Bad Autotune, replace 02?
 

Just wanted to get a sanity check from you guys. New MSPNP ecu from Brain just installed. Purrs like a kitten when it idles on the base tune Brain loaded and runs smoothly in all gear and under boost.

I tried letting it autotune today and the passenger watched it consistently hit 7.6 AFR (idles at 10.3) and there was an error message on the autotune screen saying "Sensor out of range". The tune got worse and worse as we drove and cleared up when I restarted the car back on base tune.

The reason I got the new ECU was to datalog and figure out my overheating issue. I am starting to think that the old ecu was also tuning the car too lean and that is why I was getting over heating and glowing manifolds.

So, having cleaned the WB02 sensor last year when I was trying to diagnose things, do you guys suggest I just get a new sensor? It idles rich for sure so could have very easily gotten fouled. A quick google search has people suggesting replacing over cleaning since cleaning doesn't always work.

Thanks for any advice!

Have you tried to tune it leaner? Autotune is going to work better when you're at least relatively close to your target.

No, I hadn't gotten that far. Just working off of an initial gut feeling that the sensor might be bad. If the sensor is reading 7.6 AFR under load, could that be normal?

If the engine is not experiencing 7.6 AFR, then no, it's not normal. I doubt it would run well at that AFR, does it?

it runs pretty smoothly on the base tune, just some hickups at light load/throttle. It just runs rough when it starts to autotune

Your wideband isn't giving a valid signal. If you were really at 7.6 AFR, your engine would be flooded. At 10 AFR you should be belching black smoke.

Don't touch the tuning until you figure out whats up with the wideband.

What wideband. If its an innovate then 7.6 is their full rich reading. It could be richer than 7.6 and you won't know.

Thanks! This is what I wanted to hear since I don't know much about tuning and what values mean what for the engine. I will order a new o2 sensor when i get home and verify which one I have in my car.

It is an LC-1 but I saw they have two types (4.9 and 4.2) so I will have to check. Good to know that 7.6 is its bottom range stop.

Just have an LC1 and no dash gauge?

The Dash gauge in there now is a colored bar that fluctuates back and forth between lean and rich and has never filled me with confidence on its accuracy. No actual numbers to verify with.

Buy a new wideband O2 sensor. Either an MXT-L or an AEM EUGO. Quick dicking around around an lc-1.

yeah it's 7.35-23.4:1 range.

double check the AFR calibration is set to innovate default.

Hmm, I'll check the calibration settings when I get home. Why no love for the LC-1 Shuiend?

Because they are old and shitty. You can't tell what it is actually reading compared to what TS shows. Every other wideband I have tried has had a slight offset in TS compared to the gauge for the wideband. Usually it is a small offset but you can calibrate the MS to read correctly. With the LC-1 that is next to impossible to do due to not knowing what it is actually outputting. I used LC-1 way back in the day. Now I won't touch one.

I'll look into those other two sensors then, thanks!

They are both around $150 on Amazon prime. Much cheaper then a new motor. Both will use the newest Bosch wideband sensor.

Have you done a free air calibration like Innovate reccomends you do every so often? My friend bought a car with an LC2 in it and it would read max lean aby time you weren't in boost and max rich any time you were. A calibration cleared it right up.

I did back when I cleaned it last year and I might try it after checking calibration and giving it one last clean to avoid buying new. If all else fails then I might just swap it out for the MXT-L

The manual says to calibrate it when a new sensor is installed and again 3 months after that. Then, for a turbo car, every 6 months or 10k miles whichever comes first.

Being a track car that gets driven 3 times a year, that would probably be more like every 2 years but point taken :P I bought the MSPNP to be able to datalog and figure out what was wrong with my car so the o2 sensor making the car tune lean would absolutely make sense.

Just wanted to get people's opinion on cleaning vs. replacing and I'm glad I did since Shuiend had that advice about ditching LC. I wouldn't mind having a gauge on the dash that actually gives me a number and not just a spazzy colored bar having its own little rave

Dash party raves are fun though!

For a track car, they actually reccomend calibrating it every track weekend. :P

I would look into the AEM UEGO, no calibration needed and about $160 with a number display and an led guage. I have one, seems well built, but I have not gotten around to installing it yet, so I cannot offer any review past that

looks like the price has gone up since i bought mine a month or so ago, but MTX-L is much nicer than the AEM in my opinion. I have used the MTX-L on several cars. They have a faster refresh rate than what i have seen on the AEM and look better to boot. I got mine from the link below when it was 150 then i got 15% off since it was my first order. its up a bit now but still a touch less than 150 shipped.

https://jet.com/product/Innovate-Mot...f8b404786b52ad

fwiw, Amazon $169 as well - http://www.amazon.com/Innovate-Motor...9887327&sr=8-1

Ordered, I guess I have a project next free weekend

So you feel the need to spend money on a whole new sensor without ever calibrating the old one?

:burncash:

Blame him :giggle:

I don't think it's an issue of calibration. And I had an LC-1 with an XD16 gauge, and wrote my own software to compare the serial stream to the analog output - it was never more than .1 point off.

In the LC1 you've got 2 outputs - analog1 and analog2. By default, one is a 0-5v reference for wideband output and the other is a narrowband emulator. On the Megasquirt side, you've got https://www.diyautotune.com/support/...novate-config/ . There's also the logworks software to verify what the LC-1 is outputting.

So the last 6 widebands I have used (3 AEM UEGO, 3 MXT-L) all calibrated like the link your posted gave me different values in TS then what the actual wideband gauge displayed. For everyone I had to go in an manually change calibration values in TS to get the 2 gauges to match. None of them used exactly the same values to calibrate. They were all close, but none exactly the same. I have also had 2 LC-1's randomly die on me, not the sensor, the actual controller. I also know Braineack had problems with at least 1 LC-1 controller. I also never had any luck with logworks when I tried it on my LC-1's. After those experiences I won't bother using a wideband controller that does not have a built in gauge that I can use to compare values. To me it is not worth having something shitty happen and the engine going kaboom. Then again I have a real job now and can afford to overnight parts from Japan so that might factor into some of my rational.

I'm not looking to poke the bear of "this wideband is more accurate than this one" - there's plenty of reasons why it happens with ground offsets, wiring, sensor variances. But in this case with a super rich setting at idle - somethings fucked far beyond a bad sensor. Miswired, not set up correctly are my thoughts.

My old lc1 works awesome. Probably a wiring issue

Sorry to disappoint you but compressed time-tables also play a factor. I am trying to get this car running reliably by an event in May so if I can eliminate a possible failure with a new sensor (and gain the ability to read actual AFR with an actual readout on a gauge) then that is move valuable to me than spending the time to save the one I had.

If I got advice that I should just clean it and I would probably be fine then I would go for that but the advice I'm getting is pretty soundly in the "get a better/new sensor" category. Plus the LC-1 was purchased and installed by the original owner around 6 years ago so replacing it with something I have installed and wired would make me feel a lot better about it working right.

when you wired it up, did you follow the wiring info that brain sent you with your megasquirt? If you did, then you should be in good shape for the wiring.

The Lc-1 and the Mtx-l both use a bosch sensor, but they are different bosch sensors. The mtx-l one is was never put on a car from the factory so it cannot be found at autozone (online is the place to get them). The Lc1 uses a bosch sensor from a late 90's vw jetta turbo or something. I would recalibrate that in a heart beat, try and use a little bit of break cleaner and compressed air to clean it out, I'm sure its black from running so rich. As far as settings go, my mtx-l reads closer to TS when set as an LC-1. So some widebands like different settings. I'd clean it, recalibrate it, and experiment at idle! which ones give you a closer reading between the screen and TS. If the sensor hasn't gotten really wet, its probably fine. The mtx-l is nice and I'm sure you should upgrade soon, but try cleaning it first.

Yeah the LC1 uses the Bosch 4.2. The LC2 and MTXl uses the Bosch 4.9.

Personally I'd first confirm that you're using the correct output (should be using the yellow wire) and connect to LM Programmer to make sure that output is actually outputting 7.35-23.4:1 through 0-5v.

if you're sending the 0-1v output to the MS, you're going to read crazy, crazy, rich. You'll only be able to see between 7.4 and 11:1 AFR.

Seeing 7.4:1 in any case suggests you're either stupid rich and dumping raw fuel out the tailpipes, or there's no voltage present.

MTX-L is in the car and fired it up with cold-start AFR around 15.0. Going to give some tuning a shot and see what it does!

Man I like this gauge sooooo much better than rave light bar I had before. Actual digital readouts are wonderful

The MTX-L has been sold at various times with the LSU4.2 and LSU 4.9. You can swap an LSU4.2 onto the later MTX-Ls by swapping over the sensor cable.