I admittedly have very little practical experience modifying my cars. I do plenty of mental simulations and grasp the concepts pretty well though.

I've been patiently awaiting some inexpensive and reliable means of handling the fuel and spark issues Forced Induction presents.

A friend of mine put me onto a concept I haven't seen discussed here. Additional Injector Controllers. So I got to thinking, between an AIC and an Ignition Controller kit, wouldn't it still be manageable. Let the stock ECU handle everything until you start seeing boost. Spark timing control would probably need to be removed entirely from the Mazda unit's control (at least for the leading plugs) but I think I'd be okay with that.

That being said, what about using the MegaSquirt as ignition controller and Additional Injector Controller all in one. I really don't like the idea of running a 2 stage fuel system in place of the stock 3 stage. So instead, let's go the other way, and run a 5 stage fuel system.

Of course, if you source the right injectors, you could add alcohol injection running off the same system.

What am I missing here?

 

The glory of the megasquirt is that it can control almost anything. The downside of the megasquirt is that if you don't know how to build it to do what you want, you can't get it to control anything. You are aware that the base MS will be a pile of boards and resistors that you put together yourself with any additional functionality being custom by you? Any premade board you buy would probably not be wired to do what you want.

 

The main question with AICs is - where do you mount the additional injectors?
This question is rhetorical in this context because that has been answered.
However, on the Renesis, it is not a matter of simply mounting them somewhere on the upper intake manifold.
Its a moot point anyway, because all of the same problems and issues that exist with the current engine management solutions are the same with an AIC.

 

How are the issues the same? The AIC starts dumping fuel when manifold pressure goes positive. I understand the megasquirt is a kit, at this point I'm still hashing out ideas.

 

Because the precise amount of fuel is a moving target if you are not in control of the other 6 injectors.
Simply dumping in a proportionate amount to fuel to the increased airflow will not yield anything close to a uniform A/F ratio on the RX-8.

I've used AICs on other cars in the past and they can be used safely to about 50% over base power on a boosted application that doesn't have the kind of OEM fuel control that the RX-8 does.

You might want to go and read the other engine management threads and understand the issues before you go any further in arguing the point.

 

Asking a question does not equal arguing. I've read many of the threads, that's why I asked the question, I hadn't seen this approach addressed. All the approaches I've seen involve either wresting control away from the stock ECU or fooling it. Hopefully before I'm actually ready to do my setup we can simply give the stock ECU new instructions, but given the pace of development, I have my doubts on that front. Given a set of parameters, the stock ECU is going to behave the same way time and time again. It's a computer, it's impossible for it to do anything else. Are some of the parameters it's working with historic? It seems so. That, however, doesn't change the equation one iota. That being said, I would think it possible to learn the stock ECU's behavior during tuning and, with the right setup, build a tune that accounts for the expected behavior. When real-world conditions cause an out-of-bounds situation (AFR outside expected norm) to occur the driver should detect this, by monitoring his gauges, and take appropriate action. The issues, by your own description, are not quite the same as other approaches have encountered because we're not doing the same thing We're not trying to fool the ECU. So again I proffer the question: What am I missing? Again, I'm not trying to argue, I'm trying to understand.

 

However, that set of parameters exceeds anything an AIC can keep track of.

Human reaction time is FAR slower than the duration of a single pre-ignition event and those can happen without the A/F reacting in a predictable way.
The only method that works with the Renesis is proactive tuning.
AICs and watching your gauges is reactive tuning.

I've written thousands of words on the subject here and elsewhere, so I am not going to repeat myself.
The OEM PCM is concerned with significantly more parameters than just those that constitute load and an AIC is not equipped to address these inputs.

 

that is true, however, the list of parameters is long, and the ecu does have some adaptability, which actually works against running better with fi.

and b) if you did use an AIC you'd find out why people dont really use them anymore. even on a car where the ecu is "predictable" the aic doesnt compensate for changes, be they boost, weather, etc etc

so can you use an aic? yes. is it adding more complications to an already complicated setup? yes. is that good? no.

 

Most importantly - modern intake manifolds are designed "dry" - they don't flow vapor the same way they flow gas.
The Renesis manifold is further complicated by a crazy arrangement of valves and multiple pathways.

 

Thanks for the explanations. After some more digging, I'll probably go the megasquirt route and handle the 3 injector stage in code to maintain some semblance of the the original design. The turbo design I'm working out is intended to not even spool noticeably below 3500 revs, and be making design boost by 5500, so keeping the three injector stages should work out better than 2 injector stage and a 4 injector stage. That'll allow me to stage the injector flow rates for better longevity and responsiveness.

 

The 3 injector setup is designed for precision, not peak power.
If you are looking for peak power (which it seems obvious to me, if you would even consider a turbo that magically doesn't produce push below 3500 but peaks at 5500, is your "goal"), I'd just force all of the intake valving open right from the start (you aren't concerned with the CEL since you will go MegaSquirt) and gang the secondary and primary 2 injectors together.
Before you do any of that, I'd learn how to read a compressor flow map and maybe get some injector/airflow math under your belt.

 

Or you could just go with a proven piggyback solution and save yourself a lot of heartace .
there are a few out there .

 

I don't think he is interested in going down any beaten path.

 

yeah just run the star mazda motec setup....

 

OR he can wait for the mythical Access Port...

 

Actually, peak power is only part of my concern. I'm looking at stuff that needs more exhaust flow to spool than is produced below 3500 in order to ensure a minimum of impact on light throttle cruising. Basically, I love how the car is basically three cars in one. Sedate and manageable with zero effort when driven only in the primary intake ports and injectors. A proper spirited sports car in the secondaries, and ready to gobble up road as quick as you can feed it in the tertiaries.

I do have one very serious question though. Why is that the instant someone starts asking questions that would allow him to think for himself, the more knowledgeable instantly assume they're going to go off half-cocked. The fact I'm asking questions is a pretty good indicator that I'm interested in getting MY application right. Are my goals the same inside-the-box goals most people have, obviously not. Are they realistic, I haven't a clue. A friend indicated the driveability profile I'm looking for is possible. Do I know if he's right, of course not. I understand no one here knows me personally, but I'm a wee bit smarter than the average dolt. I have no intention of just bolting a bunch of **** together and hitting the road.

This is just one part of a fairly extensive project to take an already beautifully performing car and build it into the back-side kicking GT car* Mazda COULD have built.

*GT Car = a car that allows you to pass the volvo on the on-ramp without setting down your latte`.

 

Generally, the problem is in the way the question is presented.
If I am unsure about something and I think it is possible to get some input, I will site quotes from my searches that support my question.

I gave you two cogent, concise answers (#3 & #5) and you still refused to address the logic in them. You just returned to me the typical toddler mantra: "Why?"
Have you read the engine management threads? Do you see why simply dumping fuel on load is not going to work?

 

I'm pretty sure you will make the motor less powerfull at low revs with a large turbo .......

 

Dumpin just fuel isnt the issue...AP if it ever comes out should make life easier...any day now...any day...

PS: You said megasquirt...heh heh heh...

 

Yes, I do now understand why dumping fuel at load wont work. And yes, I have read several of the Engine Management threads. I got a bit turned around when the idea of an AIC was presented to me. Honestly, I'd never heard of such a thing, and since I hadn't seen it discussed, wasn't sure how it fit in the equation. The logic on the AIC sites seemed sound enough, it just takes time for someone who's focus is driving, not building, to combine all the information from all the sources. I did address your point about a moving target. The idea was to use something that was capable as a standalone (and thus could see the results of the combined efforts) as an AIC. The complexity of that route being pointed out (I'm sure I'd have realized it eventually) makes that an unattractive option.

I'm an Enterprise Architect, so while the skill set is different, the disciplines are similar. It's a matter of FULLY understanding the current state, the goal, and all the options and issues in between. Sorry, it's been my experience that, while childish, simply asking "why" tends to be the most effective means of getting at the REAL issue. I guess it's just habit from dealing with excessive executive BS.

So the primary draw back is that because an AIC is purely load based, it can't see the effects of the ECU's tuning adjustments, making it very imprecise. This difficulty could theoretically be addressed by using a more powerful system that monitors more variables as an AIC, but that would be anything but an elegant solution. Does that about sum it up?

I really do appreciate the guidance. Sometimes it takes a little more than a "because I say so" comment to trigger that "oh yeah" moment. I'll now go back to my calculator and graph paper for a while.

 

If your a computer guy.... here this might show you why an AIC won't work. Check out the engine logs as well.

https://www.rx8club.com/showthread.php?t=118529

 

Thank you, that confirms it, at the high level. BTW, very very very nice.

 

Hey thanks. If you look at the input deviations, over time they are close, but you can def see where the PCM is taking a ton of variables to decide how much fuel to inject. Otherwise at WOT at X RPM in open loop there should not be any deviation. But the PCM is a complicated *****.

 

you can say that again.......