AFR Stelvio 8v - the truth!

[Xlratr]

After playing around with my fuel and ignition maps for the last couple of weeks, I was really pleased with the way my Stelvio 8v was running. For those of you who have had a look at the standard fuel map in Tuner Pro, you might have noticed that there's a big dip in the original fuel curve around 5,000 rpm. It's really easy to see if you copy the numbers and paste them into Excel and make a little line graph.
So I filled in the dip!

But I still had an uneasy feeling that maybe I'd messed something up, so I got myself an Innovate LC-2 kit with wideband sensor and AFR gauge and went looking for the truth!

I installed it on my bike (2012 Stelvio, Lambdas off, Zard with DB Killer), reloaded the original map but kept the Lambdas off, marked the throttle positions on the grip and went for a ride. I quickly realised that it's really only a small number of positions on the map that are really relevant, and for me that was the area between 3,500 and 5,000 rpm at steady highway speeds.

This is the AFR the standard map gives you with Lambdas off (i.e. In open loop) and with a Zard exhaust at constant speed:

4,000 rpm = 12.9:1
4,500 rpm = 14.2:1
5,000 rpm = 15.1:1

Under load, the AFR was fine, but I was surprised that the AFR is so rich at a steady 4,000 and so lean at 5,000. The 15.1 was a bit scary!

With those numbers (and a few others) I was able to redo my fuel map again, and after a couple of tries I now have it running at around 13.4 between 4,000 and 5,000. I've left most of the rest of the map pretty much standard.

The other thing I learnt was that filling in the dip in the curve at 5,000 was way too much fuel. There needs to be a dip there, but just not so big as from the factory.

Those numbers above might be interesting for any of you guys with aftermarket exhausts.

Now I can sleep at night! :-)

[Vasco DG]

Be careful relying on EGA alone the engine's head design and camming means that AFR readings alone can give a false impression of what is happening inside the combustion chamber.

Different pipes can have a radical effect on how the engine breathes and fuels as well.

Pete

[Larry]

You realize that you've just given him insomnia now.  ::)

[Xlratr]

Be careful relying on EGA alone the engine's head design and camming means that AFR readings alone can give a false impression of what is happening inside the combustion chamber.

Different pipes can have a radical effect on how the engine breathes and fuels as well.

Pete

Pete, thanks for the sleepless night!  ;D

I had read your posts on that, and it was in the back of my mind. But my theory is that the map is a bit too rich at 4,000 because the Lambdas are supposed to take care of that. 5,000 is the early stage of open loop so I think the factory probably tuned it for 14.7 with the standard exhaust for emissions reasons (the dip in the fuel curve right there is massive!), and I think the Zard just pushed it further up to 15.1. I'm thinking that at 4,000 rpm the cam design and exhaust combination shouldn't be so radical to be letting unburnt fuel through.

On the other hand, I wonder if there's room in the shed for a Dyno?  :D

[molly]

What you have done is basically how I set up my 1200 2v Sport. Does the bike feel it is running better with your changes? The best way is to check is to do back to back runs with the original map and compare throttle response and fuel consumption. Once you feel your modified map is an improvement make any further changes over a extended period using small incremental  steps.

Dyno produced maps don't always give the best on the road performance. I compared my data logged map to a generic pro tuned one and although it made slightly more power mine gave the best all round results.

[Vasco DG]

What you have done is basically how I set up my 1200 2v Sport. Does the bike feel it is running better with your changes? The best way is to check is to do back to back runs with the original map and compare throttle response and fuel consumption. Once you feel your modified map is an improvement make any further changes over a extended period using small incremental  steps.

 :+1

Really, there rare a host of so far unknown tables and adjustments the only way to progress is in small increments and try to identify what the *Unknowns* are.

Pete

[Xlratr]

The first map I made included lot's of changes to the fuel curves in areas that just didn't look right to me, and it included some quite big increases in fuel in the mid range. But I was just working on assumptions and I now realise that most of those carefully calculated changes were pointless. Apart from under load acceleration (which seems to be fine) and off idle response, the only relevant area for me to concentrate on is right in the middle of the map, i.e. 4,000 to 5,000 at half throttle.

Molly and Pete, I agree on what you say about making small adjustments. All the changes I did made it too difficult to judge what I'd done right and what was wrong. That's why I started again with the factory map and just concentrated on one small area.

Pete, I understand what you're saying about camshaft effects, but on the other hand I can imagine the AFR gauge reading richer than what's going on in the engine under certain circumstances, but I can't see it reading leaner. All those gases have got to come about sometime right?

So I am convinced that with a Zard exhaust and standard map, the motor is running too lean at highway speed (at least German highway speeds! :-).

And Molly, yes I do think the motor feels less stressed at 5,000 rpm than it did, although it's still too early to talk about fuel consumption.

[pauldaytona]

You can record the results with a bluetooth dongle and an android app,
https://play.google.com/store/apps/details?id=scantwin.scanmM5.AllinOne

For recording the lambda with rpm and tps, connect the LC2 linear output to the input of your OEM lambda, the 5am will show the values even when lambda is off. Then with logworks you can display the log in 3d. And get all sorts of views. 

 Further, at each rpm there is a corresponding tps value, with recording you find that easy. There is the steady speed value, and you can do with lambda 1 there, keeps consumption down. Under that line is when you decelerate, baove it is when you accelerate, there you want more fuel.

 

[Xlratr]

You can record the results with a bluetooth dongle and an android app,
https://play.google.com/store/apps/details?id=scantwin.

Thanks for that Paul. You're the second person encouraging me to go that route, so I might just have to do that.

One question. There are two voltage outputs from the LC-2. One of them is 0-5v, which I have connected to my AFR gauge. The other one is currently not used but should be 1.1v = 14AFR, O.1v = 15AFR. Which one should I connect to the ECU? Neither of those two value ranges correspond to what a narrow band sensor has in output values. Will it not damage the ECU if I connect a 0-5v source?




Sent from my iPad using Tapatalk

[Xlratr]

One more question! Is there anything to watch out for when buying the Bluetooth dongle, i.e only certain types work, etc??


Sent from my iPad using Tapatalk

[Xlratr]

Obviously I should read before asking questions!  ;). I just discovered that the analog output voltages from the LC-2 are programmeable!!
Should I go with 0.0v = 0.5 Lambda, 1.0v = 1.5 Lambda? I think that would keep me within the same voltage range of the narrow band sensor so there should be no problem for my ECU. Lean/rich is the oposite way round to a narrow band, but I suppose that is OK (my ECU doesn't know that! :-)

[pauldaytona]

Obviously I should read before asking questions!  ;). I just discovered that the analog output voltages from the LC-2 are programmeable!!
Should I go with 0.0v = 0.5 Lambda, 1.0v = 1.5 Lambda? I think that would keep me within the same voltage range of the narrow band sensor so there should be no problem for my ECU. Lean/rich is the oposite way round to a narrow band, but I suppose that is OK (my ECU doesn't know that! :-)

It has two outputs, one is a smallband curve, the other is linear. The linear one is the one to use. The ecu does with it nothing else then displaying it on the diagnostics, and that is what you need.   Just leave it as default that works. If you are there, just two values are needed in the App, to calculate the voltgae to lambda or afr display. If you are there ask again, I'll tell you the values. You can check if reading does give same values in guzzidiag as in to the lc2  connected logworks.

I have a cheap chineese Bt dongle. 

[makarushka]

I can imagine the AFR gauge reading richer than what's going on in the engine under certain circumstances, but I can't see it reading leaner

It absolutely can read leaner due to reverse pulses when the air is being sucked into the exhaust. This is strongly dependent on the position of the O2 sensor and how free-flowing the exhaust is... It's also one main reason that Autotune and similar systems require skilled tuning to work and are not simple "set-it-and-forget-it" tools; one needs to disable the correction in certain areas as it may be thrown off by erroneous O2 readings. Vacuum switches have been employed for disabling the reading on the deceleration. But then you still need to fill those areas on the map with a correct value, so some amount of educated guesswork and trial-and-error is still required.

[Xlratr]

@Paul: I think the LC-2 allows both outputs to be programmed between 0-5v, but I haven't been able to check as I'm still waiting for my USB/Serial adapter to arrive. But either way, I'll go with what you said and connect it to the same output that my AFR gauge uses (i.e. 0 to 5v). I'm also still waiting for my budget Android tablet to arrive so that I can test the app (I'm an iPhone user), so I'll report back when I get everything set up.

@Makarushka: Very good point! I hadn't thought that through properly. On the other hand, I assume that if the AFR stays lean over a spread of 500 rpm or more, it is probably more related to fuel than to unwanted harmonic effects. But assumptions are dangerous and for sure I'll be doing plug checks after I get the numbers to look right. At the end of the day, I'm not making any massive changes and the revised values I'm working with are definitely not going to put me in a dangerously rich zone. It's just interesting to see how the bike will work with slightly different curves.

[Vasco DG]

Charge transition at certain RPM with various pipe and filter combinations is also a contributor to odd AFR readings and peculiar behaviours.

Pete

[Xlratr]

I hate electrics!

Still waiting for my Android device to arrive so I can't test the Bluetooth dongle and App yet. Hopefully tomorrow, because it'll be the last chance to work on the bike for a few days.

But I did get my Serial to USB adapter, so I could check the output voltages of the LC-2. I reprogrammed the second output and I now have both running 0-5v. One is connected to the AFR gauge and I will connect the other to the (currently unused) Lambda Superseal connector. I checked the output with my Voltmeter and both are showing around 2.5v at start up (Lambda value 14.7), so that's fine.

I'm also still waiting for some Superseal connectors to arrive, and I hate cutting in to original wiring if I can avoid it, so at the moment I haven't been able to properly connect the LC-2 output to my ECU. However, when I strip the end of the wire and push it in to the Connector block, I can't see any signal in Guzzidiag. I know it's outputting the right voltage, and I pushed the cable as far as it would go into the connector block. Nothing!

The black wire coming from the original 4 wire narrow band sensor is the signal carrying wire (correct?) and I'm pushing the signal wire from the LC-2 into the corresponding part of the female connector block. I also tried earthing the other connector (although I don't think it's necessary), but that didn't help.

It might be OK when I get the Superseal connector on the end, but I don't think so. I'm doing something wrong. Could somebody just reconfirm exactly where the signal carrying wire from the LC-2 should go?

Did I mention that I hate electrics?





Sent from my iPad using Tapatalk

[beetle]

The black wire is signal +, and the grey wire is signal -


 ;-T

[Xlratr]

Thanks, that's what I was working with. But do I just connect the LC-2 output to where the black wire used to be? Nothing else?


Sent from my iPad using Tapatalk

[pauldaytona]

I connected two wires in the 4 pin superseal, one is the signal from the lc2, the other one is +12v, that goes to the fiat adapter. So you can leave the BT and fiat adapter in the toolbox compartment.  So pin 1 = output lc2 and  pin 4 +12v  I have the mtxl and connected his power to the navi power, that is switched too.
I too hate to cut in the original wireloom.  The ground comes from the general ground.

[Xlratr]

Thanks Paul. Just tried LC-2 signal on PIN 1 and 12v on PIN 4, but still no mV reading in Guzzidiag. The voltage is there at the cable when I connect my Multimeter.



Sent from my iPad using Tapatalk

[Hellgate]

Amazon says my cables arrived today, but I'm on turning the road for another week. I love how we can get into the ECU and tinker.

Sent from my SAMSUNG-SGH-I747 using Tapatalk

[pauldaytona]

you have two lambdas, When you have oem lambdas connected, do they display both mv in guzzidiag? The second lambda is something I never could test, because I have no testobject. Else try to connect to  the other lambda connection

[Xlratr]

you have two lambdas, When you have oem lambdas connected, do they display both mv in guzzidiag? The second lambda is something I never could test, because I have no testobject. Else try to connect to  the other lambda connection

Hi Paul. I have now removed both the original Lambdas but I believe they did both show values in Guzzidiag when I selected Lambda mV1 and Lambda mV2 from the drop down menus. I can easily reinstall one of the Lambdas in the currently unused location this evening just to make sure. You think it might be because they are switched off in the map? 

[pauldaytona]

No it was just because I wasn't shure both showed good values because I didn't test them. With lambda off, they show in diagnostics, else the logging this way woudn't work.

[Xlratr]

No it was just because I wasn't shure both showed good values because I didn't test them. With lambda off, they show in diagnostics, else the logging this way woudn't work.

Hi Paul. It seems there is something different about the way the way the twin Lambda ECU works. I just put back one of tho original Lamdbas, and I still couldn't get a mV reading. But when I reinstalled a map where the Lambdas are switched on, then I can see the voltage in Guzzidiag.

Also, when I then connect the wideband on the other side, it shows max 1.2v, athough the AFR shows 14.5 and my multimeter shows around 2.5v coming out of the LC2.

It seems to me, I would have to have the Lambdas switched on in the map, but not connected. The ECU would then default to Open Loop I think. But I still have the issue that the mV output from the ECU is completely wrong.

[beetle]

It seems to me, I would have to have the Lambdas switched on in the map, but not connected. The ECU would then default to Open Loop I think. But I still have the issue that the mV output from the ECU is completely wrong.

That won't work. With lambda on and the sensors not connected, you'll get a fault registered.

Is the output of this LC2 differential or referenced above ground?

I'll need to look at the Stelvio wiring diagram. I'll be back.....

[Wayne Orwig]

Also, when I then connect the wideband on the other side, it shows max 1.2v, athough the AFR shows 14.5 and my multimeter shows around 2.5v coming out of the LC2.

A normal O2 sensor only operates in the zero to one volt range. It never goes over one volt. Can you set the wideband sensor to operate in that range?

[beetle]

Wayne, the Innovate devices with wideband sensors are usually a 0-5V linear output. My TechEdge wideband controller has a differential 0-5V linear output and it works with my Griso. The Stelvio narrowband sensors are the same units, so I see no reason for John's setup not to work if he gets the wiring correct.

John, if I understand correctly, you are only connecting the yellow wire of the LC-2 to the ECU sensor + signal wire? The  - signal wire will need to be referenced against the yellow wire output. However, if you connect it to ground I'm not sure whether the ECU will freak out.

[Xlratr]

Mark, I'm actually connecting the the brown wire from the LC-2 to where the black wire from the narrow band used to go. This is normally the one that simulates the narrow band output, but I reprogrammed it to be exactly the same as the yellow one, i.e. 0-5v linear. I checked the output with my multimeter and it reads 2.5v, bang in the middle, i.e Lambda 1. But when I connect it up to the ECU, it only shows around half that in Guzzidiag. Very strange. I'm thinking the 2 Lambda ECU can't handle the voltage as it's outside the normal working range, but I've no idea if that's true.

And as I mentioned, Guzzidiag only shows a mV value when using a map where the Lambdas are on. Otherwise, zero!
I wonder if it could have something to do with the third Lambda flag that you can see in the map in Tuner Pro? I switched all three to off. maybe the third one should stay on?????

I'm off on a 4,000km trip to Norway tomorrow, so I won't be able to investigate any further for another week. I spent the last couple of hours with an Android tablet fixed to my Ram Mount running the Scan App, and with an action cam in my left hand going up and down the autobahn at different speeds :-) . (Alternative logging!!!). It actually worked pretty well and I think the map is good. I'll fine tune a bit more when I'm back.

Thanks for all the input guys!




Sent from my iPad using Tapatalk

[Xlratr]

A normal O2 sensor only operates in the zero to one volt range. It never goes over one volt. Can you set the wideband sensor to operate in that range?

Wayne, yes I can programme the output to run Linear between 0-1 volt, just like a narrow band. But Paul and Mark say it should work with 0-5v, or at least the single Lambda bikes do. But even If I do that I still have the issue that no mV value is received at all when the Lambdas are switched off in the map.


Sent from my iPad using Tapatalk