LC1 in the stock location but what is going to send the data to the ecu?

put it in a different location kinda like in your picture and just have the gauge in the gauge pod.

Im just confused....

I have ONE (1) WBO2 sensor. I had the AEM unit. But, I sold it for Innovate's LC-1 unit. Therefore, still only having one sensor. I mounted mine as seen in my previous picture, mounted the unit to the firewall and ran the wires through a hold in the firewall to my ECU. Analog 1 from the LC-1 unit goes to the front O2 sensor wire of my ECU. I then used the code from my first post to utilize the 0-5v input.

Technically, the stock NB sensor is useless now in my case and is just chilling in the o2 housing. I will more than likely take it out and put the plug that came in the kit in there to block it off. I talked with Jeff about how the ECU will respond to not seeing a heater circuit to the sensor. He said it will not be a problem and that you can just unplug the stock sensor since our ECU does not look for a heater circuit. Therefore, you will not get a CEL!

If you don't burn your own chips and edit the code of the ECU then you will not be able to use this feature. Since that is just one of many things DSMlink Cannot do!

 

[twdorris]

If you don't burn your own chips and edit the code of the ECU then you will not be able to use this feature. Since that is just one of many things DSMlink Cannot do!

Which feature are you referring to here? We certainly support native WBO2 simulation, if that's what you mean.

v3narrowbandsim [ECMTuning - wiki]

EDIT: FWIW, I've been running my car on this for the last 4 years or so and passing emissions each time they actually bother testing my car. The WBO2 is installed in the rear O2 bung.

Thomas Dorris
ECMTuning, Inc.

 

[->PrOjEcTGS<-]

Yes, Simulation. Not the real thing. The ECU is still seeing 0-1v. While with the code I used it is utilizing the actual 0-5v output. Internally the code is adjusted so that higher voltage is now lean and lower voltage is now rich. No simulation.

With your setup it's not the Wideband that's simulating a 0-1v output it's the ECU that's taking that 0-5v and simulating the 0-1v. Totally negating the benefits of the broader 0-5v output. How is that any different/better then having the Wideband unit itself do the simulation? Please Explain...

 

[twdorris]

Totally negating the benefits of the broader 0-5v output.

It's simply a different approach to the same problem. The ECU is ultimately just switching between moving the STFT in a positive direction versus a negative direction. Unless you actually attempt to alter the rate at which it updates the STFT based on the difference between the current WB value and the target value, the two approaches are simply different. We considered doing that too, but in the end a simple switch point works perfectly well.

You then tweak the rate at which the STFT is adjusted based on how far "away" the WB is installed from the engine and you have a perfectly operational solution. My car runs just as well with the O2 sensor installed in the rear O2 bung as it does in the front.

Thomas Dorris
ECMTuning, Inc.

 

[->PrOjEcTGS<-]

Project are you affraid of running your wideband in the stock location, just to verify the results ?

Atleast try it before you keep editing more code. We used 122 to 128 with great results at the stock location.

I don't have a problem running the wideband to the stock location. But, I already have it in the downpipe and don't plan to move it. With the new LC-1 I'm not seeing the Swings like I was with the AEM unit. I am not going to extend my wires just so I can move the position of it. Plus I'm limited on space up there anyway and would like to get rid of the stock sensor. I'm running the 2g O2 housing which places the sensor further down than the 1g O2 housing and the stock sensor is hitting my LICP to my FMIC. I'm very content with the position I have it in. I get sweeps from 14.5-15.0 at idle which is a lot better that 13.3 - 15.9 from the AEM unit. During cruise it's better and hovers right around 14.7 to 14.8. Very good.

Thomas, I agree with you it's just two totally different approaches.

 

[twdorris]

Thomas, I agree with you it's just two totally different approaches.

It's not totally different, though. That's my point. It's still a lean/rich switch. It's on/off. That's effectively a simulation of narrowband operation. You're just not stuffing a narrowband voltage in somewhere. But you are making a binary rich/lean determination. Which is just simulating narrowband operation.

It's not taking anymore advantage of the wideband's range than a single switch point. The only thing having two switch points is doing for you is working around the fact that you're not adjusting the STFT update rate. In our approach, you pick a single switch point and then adjust the STFT update rate depending on where the sensor is installed. But in the end, the results are the same. It's still a binary switch based on some WB voltage comparison.

To be a true WB closed loop system, you'd have to take into account how far off the current WB value is from the target value and adjust the trim faster the further away you are and slower the closer you are. That's something we may look into later, but for now a single switch point with adjustable update rates is working perfectly fine.

Thomas Dorris

 

[->PrOjEcTGS<-]

Ok. Maybe I'm still not getting it. This is off of your site: With ECMLink's narrowband simulation function, you can simply run the 0-5v output to the ECU and then configure the ECU to produce the 0-1v simulation internally, thus freeing up an additional input on the ECU for some other aftermarket sensor. That means that your way of doing things is still only utilizing 0-1v? Correct? Where my code the ECU switches between 127 or 2.47v and 129 or 2.51v. If I'm getting this right yours would be in the 0-1v region where the .04v variance would be a larger AFR difference than the wider resolution of 0-5v? I'm I interpreting your site wrong or what?

 

[twdorris]

Where my code the ECU switches between 127 or 2.47v and 129 or 2.51v.

Same here. It's just one point and an adjustable trim update rate. The switch in our code is not based on 0-1v values. It's based on raw wideband voltage. You select it here:

http://www.ecmtuning.com/images/forums/v3screenshots/nbo2sim.png

You then configure the rate at which STFT is adjusted (how quickly updates are made to the STFT value) based on how far away from the engine the O2 sensor is installed. This allows for very fine tuning of the behavior under varying airflow and operating conditions.

Thomas Dorris

 

[PIMking]

I think hes confused about how the ecu is going to act when you replace it with the wideband without changing anything. The answer is in the first post where he gave the hex editing to set the parameters for wideband in the ecu. He maybe doesnt do his own eprom editing and might not understand what was said. Maybe he uses dsmlink

Yes I have DSMlink

 

[PIMking]

Same here. It's just one point and an adjustable trim update rate. The switch in our code is not based on 0-1v values. It's based on raw wideband voltage. You select it here:

http://www.ecmtuning.com/images/forums/v3screenshots/nbo2sim.png

You then configure the rate at which STFT is adjusted (how quickly updates are made to the STFT value) based on how far away from the engine the O2 sensor is installed. This allows for very fine tuning of the behavior under varying airflow and operating conditions.

Thomas Dorris

Thomas,

Is that the same with DSMlink v 2.5?

 

[twdorris]

Is that the same with DSMlink v 2.5?

All direct access and narrowband O2 sim stuff is specific to V3.

Thomas Dorris

 

[the_mork]

The ecu takes the signal from the o2 sensor and uses it to determine whether the fuel mixture is rich or lean, that much we all know. I think what twdorris is saying is that it doesn't matter how sensitive and perfect the equipment you are using to determine if a room is light or dark if the only control that you have over the situation is a light switch. In this case there is some very useful potential to using a wideband natively if we also change our on/off style switch to a dimmer but since neither approach replaces the light switch both are equal and provide EXACTLY the same results. Furthermore converting the input from the wideband range to NB range doesn't reduce the accuracy of what is essentially a binary switch.

I'm certainly not a programmer though so I could be less accurate than I believe.