Let's get to the bottom of this...

[BAC5.2]

OK, I want to get a definite answer.

Will an SAFC work or not?

Now, reasoning would lead me to believe that it does. My buddy Brent (Matt, Vikash, Scott, and Steve, you all have met him. Green 93 Impreza 1.8T) has an SAFC in his 1993 Impreza 1.8T.

He's OBD1, his ECU can't be much different than ours.

So I'd figure it'd work.

BUT, people say it won't. There are reasons I have to believe either way, but everything I have heard is all heresay, so I want concrete answers as to WHY it won't work.

I'm thinking about picking up an SAFC V2, but I'm not going to waste my money if it absolutely won't work. So I want to know EXACTLY why it won't work.

I want to know how our cars vary ECU wise to an Impreza.

Phil

[vrg3]

I don't think it won't work.

It's just a very crude tool as far as tuning goes.

For one thing, the ECU does make extensive use of feedback so you might have to tweak the signals far enough to force the behavior you want.

For another, modifying the airflow signal affects more than just fuel. It changes the ECU's computed load values, which affects lots of computations. I'd say the most important one is ignition timing.

So basically you can try to get passable results with it, but you can't really use it scientifically because you can't predict the precise outcome of your mucking with the airflow signal. In the process you also lose some of the benefits of the engineering that went into the stock ECU.

I'm sure Brent's car would run way, way better with proper engine control. From what I understand, he just put big injectors in and then used an S-AFC to bring fueling back down to appropriate levels.

What do you plan to do with it?

[BAC5.2]

Well, I.. *cough* you know... *cough* *cough*



Hmm. I also wonder if the FCD would cause a problem with the readout that the SAFC gets. It can function as a boost gauge, but I wonder if it reads what the ECU sees (which doesn't do much good, as it just would read 10.xxpsi).

Also, how does it change the load values?

Would this be a better option than leaving the ECU stock and letting it guess?

$300 is a lot more affordable than $1800.

[vrg3]

If you won't tell us what you want to do with it there's a severe limit on the quality of the advice you can get.

I don't know what you mean about leaving the ECU stock and letting it guess. Guess about what?

I believe the SAFC only can act as a boost gauge on cars that use speed density fuel systems. It just displays the reading from the sensor. On our cars it wouldn't show manifold pressure but it could display airflow.

It changes the load values by screwing with the airflow signal which is the primary factor in the ECU's load calculation. To extremely oversimplify, let's say you just got your hands on some 740cc injectors, which flow about 50% more than stock. And you decided you wanted to use an S-AFC to make them work with the stock ECU. So you just go and have the AFC decrease the airflow signal by 50%. Now you end up with the ECU thinking you're getting about half the airflow that you really are, so it should pulse the injectors for about half as long. But the problem now is that the ECU won't realize you're on boost when you are, and so it'll keep its enrichments conservative and its timing advanced. The ECU's timing maps dictate that the maximum permissible timing advance at high load is much less than the minimum base advance at low load.

Uh oh. That means you risk detonation on boost. So you better make the mixture richer to try to keep combustion temperature down. So you richen it up across the board, or maybe just at high RPMs and/or high throttle positions. That keeps it from pinging. But consider what the ECU must be thinking. There's no way it can operate intelligently in such a configuration. You've stretched the maps so the shapes of their surfaces are wrong. You've now placed portions of the map that don't have much adaptiveness onto places that normally benefit from adaptiveness. You've modified the mixtures to safeguard your engine, but if you ever allow any of that to happen on closed loop the ECU will try to adjust its fuel trims.

So you can get passable performance but you lose much of the adaptivity and intelligence built into the ECU.

This is assuming the AFC is smart enough to know that hot-wire airflow meters typically have quadratic transfer functions and so turns the specified 50% drop in airflow into a 75% drop in voltage.

And this is also assuming that you neglect any differences in injector opening and closing time.

$300 is relatively cheap, so an AFC is certainly something to consider. It's your choice. I personally don't think it's worth the money. There are other cheap options. MegaSquirt can easily cost less than $300, and then you'll be able to leave the stock ECU in charge of ignition and have complete control over fueling. It's a lot more work, but I think it's worth it.

[BAC5.2]

Well, if you must know...

I'm getting a TD05-16G off of a JDM STi from a buddy of mine (Brent, actually). I want to be able to safely run boost without killing myself or my engine.

I was thinking the SAFC would be a stepping stone, to at least help the car from runnning to lean, or to rich.

What is this info about the MegaSquirt? How much more work? What does it involve, and where can I get it?

[vrg3]

Oh. I knew you were getting the TD05 from Brent already.

So you were going to try to put big injectors on and use the SAFC to make it work, like what Brent did and like I was hypothetically describing? Cuz an SAFC can't do anything about tiny injectors on its own.

If you don't want to bother going all-out and want a temporary way to get enough fuel for a big turbo like the TD05, I'd say to just get a Walbro 255lph fuel pump and buy Kelley's RRFPR.

I never told you about MegaSquirt before? It's a very simple DIY ECU: http://www.bgsoflex.com/megasquirt.html

[BAC5.2]

I was going to use it to help tune with our stock injectors. I want to call RC Engineering (not Race Comp) and find out what a set of custom big-flow injectors would cost. I would like some 550 to 600cc injectors eventually. They should be able to pee enough fuel for everything to work well under high boost.

I don't see the RRFPR on Kelley's site. Where do most people get Walbro's for a good price?

Also, you think I could build the Megasquirt? Ha. I'm sure I could figure it out if I had the time, but I don't. Does the Megasquirt actually work? How effectively would it do so? Would it prevent me from having to go to standalone while taking full advantage of a set of bigger injectors and the ability of the 16G to flex it's muscles?

Yea, I was toying with the idea before, but then he presented me with the idea of just doing it now so he could put everything back together with a new turbo in place of the 16G (VF22). I'm going to get my stuff together, and I might have it as soon as next week.

[vrg3]

Neither an SAFC nor MegaSquirt, nor any other aftermarket ECU or piggyback, will allow you to make good power with the stock injectors. At stock fuel pressure, they just can't flow enough to support more than maybe 240ish horsepower safely.

Increasing fuel pressure is the only way to get the stock injectors to flow more. It'd still be kind of a band-aid, but it at least has a chance of working.

Kelley mentioned the RRFPR in another thread. I'm sure he'd be glad to give you more info on it if you asked.

MegaSquirt does indeed work. My brother and I used it when we turbocharged the 6M in his Supra. It is a standalone ECU; in fact, it was originally developed for converting muscle cars from carbs to EFI. It's very simple (it only deals with fuel, it's nonsequential, it only has an 8x8 map, etc), but it does work.

Even if you don't want to learn how to build it, you could find people who build it for you. My brother might even sell you his, since he's now using the stock ECU for the 1JZ he put in.

[azn2nr]

ok along the lines of the same topic are the ecus in the gc/gf cars the same physicaly as in dimensions and harnesses. and i mean for gcs that are jdm ej20g,k,7 because apex as a new plug and play standalone for the gc that comes with the controlor. its a pretty good price as far as standalones are concerned but after shipping from japan it myght be the same as a perfection and it might not work. but are the gc ecus physicaly similar enough for it to plug in.

see more info here
http://www.takakaira.com/asp/template.a ... QqlnMokrNm

[vrg3]

The pinouts are almost identical. The dimensions are the same but the mounting brackets are very different. You could probably make the that APEXi thing work.

[BAC5.2]

Brent's 98 has a PowerFC. He says it seems pretty sweet. But he also has the JDM ECU.

$905 is a pretty good price, and if it works, then that's even better. I'll look at Brents ECU in the 98 and take some pictures.