Forum » General Tuning Discussion » Auto-Tuning on Engine with no MAF, ITBs and Narrow Band O2

Auto-Tuning on Engine with no MAF, ITBs and Narrow Band O2

General Tuning Discussion

Discuss all things tuning in this section. News, products, problems and results. 


Page 1
Author
165 Views

I'm looking for a bit of help validating something a tuner said to me which I think is a bit suspect and this is going to be a bit of a long post, sorry.

In this post I'm talking purely about motorcycle engines running ITB. I have a BMW S1000R to which I have fitted a full Akrapovic race exhaust system. When I fitted the system I wanted to get the bike tuned purely to correct the fueling as it's a road bike and I am not after out and out performance from the engine. Also, I couldn't find any aftermarket ECU software for the BMW ECU so in the end I used a piggyback fuel module from Power Commander.

During my search, I came across a tuner that said they could tune the bike without modifying the software on the ECU and what's more their modification would leave the bike able to self-learn or auto-tune for any modification I made now or in the future. Call me cynical but I'm always suspicious of anything that has the word 'auto-magic' in the title. I asked the guys a bit more about the system, how do they do what they do without changing the stock ECU software and most importantly how their auto-tune feature worked. This is where things started to get vague.

With regards to the software, what they said was that they install their software directly onto the chipset of the ECU utilising unused space on the EPROM. Apparently, this space is always present in all ECUs because the chips themselves are only made by a couple of companies which are then utilised by the likes of Bosch, Marelli, etc. The likes of Bosch, etc are not allowed to look into this space on the chips which why it can't be seen by anyone other than this tuner and the companies that made the chips. Now, this seems a bit suspect to me but I don't have any experience programming embedded computer systems so I can't say if this is possible/true one way or another. In my mind though, assuming what he said is correct, their system would act like a piggy-back system albeit within the ECU itself.

The auto-tuning bit is where I really started to feel like someone was trying to pull the wool over my eyes and this is the main crux of what I wanted help with. I asked them if they installed a wideband O2 sensor to the bike and made the ECU work in a closed-loop permanently and they said they didn't install anything extra on the bike.

This is what the guy said

"We don't do anything the 'normal' way, so you'll have to forget about trying to compare it to anybody else. In a nutshell it is possible to derive air velocity and density from the standard sensors on the bike and using other inputs (throttle position, temperature that kind of thing) we calculate fuelling on the fly."

I did press him for a more thorough explanation but he said: "...this is our living and so we never reveal anything that could allow someone else to copy our system."

So my question is this, given stock sensors on a bike, i.e. no MAF, is it possible to correctly calculate the amount of air entering the engine for the entire RPM range.

What I believe right now is, that it is only possible for the idle and cruise areas of operation using the MAP sensor and at high-load/high-RPM you could only revert to an Alpha-N style fuel table. The only way to around that would be to use a wideband O2.

Am I talking rubbish or is he?

It does seem to be techo-babble BS - I suggest you see what others have found when they tried the products this company is offering.

I expect there will be a BMW m'cycle club or three that can give some informed comments.

Hi Gord,

That's the thing, they say their software works on any machine and I see many people have said how their bikes 'feel' loads better afterwards. I've also seen dyno printouts that suggest the bikes are making more power but there is never an AFR graph on the printout. The company in question said they are happy to show this information to customers directly but won't offer anything to non-customers who are just interested in their solution.

My issue is, watching the content here, which I put a lot of stock in, it seems to me that directly measuring the air entering the engine using a MAP sensor is not possible with ITBs. You would need to take readings after the fact utilising a wideband or am I wrong about that.

Thanks for the response.

It's possible to do some self-learning with a narrowband O2, the problem is your learning would require working around the air/fuel ratios the narrowband sensor can measure. If the narrowband is essentially accurate near 14.7 AFR and then a wild-guess as you get much richer or leaner, the self-learning might slowly remove fuel until the sensor measures 'leaner than 14.7' and then add a large percentage like 20% or 30% on top of that to hit the richer AFR target you are hoping to actually achieve. It might be safe to perform this sort of method on a production car at part throttle with mostly factory parts, but I wouldn't personally feel comfortable doing that to a modified engine at high RPM or full load.

Remember that the OEM wants that bike engine to run safely, and it's very common for motorcycle guys to take their bikes to racetracks and drive fast up up mountain roads so the stock programming is likely going to handle that OK on a bone-stock bike. Using a MAP sensor and Throttle sensor combined should get the OEM plenty of into to run a stock engine well, even with individual throttles. If they also have a pressure sensor in a single intake runner (after the throttles), checking the pressure pulses in that runner carefully should get the OEM ECU a lot of information for how the engine breathes.

Agreed, did a little more digging and it seems that some of these 'tuners' do little more than remove, or alter, the limits in throttle opening in the lower gears and mid-range with some folks who have done before and after dyno' tests showing the same power after the chipping. However, that is going to be dependent on the actual company offering the 'yuning'and some may actually show an increase - going to have to check those you are actually interested in.

As for this company, as per the above, power/torque increase from improving breathing is most unlikely to be extrapolated from the listed sensors. Worse, as narrowband is normally low rpm/throttle, and some mod's will actually reduce the efficiency of the engine under those conditions, the engine is more likely to be running compromised fuelling and ignition when those improvements do start to take affect.

Some aftermarket suppliers will actually have a re-flash/map as part of the product package but the best thing is probably to do all the mod's you are planning to do and get your bike tuned to your specific setup - especially as modern high-performance motorcycles are already highly stressed machines.

Sounds like they flash an Alpha-N based bypass algorithm into the stock control unit.

Thanks for all the comments on this post, its really helpful information.

Just to be clear, I've sorted my bike with a Power Commander. as I said, on that bike, I'm not looking for ultimate power. I just wanted safe fueling given the switch from a stock system with a cat to a race system without a cat.

After speaking to this guy I just wasn't convinced about what he is selling. After he was vague he started to basically say I didn't know what I was on about because I don't tune bikes for a living. Which is true I don't and I just wanted a bit of confirmation that my understanding of this particular type of tuning was correct, which I think you guys agree it basically is?

It will be a "known safe" tune they have done with more comprehensive measurement gear on that engine with pipes/filters. They will be relying, at least at full load on manufacturing tolerances of the engines/limited variance is very open pipes. I'm sure if you chucked a helmholtz resinator in the exhaust you could generate a kick in the afr trace despite their majic software, almost certainly you could make the whole top end rich with an exhauat baffle too.

If you could do off stoic learning (beyond the hold it at lambda briefly then add 18%, dodgyness) with that hardware the manufacturers would be doing.

Slides - that's my point with it really. If this was possible why aren't the OEs doing it?