They veeeeery carefully map it out. They test it in as many altitudes and atmospheric conditions as they can. Mind you, they have to use the same tune for EVERY car of a particular engine they make regardless of the climate in which it will be used, and that doing so results in over 1,000 maps for every conceivable environment. (Of course, T7 also has an electronic and adaptive throttle, which requires a few extra maps.)

You could probably get away with much less testing if you know the climate in which you'll usually drive your car.

T5 (older ECU 1990-1998) is much simpler and has a much larger margin of error as far as EGT, but it still does something similar. You could probably mimmic the T5's setup with the Evo 9 ECU.

Disclaimer: I'm crazy and you listen to me at your own risk. Nothing I say is gospel, just suggestions and thoughts that you can apply for yourself at your OWN discretion!

If you have an EGT probe, you could tune for the EGT something like this: start at the lowest psi you could imagine knock retarding to and adjust the timing until you reach whatever you decide is your EGT limit. Then raise the boost in small (.5-1 psi intervals for accuracy), and do the same until you get to a point where you cannot raise the boost without hitting knock, and cannot pull or add timing without either hitting knock or reaching your EGT limit. The tricky part is that you'd need to do this for as many load points as you can AND ... you'll still want a good margin of error so that you are not pushing the knock threshold constantly. (Only Saabs can do that. Neener neener!)

Anyway, I just thought it was something worth bringing up. A lot of Subaru tuners are much more conservative with the EGT than they need to be because they cannot lower boost with knock. So, if the IAM drops too low, they need to make sure it won't cook the exhaust valves. Now that the Evo 9 ECU has this ability, I say use it!

-Adrian

 

That sounds very interesting indeed.

The factory EMS I'm working on uses a mathematical model to predict EGT. When EGT reaches a critical threshold, fuel is added until the model output temperatures have come back down.

This sounds equivalent to the SAAB system but instead of look up tables, we're using a model.

BTW, any links with technical insights into the SAAB system you could send me?

 

A mathematical model and a large number of maps amount to the same thing. Except that, when mathematically modeled, the ECU has to "calculate" the values for the fuel maps needed to accomplish the lower EGT. The Saab ECU essentially follows what COULD have been turned into a mathematical model. But, all the output values for that model are stored as maps so that they don't need to be calculated later. Sometimes I could see mapping it out being more reliable as it's easier to crash a "thinking" process than a "regurgitating" process.

No interesting Saab ECU links really. The information I gave has was from a friend who works with Saab's lead ECU programmer.

I do have an interesting link about the workings of Saabs rather unique ionized-gap knock sensing system: http://www.vehicular.isy.liu.se/Publ...PhD_580_LE.pdf

You could, theoretically, use the ion-sensing system to actually measure the EGT of each individual combustion. But that's rather off-topic in an Evo9 ECU discussion thread.

Anyway ... you could make a mathematical model for a particular setup's EGT which people could apply to their maps if they have a similar setup, even if they don't have EGT probes. That could turn out fairly cool if done properly.

-Adrian

 

Is it possible to use Saab's ionized gap knock sensing system in a Mitsu standalone/piggyback ECU?

 

Not with your stock ignition system. You'd need a very special direct ignition system and an ECU capable of running it. The Saab DI rack is a single-piece and isn't spaced right to fit the Evo. Harley uses the Delphi system, which has individual coils which could be spaced to fit the Evo. Read more about that here: http://www.delphi.com/pdf/ppd/pwrtrn/gas_ignics.pdf

No idea how you'd hook that up to any of the existing EMS systems.

 

Why is a Saab Tuner so involved in this Evo (tuning) thread

Just curious

The Viggen guys I meet at the track make it a point to be very uninterested in Evo's

 

Sorreh

 

There are a couple coil on plug ignition systems out there. With a standalone EMS like the AEM, it's easy to drive something like this, but what I'm more interested in is what you do once you have COP set up in order to monitor knock through the plugs.

Worth some research, I'll read up on that pdf you posted and get back to y'all on some ideas. You should be able to get it done, but it's just a question of how.

Meanmud, quit hating. This is one of the more informative posts I've seen on here in a while with Shiv and him sharing ideas and it's always good to get a fresh voice on here to talk about tuning possibilities that are a bit outside the box. Besides, it's better than all the other posts on here that bump a thread because it once had some shiny pictures in it or talking about which blowoff valve or exhaust is "the best."

 

No, unless your standalone has the software on to use the ionized gap detection signal.

I believe that Harley uses a neural network to process the signal to determine if knock or misfire is present. I can dig up a technical paper on it if you wish.

It's not a simple threshold that you're comparing to, but you're looking for specific characteristics in the voltage/current trace. It's cool stuff. I wish I were working on something like this.

Does anyone know how the stock EVO ECU detects misfire? Does it use crank based signal processing?

 

Does anyone know how the stock EVO ECU detects misfire? Does it use crank based signal processing?[/QUOTE]

YES

 

The stock ECU diagnoses misfire by constantly monitoring the angular velocities of successive crank rotations. If it falls outside a certain window, it cries "misfire!"

Shiv

 

SaabTuner-- What kind of output does the ion detecting system feed into the ECU? Is it a 0-5v signal? Frequency signal?

shiv

 

I was wondering, how is taht mivec tuning coming along? Something rather like honda stuff?

 

I was only curious - no need to be Sorreh

 

I'm not sure about the Voltage. I think it's 80v in the Delphi system. The ion "signal" however, is measured by the current flow across the plug gap during the combustion process. Idealy, it looks like this:



A knocking signal looks like the blue line in the image here: http://img.photobucket.com/albums/v1...ngIonTrace.jpg

In the paper I linked-to earlier, on spark advance modelling and control, they were able to extract the burn angles for their test engine and continuously maintain the PPP @ 14* ATDC to keep the engine at MBT even with varying burn rates. It offers a number of possibilities for heavily modified cars which might have different burn rates since you can dynamically maintain MBT through a feedback system.

Anyway, I would definitely give that Ignition Modelling paper a read. It really gives one a good background on the how's and what's of ion-sensing.

-Adrian

p.s. Until recently, Saab's DI racks weren't just any DI, they were capacitative discharge. The "red" DI's used a thyristor to control the discharge, and the "black" DI's used a triac. Ridiculously high ignition voltage as a result. Crazy swedes.