guitarjunkie28

it's not the big dips, but hte small variances in hp--the thing that makes it look all squiggly.

there is a smoothing factor that averages those out to make your curve smooth.

ModMech

Since they are significantly "in bed with" Ford Motor, that should not be any shock.

One technical point I can offer is that the EEC uses the MAF as the "god sensor", EVERYTING depends on the MAF. Other sensors allow the EEC to fine-tune, but at WOT the MAF is THE sensor.

At WOT the EEC uses only the MAF to calculate load. It completely ignores the HEGOs. So, if the MAF is off or the signal is not "true", you will not get to 100% load in the EEC, and eventhough you have squashed the go-pedal fully, the EEC will not act as though you have. It "knows" the VE of the engine at any given TP% and RPM, the EEC compares that to actual MAF to calculate load. As you can see, if you do not get to 100% load, you will not get to WOT enrichment, you will knock, it will pull spark, and every 1* of advance is good for 3-4 RWHP, you can get hurt pretty fast above 6500 RPM by losing spark advance.


Tell that writer you have a present for him, and that you'd like him to bendo over as you deposit that 8 right were it belongs.

Richard Paul

iTrader: (5)

I already told the service writer what to do. Second time I met him.
That's why I'm at MazdaUSA. If I'm my normal nice self there and burn that bridge where will I go? FoMoCo?

Reminds me of the time I got stopped in my Testarossa up on the 5. I wasn't in the best of moods anyway, so when the cop asked "why are you driving a Ferrari and me a truck" Well, I told him why, "because I have a real job and you don't". Guess you know how that worked out. I got to sit in the Magic Mountain jail for 4 hours.

On the other note, if what you are saying is true then drawing throgh the MAF will cause all sorts of problems if there is not enough headroom in the original map. Lets say that originally they figured the 8 will never flow over 600 CFM so they only program to 650. If I exceed that it instantly thinks the sensor is crazy and goes to limp home. Nice.

ModMech

RP,

It would be GREAT if the EEC went into "limp home", but it's worse! It just guesses! When you peg the MAF, the EEC simply uses 5.0V, or 100% load. You might very well be at 125% load, but as far as the EEC knows, you are only at 100% because that's all it "knows".

You mention drawing through, and I do sincerly hope that is your plan, because I know of ZERO GOOD blow-thru automotive applications. When you pressurize the inlet air is REALLY screws up the MAF, one part is that it is under pressure, the other is the turbulance in the inlet stream that was not engineered into the transfer function at the factory. So, a suck-thru is REALLY the only GOOD way to go.

Also, the MAF may, or may not, have a linear (or even close to linear) transfer function. It is common for MAFs to flow 70% thru 4.0-4.2V, and the remaining 30% or so using the last 1.0-0.8Vdc of range. With all your test equipment, it would be EASY to for you to map flow vs Voltage to get a very good idea of the transfer function. After all, that's all a TF is, Flow vs Volts.

Richard Paul

iTrader: (5)

That's all good except it will run out of volts before it reaches the range where I need to run. If it were a perfect world with a draw throgh system it would continue to function with the extra air. How can that be changed. Never mind I already know, ask Mazda.
But if as you say it reaches the max fuel supply and just stays there when extra air goes in I can live with that. If it fuels the flow up to wher the manifold goes positive then I just add fuel for the extra mass which I can calculate.
That is what I have in mind. I want to leave the stock set up alone and add fuel via aux nozzles before the blower based on manifold pressure. Said aux nozzles only feeding for the extra air. All air below that level is handled by he stock system.

ModMech

Actually, you can change MAFs to get the flow range you need. The only caveat is that you also need access to the EEC to program in the new transfer function. This is NOT difficult, but you do need to know how to access the EEC tables.

When you program it this way, the stock system can handle the extra flow as long as the injectors will accomodate it. The EEC, as I said before, know the THEORETICAL VE, and with good MAF data (even at 125% of "theoretical flow") the EEC will properly calculate load. That means that EVERYTHING the EEC controls will be exactly correct for loads over 100%. Best of every world.

The other way to deal with the additional flow without getting into the EEC is with an FMU. Given the electronic limitations of this project, this may be the most elegant method available.

ModMech

Does anyone KNOW what ECU model is in the 8?

If it's an "M", there is support for it, and we can open pandora's box on the programming.

Richard Paul

iTrader: (5)

I can't recall ever hearing anyone on the forum mentioning that they knew this. How can we find it? Could we take one apart and be able to tell?
I can try and ask Mazda when I'm out there next time. That will be the week of the 11th. I really doubt they will share much with us. They just don't want the liability of someone screwing with it.

StealthTL

iTrader: (1)

From old enquiries, this is all we have on the ECU chip....

It is an Hitachi,
the Hitachi SuperH processor by "Renesas" (the combined chip divisions of Hitachi and Mitsubishi)

High-performance single-chip RISC with SH-2E core
52 MIPS/40 MHz/3.3 V
High-speed multiplication/accumulation operations
Built-in 32-bit multiplier
Built-in single-precision floating-point operation unit
Built-in large capacity flash memory with a single power supply and large capacity RAM
Write and erase operations available with the single power supply 512 kB Flash ROM/32 kB RAM
Timer: ATU-II (a maximum of 65 input and output process) Compare-match timer 2 ch
A/D: 10 bit x 32 ch
Serial: 5 ch DMAC: 4 ch
HCAN: 2 ch (1 ch is shared with a serial interface)
Package QFP-256

Other than that, we know nothing!

S

guitarjunkie28

hey, i used a super h compressor on a hitachi turbo on my old fc :D

policyvote

Small point of interest: the Sega Saturn used dual Hitachi SuperH processors for its CPU.

Peace
policy

rotarygod

Richard, your idea to run an ecu as purely an additional injector controller should work just fine. Megasquirt could be a viable alternative for this. Get the supercharger mounted in the car and try one. They are $140 for the kit or $325 already assembled. The good news is that since it uses an onboard map sensor, you can dial in different boost level maps so changin boost would be just a change of pulley for the consumer which would require no additional tuning on their part.

Richard Paul

iTrader: (5)

We agree that the extra fuel can be controled OK. The thing we don't know is what the stock system is going to do when all that extra air is sensed by the MAF. For a small amount it will probably fuel it because there is some headroom in the program. When it exceeds that map what will it do? Shut down thinking there is a sensor failure?
Something needs to be done to tell it to just shut up and act like it was still within it limits. Something might be done by using manifold pressure to stall the stock system at that pount.

Stealth, all that is great but does it mean we can get to modify it with an available tool?

rotarygod

Supposedly the ecu doesn't care what the maf readings are so long as the O2 readings are fine. The ecu always readjusts the maps because it senses some weird O2 readings. If you add fuel to keep them inline with where they should be, the car shouldn't care.

Richard Paul

iTrader: (5)

Fred we are talking back and forth on two threads. Anyway this is the road we shall go down. MazdaUSA gets the car next week and hopefully it will come back OK. I'm again starting to test some new blades. I hope they are machined by the time the car gets back. That doesn't mean one will be together just maybe close to getting on the test bench. I can use a mock up for the fab work.

rotarygod

Did you stumble across a more efficient blade design?

ModMech

Unless Mazda has done something VERY unique with the 8's ECU/EEC, they IGNORE the HEGO(s) at WOT (over about 70%-80% load/throttle), and rely SOLELY on the MAF sensor signal and command AFR look-up tables.

How much "headroom" does the stock MAF have, or do we not know?

Richard Paul

iTrader: (5)

You don't "stumble" across blade designs. :p
Without the experiance and high powered computer programs of the big engine manufactures we have to just thrash out a design by hand and then test it. Right now we don't have the fixtures to do single stage testing so we build a whole comp section. If I were starting over and had the time it would be nice to study each blade by itself then in a cascade. Maybe down the road we can set up and do tests that way.

There has to be a limit on the airflow through the MAF as there is a limit to the nozzle capacity. If that were the limit It would be great because there is enough nozzle in there for over 300hp.

Jon

If I missed a post talking about an o2 simulator then I'm sorry for this post.

Whats to keep the ECU from trying to hit the target air/fuel even if there is a seperate independent injector. The stock computer could compensate for the increased rich mixture with the stock injectors.

One more question if you dont mind.

Are you intending on leaving timing stock and just running a conservative boost level or are you going to make a timing control unit.

Jon
PTP Motorsports

Richard Paul

iTrader: (5)

For sure I'd like to do a min of spark retard based on man press. It will be very basic and any controler could do it. It is just that we need to not get intwined with the stock priograms. I don't think the timing is going to be that hard. If it comes down to it there are boost retards for regular ignition systems that if multiplied could be used here without the stock ECU knowing it.

Back to the nozzle question, does the stock unit mod the mix even in open loop by way of the O2 sensors. Or does it just follow a map based on speed and throttle?

rotarygod

The problem with using a simple knock retard unit is that it needs to retard both leading and trailing independently on different channels. This can't be done with something like an MSD 6BTM. A J&S knock sensor would do it but this is only after it hears knock. I don't like that idea.

slaxer07

I'm trying to find some documentation to back it up but can't as of the moment but, as ModMech said, it looks at the MAF and then will reference a A/F table according to that voltage. It ignores the O2's until it goes back into closed-loop. Look at it this way if the PCM was to look at O2's during open-loop everyone that has installed a canzoomer would just be adding fuel that the fuel trims would compensate for and turnaround and subtract because it would see the rich mixture that the O2 would be reporting. That is why they can't start adding fuel until higher RPM's.

As far as headroom for the MAF if I get a chance tom. at work I'll hook the WDS to see how high the voltage gets.

slaxer07

I was just reading back on ModMechs comment, and one thing to add about load is the PCM will also look at TP. It will also look at the IAT for almost redundant info.

Jon

I was under the assumption that the ecu always took the air/fuel mixture reading and adjusted to compensate when the engine was warm except at a certain load and rpm. I would not worry about being able to compensate for full throttle high rpms I think the major issue is in the lower rpms and not quite full throttle. Or maybe just rpm dependent. I have noticed that many people are having problems in the midrange rpms and maybe thats why greddy's coolant temp switch only operates between 1200rpm-4900rpms. And maybe thats the problem the SSR turbo kit has at the low rpms with pinging. But I am far from being an RX8 expert especially on the factory system.

Jon
PTP Motorsports
512-834-8771
Jon@ptpmotorsports.com

slaxer07

You are right but that certain load and rpm you talk about is when it goes into open loop that is why you see it go pi rich there otherwise it constantly monitors the O2 sensors for the most effiencient A/F ratio. As far as the low rpm pinging with the SSR kit its obviously to lean which would make since because they said thats with the ball bearing kit and it would reach maxium boost at a lower RPM.

I wish I could look at the pid data on a turbo car to see how the MAF, and O2's react. They are the key toall this mess.

One more thing, didn't I read on here somewhere that some greddy or canzoomer owners had to reset kam every so often. (which would reset your fuel trim tables) what was the deal with that.