Saabscene Saab Forum - Saab Technical Information Resource banner
1 - 13 of 13 Posts

·
Registered
Joined
·
5 Posts
Discussion Starter · #1 ·
First post on this site and wanted to say hello to all.

I'm an engineer in the US in the process of collecting stock .bins for T7 and thought I'd check on your side of the pond if anyone was doing the same.

Is anyone working on T7 here? Has anyone else (other than the current retailers) read a 29F400? I imagine there are differences in the calibrations here/there, it would likely be a good learning experience to take a look at those differences.

Have a great day and send me a message if you'd like to coordinate.

 

·
Registered
Joined
·
14,305 Posts
Hi Agent9 and welcome .

AFAIK the only folk this side who have read T7s are the aftermarket tuners, who probably won't be that keen to release what they know!
 

·
Registered
Joined
·
5 Posts
Discussion Starter · #3 ·
Thank you for the welcome!

As near as I can tell the situation is the same over here in US.

In fact, posting what I have already could might-possibly get a mark put on my head.

Here's one of the many 16x16 "maps" from a T7 lpt:

 

·
Registered
Joined
·
5 Posts
Discussion Starter · #5 ·
Well there's the problem!

I'm certainly in the infancy of "tuning" these things--and by comparing the differences between data I can more easily pinpoint where what controls are located...so I posted here to hopefully gain another source of files that I'm not already exposed-to.

Anyway, compared to the drivers of other car makes, us Saab guys have it pretty bad because the options are so very limited--and that's on purpose. Take the Nissan 300Z for example, their coding is open source. Imagine having your friend do a few long pulls down the road while you sit in the passenger seat and adjust your timing/duty cycle/whatever, on your own...imagine not being tied directly to a tuning house that has set stages for set software...imagine going to a junk yard and pulling the MAF off of another car, or running whatever high-flow injectors you can find for a good price...all for the cost of the chip and some burning equipment, adpters, and time of course. Imagine having custom software for your very own car that you tweaked all by yourself to match the hardware that you darn felt like bolting-on. You should see what Honda drivers have available to them!

So to answer your question, what is the above graph? I don't know exactly (yet). That's why I'm writing here.

Here's a great article that describes what tuning actually is at a base level (certainly T7 is more advanced than the Z stuff, but very similar concept):

ECU Basics

As I see it, yes, certainly there's an art to it, but ACCESS and data location are the main hurdles.

T7 uses an AMD manufactured AM29F400BT90-SI 4Mb chip to store the engine calibration files. Devices to read/write those chips can be had for a few hundred.

BTW, the picture of the ECU at the top is of a Saab T7 unit with the AM29F400 removed (top right corner).
 

·
Registered
Joined
·
14,305 Posts
I know a little, (dangerous eh?) and it goes something like this...

Basically the 3d map shown correlates desired boost against revs and engine load. The ECU tries to match them up.

"Engine load" as I understand it is a very complicated parameter. It's not just about load as we might understand it but also takes into account throttle position etc. You can produce big increases in top end performance but also improve fuel economy by increasing the max desired boost whilst backing off the lower load desired boost.

Of course this is a very simplistic view- you also have to take account the APC functionality. It always used to be the case that Trionic was the most advanced, completely integrated, performance management system. I don't know if that is still the case, but it might well explain why it's not so easy to read.
 

·
Registered
Joined
·
5 Posts
Discussion Starter · #8 ·
You have to make sure you're thinking of the control as independent devices.

There's the brains (cpu) where the logic resides and then there's the storage area, where the calibration/map data is kept. The CPU collects data, (the sensor inputs), thinks about that data, makes some decisions, and then based-on those decisions fetches the data from the storage area, which then determines the magnitude of ouput in response.

It would be massively complicated to de-compile the entire logic in the CPU...which may ultimately be necessary, but I don't know for sure and don't think-so at this stage. In any case, it would be very interesting to take a look at the interface of one of those PPC devices. The way they were likely designed was by doing a data trace on the communications between a TechII and an ECU and then reverse engineering the communications protocol...amazing, but still much more simple than de-compiling the CPU logic itself.

As for the meaning of the data itself, at root, the x and y coordinates are actual data address on the chip itself the z direction is the magnitude of the value contained in that address. To actually derive meaning out of the data, you have to figure-out what data is represented in the x, y, and z directions. There's a couple of ways to do that...and decompiling the cpu (if it is a mass produced unit with info available) is one way to do that.

There's a bunch more to it, but once you spend some time looking at the raw data it isn't difficult to pinpoint the data addresses of some "maps" pretty quickly, and there's a bunch of them that are accessed for different reasons based-upon different operating conditions that control different things---but are basically just numeric patterns with the "3d" portion a simple result of address locations/data magnitudes---there are also 2d maps, just strings of numerically increasing data in one dimension address-wise. The calibration data? That takes another tactic to decipher.
 

·
Registered
Joined
·
1,448 Posts
I think the tuners would be wise to dish out what information they have on how T7 works and only to keep the actual process of how to tune it proprietary.

I also seriously doubt any of the tuners would lose any money in the long run by making public how the T7 ECU works. If anything, I think it would give self-tuners respect for the professionals as they discover how hard it is to get the horsepower numbers they want.


People in the Subaru-tuning community still pay a thousand dollars or more to have an expert tune their ECU even when options like UTEC and Street-Tuner exist and would allow them to tune their own ECUs.

And what the Saab tuning community needs now is an infusion of new consumers before this tuning secrecy stagnates the growth of the consumer base. It isn't much "fun" to be totally dependant on some tuner simply because you do not even have the option of trying to do it yourself.


Anyway, sadly none of that is likely to happen in the near-future. But it sure would be neat if it did.

-Adrian W
 

·
Registered
Joined
·
16 Posts
hi there,
T7 is quite simple,
you request an airmass per combustion via the throttle body, which refers to an airmass request map. t7 tries to meet this request via the throttle body if it cannot acheive the request with the throttle alone it allows the turbo to increase boost pressure untill the requested airmass/combustion is reached.
Fueling is simply a calculation based on airmass/combustion and injector size and is therefore not mapped. However this calculation is based on 14.7:1 so a corection factor is also used on load to alter the ratio, this correction factor is mapped.
Ignition is mapped, with correction factors for air temp etc.
This may be a little simplified.
bye.
 

·
Registered
Joined
·
14,305 Posts
Welcome to Saabscene PJ .

I'm not sure I understand what you mean about T7 requesting an airmass from the throttle body, controlled by the throttle, as in ptractice the throttle (accelerator pedal) is controlled by your foot.

Or are you saying that if their isn't the right amount of air flowing through against the preset map for for throttle position, rpm and engine load, then Trionic varies the boost?
 

·
Registered
Joined
·
1,448 Posts
He's just basically saying what the WIS says in the first part; you request torque with the pedal, the ECU tries to meet that torque first with the throttle then, if that isn't enough alone, with boost pressure beyond base-boost. (You can get base boost with nothing but throttle.)

The latter part is just a different way to say that the ECU has a renormalization map and divides by the lambda to get the quantity of fuel to inject. (Less lambda = more fuel, an inverse relationship.) Since a narrow-band O2 sensor can only sense the lean/rich transition, you can only use it to renormalize to 14.7:1. However, with a wideband, some ECU's renormalize to a specific fuel curve even during full-load.

Fairly well known information though. The hard part about T7 is knowing what each of the thousand or so maps is, and what to do with it.

-Adrian W
 

·
Registered
Joined
·
16 Posts
hi again,
what i ment by the air request via the throttle body is that the throttle cable is attached to the throttle body which is where the ecu gets its signal for the airmass requested although the butterfly is not directly controled by the cable.
bye
 
1 - 13 of 13 Posts
Top