Resolved SAFC, how does it work

[Enigma_Man]

I have a question about the SAFC.

I understand that it "intercepts" the MAF signal, and modifies it before it reaches the ECU to allow tuning and all that.

But why are the adjustments on it by RPM? Why would you want more or less fuel at one rpm than another? And say for example: if you added a lot of fuel at like 3500 rpm, and you cruise on the highway at 3500 rpm, far from WOT, would it still be adding the extra fuel?

I just don't understand why you'd want RPM based adjustments, rather than say flow, or PSI based adjustments.

-Jesse

 

[DrZiplok]

Originally posted by Enigma_Man
I have a question about the SAFC.

Good questions, too.

I understand that it "intercepts" the MAF signal, and modifies it before it reaches the ECU to allow tuning and all that.

But why are the adjustments on it by RPM? Why would you want more or less fuel at one rpm than another?

The design problem that the Apex folks faced is that a truly accurate correction map would be multidimensional. As you change the characteristics of the engine, you really need to take into account a whole bunch of factors.

Producing a $300 product that could be used (or at least purchased) by semi-literate teenagers and backyard tuners required simplifying the equations as much as possible.

As you do that, stripping away one thing after another and looking for the single most significant factor, engine speed is prettymuch a standout. Given a product design that evolved from zero factors (old AFC) to one (Super-AFC, rpm) to two (new Super-AFC, rpm and throttle position) you can understand why they've kept it simple.

And say for example: if you added a lot of fuel at like 3500 rpm, and you cruise on the highway at 3500 rpm, far from WOT, would it still be adding the extra fuel?

"Yes, but". Remember that the ECU will go closed-loop while you're cruising, and as long as the trims can compensate for the mis-correction the AFC is applying, you won't actually end up using more fuel.

A couple things to remember:

1) The AFC doesn't add "more fuel". It lies about "more air", which is not always the same thing.

2) While you remain inside the ECU's ability to compensate in closed-loop operation, you are only interested in tuning the open-loop maps with the AFC. Since these maps on the DSM (basically) just use airmass and engine speed to compute fuel mass, you really only need to tune with those two factors.

As an aside, one of the drawbacks of the AFC is that it only lets you tune the air volume signal, not the air mass signal. This is why you need to re-tune for different weather and altitude conditions, even though your car has both pressure and temperature sensors - the corrections that the ECU applies based on those sensors don't necessarily match up with reality once you move away from stock.

I just don't understand why you'd want RPM based adjustments, rather than say flow, or PSI based adjustments.

Flow-based adjustments would give you flow-curve correction, which I have to agree would be nice (although perhaps hard to tune). Some people do use pressure rather than throttle position to switch between the maps on the new S-AFC.

However, it really boils down to the simple issue that the S-AFC is meant for small adjustments and ease of use. For that, engine speed is more significant than any of the other factors.

If you want more adjustability, then there's a wide range of options; the next tier of piggyback units like the e-Manage or the SMT6 add more dimensions to the equation, but at a corresponding increase in tuning difficulty.

 

[DCJ98GST]

I am curious what the new SAFC II will have for additional features. I know it is going to have at least 12 rpm adjustments.

 

[Enigma_Man]

DrZiplok, thanks for the good explanation

I wanted to know, because I've been thinking of a way to design something along the lines of a combination VPC/SAFC deal. As is obvious, I don't have a lot of experience at tuning.

What I was planning on doing is setting up a speed-density (I think it is called) setup, like VPCs do, but instead of having to then take the VPC signal, and modify it again with the SAFC, I'd just make it so the signal is adjustable from the get-go, within the original unit.

How I was planning on generating the signal would be basically (neglecting coefficients) rpm*boost/temp.

To adjust everything, either the coefficients would be adjustable, or there would be adjustable maps based on the pressure, so you could add more fuel at a higher boost pressure to help stave off detonation, etc.

-Jesse

 

[DrZiplok]

What you're talking about there then is the sort of multidimensional correction map I alluded to earlier.

You're going to run into a number of problems, and they all boil down to the same thing; you'll be fighting the ECU. The DSM ECU is actually reasonably smart, which means that you have to take everything that it does into account when you're tweaking things.

That makes life pretty hard. I'd recommend doing a bunch of reading; the talon digest archives are a good place to start, and then just read everything you can find on the DSM tuning process.

The bottom line is that if you really want that sort of fine-grained adjustment you need to go standalone so that you're in complete control. However, there are a lot of intermediate steps one can take, and a lot of learning experiences to be gained from them first.

Map out what you think you want to do early on, and then read *everything* you can find on the parts and combinations you're considering. I've discarded countless upgrade paths based on other peoples' experiences, and I'm always looking several thousand dollars ahead just to make sure that where I go next makes any sense. 8)