[OpenFlash]

Just got back from Bimmerfest and I'm beat. Just thought I'd throw my 2c in regarding diagnostic invisibility/visibility with respect to tunes (piggyback or otherwise).

First let's start off with piggybacks. As they say, we've come a long way baby. First generation piggybacks simply attenuated the map sensor and allowed the ECU to do the rest. By "the rest" I mean raise boost, add fuel through closed loop short/long term correction and retard timing through adaption. Clearly they made power. But they left a few tell-tales signs. Signs that were not officially interpreted as 'proof of tuning' at the time. But could very well be in the near future. Here's a list of the issues and the telltale signs of tuning associated with them:

1) MAP attenuation
There are two MAP sensors. One one each side of the throttle body. The downstream sensor only reads up to 4-5psi of boost. The upstream one up to 22psi of boost. Unless you are adjusting both signals to jibe with each other, you risk an implausibility code. There was no such code in the past but times are changing.
The other solution is to only adjust the upstream sensor (the one that actually dictates the boost target) when the downstream sensor is maxed out beyond its response range (above 4-5psi).

2) Raising boost
With the MAP sensor adjusted to read lower boost than there really is, the ECU sends a command to increase boost control solenoid duty cycle. All is good until the ECU starts adding in more duty cycle than it believes to be plausible. In the past, this would cause a limp mode caused by not achieving the boost target within a plausible boost duty cycle. Soon, it may be conceived as proof of a tune.
The solution is to keep the perceived boost duty within the range of a stock car running stock boost while driving the solenoids directly at whatever necessary duty cycle to achieve the desired boost target.

3) Fuel Enrichment
Historically, piggybacks achieved this by biasing a circuit on the wideband sensor. This would tell the wideband sensor that the AF mixture is leaner than it really is. As such, the ECU would add in fuel to compensate. History, there was no code associated with this perceived "lean-ness". So it was/is common to map the wideband bias to induce all the positive fuel correction the ECU could deliver. Unfortunately, these gross positive fuel corrections will soon be intepreted as proof of tuning.
The solution is to find another way of fuel enrichment that does not involve big positive short/long term corrections which result in implausible injector on-times. The ECU knows how much fuel it takes to make 300bhp/300lbft and it will interpret any big deviance from that as proof of tuning.

4) Timing retard
Running more boost with no proactive changes to ignition advance will force the computer to actively retard timing. Just as it would if you ran your car on 87 octane. Seeing excessively negative ignition trims could certainly raise some eyebrows.
The solution is to actively retard timing so that the factory ECU doesn't need to. Of course, doing so would have to involve proper processing/reproduction of the crank angle signal so that there are no losses of sync or missing teeth.

So it boils down to this:

To have a diagnostically invisible tuning sytem, you have to keep the following parameters "stock-like":
1) Map signal in (perceived manifold pressure)
2) Boost duty out (ECU drive)
3) ST/LT fuel correction
4) Perceived Ignition advance



As far as re-flashing goes, I suspect Siemens/BMW also made it hard to hide discrepancies induced by either ROM editing or the flashing process itself. Those who have been inside the ECU, and are willing to share their knowledge, would have more to offer on this subject. My experience with flashing Mitsu and Subaru ECUs might not be relevant here. Good knows not much with regards to N54 tuning has been relevant anywhere else!

Needless to say, this is an exciting time for everyone. BMW is certainly throwing us some curveballs. It's up to all the players to aim for the fences or just sit out the rest of the game.

Cheers,
shiv

[cgillio]

Nice write up

[Amir87]

A good read!

[5soko]

nice stuff shiv.

[e90AW335i]

+1. Another great thread by the Shivster....

[Mondolemite]

Thanks Shiv, that was helpful. I may still be sitting this one out for a while since I have a later build car.

(I'm also sitting it out because I don't have the car in my hands yet but when has that ever stopped bimmerphiles from buying mods?)

[ezatnova]

Do you know any more as to whether pre-March 08 cars should be worried about PROGMAN updates? Is this an issue at all for pre March 08 cars with the V3?

[e90AW335i]

Originally Posted by ezatnova

Do you know any more as to whether pre-March 08 cars should be worried about PROGMAN updates? Is this an issue at all for pre March 08 cars with the V3?

From the information out there from creditable sources the new v29.2 progman update will detect torque output and thus tell if there is a piggyback....

[sparkyg]

Thx Shiv, good read.

[OpenFlash]

Originally Posted by e90AW335i

From the information out there from creditable sources the new v29.2 progman update will detect torque output and thus tell if there is a piggyback....

From the sounds of it, the new software monitors fuel demand and estimates torque output. And yes, I believe there is a way around that.

Shiv

[ezatnova]

Originally Posted by shiv@vishnu

From the sounds of it, the new software monitors fuel demand and estimates torque output. And yes, I believe there is a way around that.

Shiv

May the force be with you in figuring this one out. Man, what a bummer.

[e90AW335i]

Originally Posted by shiv@vishnu

From the sounds of it, the new software monitors fuel demand and estimates torque output. And yes, I believe there is a way around that.

Shiv

I knew you would figure out a way. I was just saying if someone takes their car in and it gets updated with v29.2 there is a chance that their current piggyback will be detected.

Keep us updated on the loophole when you figure it out.

Good work as always.....

[remydlc]

Originally Posted by e90AW335i

if someone takes their car in and it gets updated with v29.2 there is a chance that their current piggyback will be detected.

shit's scary.

[e90AW335i]

Originally Posted by remydlc

shit's scary.

It is what you would expect from a company that doesn't want to pay for broken shit caused by a piggyback.

Still sucks though....

[ArmyBimmerDude]

Originally Posted by shiv@vishnu

From the sounds of it, the new software monitors fuel demand and estimates torque output. And yes, I believe there is a way around that.

Shiv

Yeah, I figured you would be coming up with something.

[OpenFlash]

Originally Posted by ArmyBimmerDude

Yeah, I figured you would be coming up with something.

Already have come up with something Just testing it right now.

shiv

[bronco]

Originally Posted by shiv@vishnu

Already have come up with something Just testing it right now.

shiv

hehehehe NICE

[2008335icoupeblack]

Two questions:
1. Is the workaround tha you are working on going to be software based or will it require mre hardware (like another wire kit) to work?
2. What is your suggestion for those of us with procedes that go in for a service that may end with the new program being updated in our pre march ecu? Do we reinstall are procede or do we wait? Is there going to be a way for you to see whether this hidden code has been tripped ... and once it has can it be removed ?

I guess that is really more than two questions... but hey,,, who said I could count anyway

Thanks in advance!

[Mondolemite]

Originally Posted by shiv@vishnu

Already have come up with something Just testing it right now.

shiv

How will you know if it "works". In other words, how are you testing the invisibility of a piggyback?

[e90AW335i]

Originally Posted by Mondolemite

How will you know if it "works". In other words, how are you testing the invisibility of a piggyback?

Get one of the BMW techs to update his car w/ 29.2 and then find a workaround....

[oc24]

Originally Posted by shiv@vishnu

From the sounds of it, the new software monitors fuel demand and estimates torque output. And yes, I believe there is a way around that.

Shiv

Originally Posted by shiv@vishnu

Already have come up with something Just testing it right now.

shiv

Shiv, I believe fuel had something to do with the new DME too. Is this workaround applicable to the new DME as well or just the 29.2 progman?

[OpenFlash]

Originally Posted by 2008335icoupeblack

Two questions:
1. Is the workaround tha you are working on going to be software based or will it require mre hardware (like another wire kit) to work?
2. What is your suggestion for those of us with procedes that go in for a service that may end with the new program being updated in our pre march ecu? Do we reinstall are procede or do we wait? Is there going to be a way for you to see whether this hidden code has been tripped ... and once it has can it be removed ?

I guess that is really more than two questions... but hey,,, who said I could count anyway

Thanks in advance!

1) Yet to be determined. But I suspect it will be integrated into the PnP harness. We may update the v2->v3 wire pack to include the necessary loom as well. Will know a lot more by the end of the week.

2) For now, if you go in for a service, request for them not to update your software. If they can't honor your request, don't go in for your free oil change. Do it elsewhere As for the code, it looks like it is viewable at the dealership level (and presumably by the end user as well). Again, we'll no more in the next few days.

Cheers,
shiv