NefMoto

Hello,
Fairly new to this so here I am seeking guidance... Car is a 2001 Audi Allroad with the 2.7T APB engine attached to that "amazing" tiptragic transmission.

Im wondering if anyone has a road map / hardware guide to the 7.1 ECU.
My goal here is to replace the primary O2's with wideband units ( yes im aware that I will need a controller ) and add an octane / ethanol sensor

typically, people will code out the rear o2 sensors and put widebands in the place of the rear.  You can feed the WB signal into the rear o2 sensors for logging purposes.

To try to replace the primary o2 sensors with a wideband wouldn't really benefit you and would likely require a lot of reverse engineering to make lambda control work at WOT as it does in wideband ecus. 

I have 3" Catless downpipes on the AR...

Idea was to code out the rears as well as disable the Catalyst Efficiency Monitors and a slight tune to start.

However the idea is to unlock the full potential of the ecu which as much as it sucks that no one has fully figured out the ECU.
I know the GM boys have HPTuners suite that allows you a ton of flexibility and tunability compared to our ECU's which are still a grey area from what i see so far.

Big plans.
Could you elaborate a little bit on your background? How is your knowledge of assembler?

Also, these ECU's are freely tunable. The full factory descriptions as well as full map descriptions are available.
They are just an order of magnitude more complex than jap and usa ones.

Im a certified Automotive service technician with a little over 20 years of experience working on the VW's and Audis owning a fair shar of both over the years. With that being said I have some experience with domestics as well...

Plans for the AR are fairly standard...
I call it Project Stealth Wagon

Drop the tip out and replace it with an 01E, K04/RS6 hybrid turbos, AWE SMIC, AWE bipipe, 2.8L heads and cams, tubular manifolds, list goes on...

My biggest concern is the fuel around here... with extremely cold winters i still want to be able to DD the car. With that being said the fuel quality varies with additives and ethanol content. therefore these options really can make or break an engine out here.

I want to retain the stock EGT's , replace factory primaries with widebands and as mentioned add an octane "self tune" based on fuel rather than swapping ECU's based on fuel type. Tenatively also one of the inputs could be used to switch boost tables between street / track modes.

I know i still have a long way to go and lots to learn...

There is no way you can easily convert ME7.1 to wideband without extensive assembly experience, which you did not mention you had.

Same with the "self tune" based on octane, if you intend to completely rewrite the knock control code.

You'll definitely not be able to alter boost based on octane, unless you are talking about flex fuel.

And then we are back to the first question about assembly.

why does everyone think the 2.7t is so legendary? its a 90* wedge shoehorned into a engine bay thats too small. power per footprint is really crappy. i honestly would rather have a vr6t

Lots of talk then without any clue how the ECU works nor any actual code knowledge.
Good luck.

Well... lets disect this:
 

 - thats why I'm here to learn and hopefully spark some interest and get a few people on board as more heads are better than 1


 - idea here is to try and keep fueling in line with ethanol content as every station out here as different "average" range which in fact is sorta like flex fuel. the boost would stay stationary however the timing can be adjusted to be more aggressive. As for altering boost i want to use a unused input to change the boost level to a secondary boost table on the fly.

As far as the coding goes...
I been dabling a little bit but can barelly consider myself as an amateur at best.

As far as the ECU goes i quite versed on the hardware side as I do most of the diagnostic in the shop i work in and undrstand the I/O quite well. My original question was though if anyone has figured out the min/max to every I/O to the ECU and where is it mapped to.
'
Factory ECU's have came a long way and they have more than enough "muscle" than their predecessors when the maps were loaded on ROMS which were soldred in and the ECU's had to be socketed and run emulators like the Ostrich to tune and then burn a coresponding rom for each specific fuel / even that they were competing.
So yes... technology came a long way!

If you are comfortable with diassembly and reading the FR, as a beginner you will likely be able to code around enough to find a spare input to use as a wideband signal for logging, and presenting that value to me7logger.

but altering ME7.1 to be wideband would be a nearly impossible task, even for an expert, and for negligible gain.

flex fuel can be done with multi map switching and finding a spare input for the ethanol sensor.

but thats more yes/no than the much bigger question of detecting varying levels of ethanol (vs e85/gasoline detection).

Its easy to come up with ideas. When you need others to do the coding for you it will be expensive and if you do some digging youll see youre not the first whos had this thought. Anyone whos able to make such mods surly will not have free time for a open project like this. This isnt romraider

Dissasembly / Coding is a good skill to have... Specially when your messing with your own vehicle your forced to learn quicker and make less mistakes because thats a money out of your pocket LOL

Either Way...
I already have a bench rig ready as well as a Galleto 1260 Cable ( Just in case i FA** it up royally ) and have a pretty solid background in electronics ( have my own reflow oven and hot air station ).
Also bought a spare ECU to mess with...

I guess option B would be to use the simulated narrow band output as primary input and route the wideband signal to the secondaries for logging. I guess for fuel sensor tap a 5V feed from one of the other sensors and use a unused input to feed signal to.

Has anyone messed with speed density tunes on these ECU's ??

An alternate solution would be to leave the narrowbands in place and add a wideband bung, or code out the rear o2s and put wideband(s) in the rear bung(s)

Note that ME7.1 doesn't really like simulated narrowband output much. YMMV


Yes, but most that claim to have done so (incorrectly) think alpha-n is the same thing as speed density. True speed density on ME7.1 would be as complex to implement as wideband O2 support.

its not much of a simulation rather than a scalled output of the actual wideband output. I have a pair of LC2'shttps://www.innovatemotorsports.com/products/lc2.php (http://LC2) with the Bosch LSU4.9 sensors waiting to go in...



Fair enough... I guess your not making any headway unless your blowing shit up LOL

I don't really see the point, honestly. The auto trans can't handle more than about 400wtq anyway, which you can get to trivially with entirely stock O2s, MAF, larger injectors and a pair of K04s.

This entire thread can be summed up as blowing hot air (pun intended).

I'll be damned if anything ever actually comes to fruition.

First thing - discovering small variations of ethanol content (like around ~5%) - this is what lambda sensor is for and I cannot imagine anyone to be bothered to adjust other running parameters (ignition / boost) based on that. Even better, I have seen cases of desperates running E85 on stock maps and relying on long term fuel trims to adjust to the new fuel.

About doing code changes: I do not see how converting to wideband is "almost impossible", I have done it on an IAW ST10 based ECU, was not that much work - I used original lambda flag for enabled-ness of correction and added a simple proportional controller over what in Bosch would be rkte I guess (the required fuel mass). ME7 have very well defined task intervals and making it into a full blown PID controller is also doable (which I have also done, but for a boost controller). It does require serious programming skills, doing serious disassembly and finding the places in the code to hook into, but I would not say this is much more difficult than doing a 5120 hack properly, if not easier in some sense.

Proper flex fuel I have also done, and it would have been reported on here already, if not for the fact that I have problems getting my shit together to do live car tests. I have a full direct sensor plug-in implementation on my bench ECU, which I have dry tested inside out, it has an interrupt procedure to measure pw-s for frequency and fuel temperature straight from the sensor, DTC reporting, limp mode on errors, and what not (hat off to prj for some ideas and hints). Skill wise, this one certainly goes beyond 5120 hack / wideband addition, requires going through the whole code to get the interrupt priorities right, knowing the CPU interrupt architecture, and all that. Would not advise it to anyone that wants to stay sane. There are of course solutions possible with intermediate hardware that prepare an easy to handle 0-5V signal, it should be then a tad easier.

But all this is indeed a very very very long road, I can promise you that. I did my first disassembly in 2007 IIRC, "simple" (have not thought of them as simple at the time) code patches in 2009/2010, and more serious stuff (PID based boost controller added to an ECU) in ~2013. That should give you the idea of effort required.

Would not have even start to write all this if I did not have a question for nyet: which particular cases of ME-s and narrow band simulations do you know of to cause problems? I am yet to try Zeitronix to feed narrowband signal to my ME7.9.10 ECU long term (I did some experiments last year, but only stationary and on idle), I never had problems with it on other ECUs. I have heard (from unconfirmed sources though) that Innovate is not so good for this. So I just want to know if this is general truth, or particular setup issue.

Innovate and spartan both do not work correctly, both have impedance matching issues. Not sure about zeitronix.

OK, so there could be an issue, because during my experiments I got "P0130 - Intermittent lambda signal disruption, short to GND" (at least that's what MultiEcuScan said), I attributed this to power cuts to Zeitronix (had it connected to the lighter socket then) when cranking. But impedance sounds to me like ECU picking up shorts / not seeing the circuit. I also noticed that the signal was more jerky / discrete, while the one from the stock sensor was floating nicely. My only preference to plug it in instead of the pre-cat lambda was because of accessibility, will have to rethink that...

Yep. The problem is the ECU has very low input impedance to mitigate noise/crosstalk, and neither the spartan or innovate can drive enough current. Somebody would have to fab up a simple op-amp driver circuit, but for most people who just want it to work, that is just another headache.

I've had no issues with the LM-1...
Zeitronix seemed bang on as well.

Meaning narrowband o2 and wideband o2 agreed perfectly in both cases.

Another option is to use a different ECU.

After some research and a bit of digging I found that the ME7.1.1 from a B6 A4 ( 3.0L ) has dual primary wideband sensors, however it will require the matching harness a bit of rewiring will be needed to make it work. This may be a better suited option as it would not only allow me to do away with the Ignition drivers that are prone to fail on the 2.7 but also would be less of a "hack the harness" deal. I already removed the SAI and secondary O2's as I have installed 3" catless down pipes. i will have to dig around and see what I can find and start hoarding parts and go from there...

I have an RS6 rear subframe with the axles and differential headed my way as a starting point for my 01E swap, and currently may have found an 01E that fits the bill butmay need some fresh syncros and a quick look over.

The only reason why I mentioned it's auto is because I'm planing to start this process now rather than wait after the conversion...

3 liter engine:
1) has continuous cam adjustment on all 4 cams.
2) is NA

Good luck making that ECU work properly on 2.7.


Seriously, this whole obsession with widebands has no practical meaning for a moderately tuned car. Hell, even 600+hp stage3 cars get away with narrowbands somehow.
The stated concern about "being able to DD the car" looks even weirder in that context - because "normal driving" is exactly, where NBO2s operate fine: ECU will mostly try to run at stoich.
Everything that you mentioned can be reliably achieved with stock(ish) ECU configuration.
You can just have an external WBO with a gauge, if that makes you feel safer - but properly tuned car won't really need that.


Regarding all the other requests/questions:
Speed-density for ME7 series has been done, it's relatively easy to implement at this point (won't be very cheap, though).
Map switch for several pre-sets - even is posted here for free.
Map "blending" based on ethanol content sensor - again, can be done.
Hell, you can also have WBOs if you absolutely insist. But:
- It'll be a hack, relying on external controllers, since 7.1 lacks hardware for that.
- You'll have to cover development costs, which are going to be quite significant for something, that basically only you want & could use.

i just tuned srm hybrids on stock narrowband sensors today. used wbo2 for logging. made 517whp had no problems

Totally doable, it's actually on my list to implement WBO2 -> 2.7t softwares.. :P

Just craves for alot of time in terms of custom code, but can certainly be done ;)