The Top DSM Community on the Web

For 1990-1999 Mitsubishi Eclipse, Eagle Talon, Plymouth Laser, and Galant VR-4 Owners. Log in to remove most ads.

Please Support JNZ Tuning
Please Support STM Tuned

Resolved Is this normal AFR reading / wideband behavior during cruise?

This site may earn a commission from merchant
affiliate links, including eBay, Amazon, and others.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Kryndon

Proven Member
944
591
Jan 10, 2014
Bulgaria, Europe
Still getting used to having a wideband O2 and gauge. I noticed that my AFRs on the gauge switch very quickly between 14.6 and 15.1 AFR during cruising at 60 mph and a mostly level surface. It measures very close to that during idle, too. I am just unsure if it's meant to do that or just stay at a constant 14.7 for example (if everything is running fine). Could a tiny air leak at the bung weld or anywhere before the O2 cause the reading to go to 15 like in my video? Or am I simply not understanding how the O2 operates? I know it's based on a 0-5 volt range but still. Please take a look just in case, sorry for the shakiness! Also just to clarify, I'm not asking if the actual AFRs are okay, but rather the constant switching on the readings.

You must be logged in to view this image or video.
 
I have no experience with digital gauges but always assumed that a wide-band would kind of smooth out the oscillation of your O2 so you get a steady readout. What good is a number that changes so fast that you can't see what its saying? If my analog gauge did this I would have returned it.
 
Thanks guys, I'm glad it's looking normal. Hopefully this year I'll finally integrate ECMlink into my car and get more use out of the wideband! (and tons more questions to follow with that, hah)

P.S ----- Blinky I just saw you posted a message while I was writing mine. I think 2g2slow97 raises a good point. The main front O2 is running in closed loop and cycling constantly. It's impossible to have a 100% always accurate and equal air fuel ratio going into the engine and coming out, at least not with these old cars. There would be tiny differences of fuel and air particles flowing in and out, and apparently the wideband Bosch unit is so accurate and sensitive that it picks these changes up immediately as they flow through the pipes. I don't know if there's a way to adjust the wideband's controller to log the readings in longer intervals and how that would affect tuning later on.
 
Thanks guys, I'm glad it's looking normal. Hopefully this year I'll finally integrate ECMlink into my car and get more use out of the wideband! (and tons more questions to follow with that, hah)

P.S ----- Blinky I just saw you posted a message while I was writing mine. I think 2g2slow97 raises a good point. The main front O2 is running in closed loop and cycling constantly. It's impossible to have a 100% always accurate and equal air fuel ratio going into the engine and coming out, at least not with these old cars. There would be tiny differences of fuel and air particles flowing in and out, and apparently the wideband Bosch unit is so accurate and sensitive that it picks these changes up immediately as they flow through the pipes. I don't know if there's a way to adjust the wideband's controller to log the readings in longer intervals and how that would affect tuning later on.

I see this a bit using the narrow sim in link.
It cycles the O2 at .20v to .80 with squared edges. Unlike using just an O2 it gives more Gradual drops.

Lock your car into open loop for a moment while driving. You’ll see your wide band act differently.
 
I have no experience with digital gauges but always assumed that a wide-band would kind of smooth out the oscillation of your O2 so you get a steady readout. What good is a number that changes so fast that you can't see what its saying? If my analog gauge did this I would have returned it.

WBO2, especially those that are digital gauged, are meant for instantaneous readings.. an average/smoothing would delay your ability to read what your O2 is at the moment you look at the gauge or distort the actual number because it took other previous readings into account. A WBO2 has no understanding of open or closed loop operation.

An analog gauge is different because the speed of the stepper motor. It's not as fast as a digital readout, hence you always get either delayed or "averaged" readings. Not as accurate but as long as it has a quality sensor and stepper, still usable.

When in closed loop your ECU is always chasing the sensor's last readings to adjust for ~14.7:1 so it will oscillate. The two are never in perfect sync in CL and you blow past 14.7 and go rich or lean, then it has to backtrack. Perfectly normal. However, when it doesn't oscillate, you're likely running in open loop and it's up you to and your tune to determine whether that's correct given how you're operating the car at the time.

A slow (old, poor quality) sensor will cause a longer oscillation period which will affect how the engine runs because of the increased length of time it's outside 14.7. If it's well off, you may notice decreased or inconsistent torque output.

A pre-sensor leak, depending on how bad, may still allow you to oscillate around 14.7 provided you have enough space in your fuel trims +% for the ECU to adjust with extra fueling but your trims would read higher than 100%. If you have a constant high AFR on cruise in closed loop, you probably blew past the adjustment % of the ECU for extra fueling which depends on ECU. Look at your trims in that case and fix the leak.
 
I want to come back to this as this originally was my thread.

In that video I was still running the stock turbo ECU without any retunes. Now I am running ECMlink + injectors/AFPR/Pump and after comparing, I can see there is a clear difference.

Right now at idle my average AFREst is 13.5 while at the same time my Wideband is reading 16.1. In addition, the gauge might be reading 15.3 while ECMlink is logging it as 16.1. Both the controller and gauge share the same ground as the ECU ground under the radio.

My AirflowPerRev is 0.26 g/s so that's almost spot on and my STFTs are within +2% but LTFT Lo is +9.2%. Could my Long Term Fuel Trims requiring an additional 9.2% fuel be causing my idle to be so lean? If so, is it better to bump up Global Fuel (essentially downscaling my 1000cc injectors) or moving the MAF Comp sliders up?

Conversely, moving the MAF sliders up artificially raises my AIrflowPerRev so that will again take my tune out of whack.
 
I want to come back to this as this originally was my thread.

In that video I was still running the stock turbo ECU without any retunes. Now I am running ECMlink + injectors/AFPR/Pump and now after comparing, I can see there is a clear difference.

Right now at idle my average AFREst is 13.5 while at the same time my Wideband is reading 16.1. In addition, the gauge might be reading 15.3 while ECMlink is logging it as 16.1. Both the controller and gauge share the same ground as the ECU ground under the radio.

My AirflowPerRev is 0.26 g/s so that's almost spot on and my STFTs are within +2% but LTFT Lo is +9.2%. Could my Long Term Fuel Trims requiring an additional 9.2% fuel be causing my idle to be so lean? If so, is it better to bump up Global Fuel (essentially downscaling my 1000cc injectors) or moving the MAF Comp sliders up? Conversely, moving the MAF sliders up artificially raises my AIrflowPerRev so that will again take my tune out of whack.
First thing is first.
Your wideband gauge has to match the wide band value in ecmlink or else your fuel trims will never be right consistently.
 
First thing is first.
Your wideband gauge has to match the wide band value in ecmlink or else your fuel trims will never be right consistently.
That's what I don't get, why is my gauge and ECMlink WB value not the same? At WOT they both show 11.7 for example and my AFREst and WB matches but on idle and cruise they both have almost a point difference. Should I just set it up as Linear wideband and plug in the voltages or keep it as LC-1 under the ECU inputs tab?
 
That's what I don't get, why is my gauge and ECMlink WB value not the same? At WOT they both show 11.7 for example and my AFREst and WB matches but on idle and cruise they both have almost a point difference. Should I just set it up as Linear wideband and plug in the voltages or keep it as LC-1 under the ECU inputs tab?
Post a log Stefan
 
That's what I don't get, why is my gauge and ECMlink WB value not the same? At WOT they both show 11.7 for example and my AFREst and WB matches but on idle and cruise they both have almost a point difference. Should I just set it up as Linear wideband and plug in the voltages or keep it as LC-1 under the ECU inputs tab?

I’m not sure exactly how the scaling works with ecmlink, but (I know with transducers and other analog signals) the scaling is so large that the signal gets skewed toward the middle then gets closer as it gets to the lower and upper limit.

Try to close the lower and upper in the wideband preference settings to 16.5 and 10.5. The numbers above and below those you’re not worried about anyway..
That might help the wideband match the ecu wideband.

Also, during idle and cruse you’ll have a delay between the two.
The gas has to get through the exhaust to the O2 then send a signal back to your mixture.
So if it’s fluctuating there could be a difference in what you see instantly.
 
So this log was from the other day when I had to replace my AFPR since the diaphragm tore itself. Naturally it messed up my tune and now I have to re-add some more global fuel to compensate for the fuel loss as well as re-tweak my injbattadjust table since I have too much deadtime set at starting voltage and she floods with fuel. So in this case don't worry about my AFRest and WB not matching at WOT.

The problem is, during idle and cruise like I said, the wideband reading in ECMlink maybe says 16.2 while at the same moment my actual gauge is reading 15.4. But there is no way for you guys to know how much my gauge is reading. Maybe I can try taking a video of the laptop next to my gauge and you can catch it.

I did recalibrate the LC2 the day I fixed my AFPR and cleaned up its tip from soot. My Wideband is positioned about 40 inches downstream of the factory front O2 narrowband but in ECMlink the delay between AFRest and Wideband I have noticed is roughly 0.5 seconds.
 

Attachments

  • Idle + medium pulls 08.13.2022.elg
    161 KB · Views: 22
Both the controller and gauge share the same ground as the ECU ground under the radio.

Just something to consider: The ECMLink wiki has a page on wiring of these gauges, and other sensors. The biggest key is to ensure the wideband signal ground is tied to a signal ground pin on the ecu. Not chassis. If you do not do this, the wideband and ECU are basing voltage measurements on inconsistent "perceptions" of 0V.

 
Just something to consider: The ECMLink wiki has a page on wiring of these gauges, and other sensors. The biggest key is to ensure the wideband signal ground is tied to a signal ground pin on the ecu. Not chassis. If you do not do this, the wideband and ECU are basing voltage measurements on inconsistent "perceptions" of 0V.

I have gone through that page a few times and it says you CAN tie the ground to the ECU ground pin however you can also tie it to the main ECU grounding location which is under the radio on the chassis. Thats how i have it set up. My wideband runs all the time and i fear it may put too much load on the ecu ground paths. The issue may be due to something else.
 
I have gone through that page a few times and it says you CAN tie the ground to the ECU ground pin however you can also tie it to the main ECU grounding location which is under the radio on the chassis. Thats how i have it set up. My wideband runs all the time and i fear it may put too much load on the ecu ground paths. The issue may be due to something else.
I am going to disagree with you here. The ECMLink page specifies to not connect a heater ground wire on an LC1 (blue) to ECU signal ground. The sensor ground (white) in their guide is specified to be tied to ecu sensor ground. This is a must. That is how analog sampling works. Chassis ground is simply not good enough which can manifest itself in what you are seeing. I'm not saying there aren't other issues but this is a pretty obvious variable I'd remove as you try to get this resolved. Which specific model of gauge/controller are you use? You mention LC2 in your car profile but talk about a gauge reading, which the LC2 doesn't have, unless you are serialed into it on a laptop with the Innovate software.
 
I am going to disagree with you here. The ECMLink page specifies to not connect a heater ground wire on an LC1 (blue) to ECU signal ground. The sensor ground (white) in their guide is specified to be tied to ecu sensor ground. This is a must. That is how analog sampling works. Chassis ground is simply not good enough which can manifest itself in what you are seeing. I'm not saying there aren't other issues but this is a pretty obvious variable I'd remove as you try to get this resolved. Which specific model of gauge/controller are you use? You mention LC2 in your car profile but talk about a gauge reading, which the LC2 doesn't have, unless you are serialed into it on a laptop with the Innovate software.
I have and am using the Innovate LC-2 wideband which came with a Bosch LSU 4.9 O2 sensor. Since ECMlink only has an LC-1 option, that is the one I picked. But as for wiring it up, I went with the actual LC-2 instructions, such as getting power from a 12v headlight switch and merging the controller + gauge grounds into a single wire which then is grounded under the "E" bolt for the main ECU ground under radio. The LC-2 comes with it's own gauge as you can see in the video in my very first post.

@TK's9d2TSi That's the issue Tony. My Wideband gauge reads lower than the value it sends to ECMlink through the controller. I have adjusted my biss properly however the new value for it is closer to 45 since I am running a selected idle RPM of 850. The 30 value for ISC is related to stock setting at 750 RPM +-50. The higher you go, the higher the ISC will need to be adjusted to keep a constant good idle, accessories and fans off.
 
I have and am using the Innovate LC-2 wideband which came with a Bosch LSU 4.9 O2 sensor. Since ECMlink only has an LC-1 option, that is the one I picked. But as for wiring it up, I went with the actual LC-2 instructions, such as getting power from a 12v headlight switch and merging the controller + gauge grounds into a single wire which then is grounded under the "E" bolt for the main ECU ground under radio. The LC-2 comes with it's own gauge as you can see in the video in my very first post.

@TK's9d2TSi That's the issue Tony. My Wideband gauge reads lower than the value it sends to ECMlink through the controller. I have adjusted my biss properly however the new value for it is closer to 45 since I am running a selected idle RPM of 850. The 30 value for ISC is related to stock setting at 750 RPM +-50. The higher you go, the higher the ISC will need to be adjusted to keep a constant good idle, accessories and fans off.
Don't ever share power with your innovate wideband and another device. Always always run a dedicated power/relay for the wideband. Run your ground to ecu chassis ground. The end.
 
I have and am using the Innovate LC-2 wideband which came with a Bosch LSU 4.9 O2 sensor. Since ECMlink only has an LC-1 option, that is the one I picked. But as for wiring it up, I went with the actual LC-2 instructions, such as getting power from a 12v headlight switch and merging the controller + gauge grounds into a single wire which then is grounded under the "E" bolt for the main ECU ground under radio. The LC-2 comes with it's own gauge as you can see in the video in my very first post.
From ECMLink's wiki page on LC1 installs (but applies generally to Innovates of other part numbers and really sensors as a whole):

"The LC-1 install notes suggest that you tie the BLUE and WHITE wires together. They suggest this to keep things simple across all their different customer installations. In our case, we want the best possible reference point for this wire. And that's the ECU's sensor ground.

The ECU pins listed above are good choices or you can use any of the sensor ground points under the hood too (MAF connector, TPS connector, MDP connector, etc.). But, really, the best place would be right near the ECU since everything you need is right there anyway.

The only major concern with connecting something to the ECU's sensor ground is that it might draw too much current and blow the ECU's internal sensor ground track. But we've measured current draw on the WHITE wire to be a measly 45-50 mA. This will not cause any problems on the ECU's sensor ground circuit.

NOTE: Do NOT connect the BLUE wire from the LC-1 to the ECU's sensor ground. You will damage the ECU's sensor ground track if you do this. ONLY connect the WHITE wire from the LC-1 to the ECU's sensor ground, NOT the BLUE."


You don't have the same colors as noted above an LC2 (setting aside the LC2's manual has 0 mention of a gauge and has no provisions or wiring for one, aside from the serial out and the yellow 5V we tie to ecu). It only has 4 wires, one of which is ground. This likely means that you will be unable to attach the ground wire from your LC2 to the ECU pin, as ECMTuning notes, due to the power requirements for the heater.

Given that, I think you will always see a shift in this value at some point along the sensing curve. You can try to tune this out with the primitive linear scaling ECMLink has, but I think you'll end up with it reading off somewhere else instead.

There is no way to get around the fact that chassis ground and ECU power ground are not signal ground. The signal ground is tied to the ADC (analog to digital converter) chip inside the ecu that is sampling this 5v output from the LC2 and converting it to a digital value for the ecu to read. It's frame of reference on 0V and the LC2's are not the same as there are most certainly other components along the circuit until that ADC ground sees chassis ground.

This is the same reason ECMTuning notes above that while you can ground the sensor under the hood, the wire length along that line will throw off the signal as well. And suggest grounding directly at the ECU.
 
The 30 value for ISC is related to stock setting at 750 RPM +-50. The higher you go, the higher the ISC will need to be adjusted to keep a constant good idle, accessories and fans off.
I never seen this anywhere. I’m at 1k rpm’s idle with isc at 30. Idle is butter.

I run the wb power (w inline fuse) and ground straight to battery. Gauge and link readings are identical.
 

BLACK wire - notes​

A good, solid high-current grounding point here. The ECU pins listed above are good choices or you can try a good, clean chassis ground point too.
NOTE: Do NOT connect the BLACK wire from the LC-2 to the ECU's sensor ground. You will damage the ECU's sensor ground track if you do this.

So. I was unaware that ECMTuning had an LC2 wiki page. As I assumed, you are unable to tie the ground to the ECU sensor ground. Given that, I am unsure you will be able to 100% dial this in to perfect at cruise and not sacrifice accuracy elsewhere. They seem to suggest that you can use the LC1 preset and leave the black wire at chassis or ECU power ground. But there is a reason, when possible, they recommend using sensor ground.
 
Support Vendors who Support the DSM Community
Boosted Fabrication ECM Tuning ExtremePSI Fuel Injector Clinic Innovation Products Jacks Transmissions JNZ Tuning Kiggly Racing Morrison Fabrications MyMitsubishiStore.com RixRacing RockAuto RTM Racing STM Tuned

Latest posts

Build Thread Updates

Vendor Updates

Latest Classifieds

Back
Top