beepy

Okay, I have been conversing with a few of my more intelligent freinds, and this is an idea that we came up with and I may try tomorrow:

Background. It's me, the guy with the turbo Kia. Just ignore that for now, because this idea should work for any newer car with OBDII. Basically, the problem with OBDII is that it is too smart. If you use the hack, the short-term fuel correction built into the ECU will return the AFR to stoich no matter what you do. This is even more the case since the hack reduces MAP/MAF signal and therefore reduces calculated engine load under all conditions. The hack is also RPM-dependant while boost (and fuel requirements) are not. It also jacks up timing. But ignoring all of those faults, what exactly is it that the hack does? For you MAP guys, as well as myself and the MAF crowd, the hack reduces the signal voltage from the sensor therefore avoiding tripping a CEL code when in boost. (For me, it is a "MAF sensor voltage too high", basically the same thing)

My friend, Matt, pointed out to me that he could do the same thing with a simple .5-watt resistor. I doubted him, he proved it by running 30# injectors in his Rio N/A, with no hack, and purring like a kitten.

Boy did I feel dumb.

He also pointed out what I already knew: There just is no way to "trick" the powerful OBDII ECU. But then we uncovered its Achilles' heel. The fuel trim corrections are made based upon the primary O2 sensor signal. If you could increase the voltage slightly, you could make the ECU think it was running lean, and it would add more fuel. Morover, as it went rich, the ECU would not know. (Well, you would still get a "catalyst system efficiency below threshold" code, but let's ignore that, as it isn't a mortal condition)

So what could I do with that SAFC that I no longer need because a simple resistor does the trick?

I think you know what I'm getting at.

Most MAP/MAF sensors are a 0-5v signal, and O2 sensors are a 0-1 volt signal. But the SAFC is designed to work with most if not all MAP/MAF/Flap/Kharmann sensors, so I believe it could learn how to correct a 0-1v signal? That is what this experiment hinges on. Basically, I plan to setup my fuel system as such:

1) Calculate the proper resistor and "calibrate" my MAF for my injectors. This should have the same effect as a SAFC with a flat baseline correction.

2) Restore the MAF sensor signal wire to stock condition (with resistor), and instead splice the SAFC into the FO2S signal wire.

3) Reset the SAFC and start from scratch... Let it "learn" my sensor just as it learned my stock sensor. I will likely experiment a bit to see what "type" of sensor it likes the most.

4) Empirically calculate a correction % for making a safe 11.5:1 AFR. I should just have to play with the SAFC and keep an eye on the wideband.

5) To maintain fuel economy, I would then set a 0% correction at "low throttle", with the enrichening correction at high throttle, with the high throttle set at ~25% throttle, right about where boost kicks in when cruising.

I think that this would solve my part-throttle boost issues, and would allow me to run a safer AFR without switching to a standalone.

Thoughts?

Joseph Davis

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by beepy &raquo;</TD></TR><TR><TD CLASS="quote">My friend, Matt, pointed out to me that he could do the same thing with a simple .5-watt resistor. I doubted him, he proved it by running 30# injectors in his Rio N/A, with no hack, and purring like a kitten.</TD></TR></TABLE>

Welcome to the MAPhack. http://www.carolinahondas.com/...=1924

It's not the greatest idea, and if he manipulated the signal instead of the refference voltage, he did it wrong. Output gets really, really skewed that way. Voltage divider on the referrence side is even worse (should be regulator) because the amount of voltage shunted off varies with the variable load of the MAP/MAF under dynamic operation. I reregulated the referrence voltage down to 3v input to the MAP, and fiddled my fuel and ignition maps before there was freeware boost code... that ran fairly well for a couple of years.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by beepy &raquo;</TD></TR><TR><TD CLASS="quote"> Basically, the problem with OBDII is that it is too smart. If you use the hack, the short-term fuel correction built into the ECU will return the AFR to stoich no matter what you do. </TD></TR></TABLE>

I've been through this before multiple times - that is not a bad thing. You only think it is because you don't understand what is going on. The only time the ECU recognizes O2 input is at idle and cruise, and it manipulates fuel trim to acheive stoich. My OBD2 experience is limited to Hondas, but they ignore closed loop fuel trims when in open loop aka WOT. Just like every other ECU ever.

If you are worried about part throttle boosting + closed loop - don't. Your fix is as simple as a using an opamp + free MAP sensor to switch TPS input to 4.5 volts under boost. WOT to the ECU is, quite literally, WOT. Call me crazy.

beepy

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">If you are worried about part throttle boosting + closed loop - don't. Your fix is as simple as a using an opamp + free MAP sensor to switch TPS input to 4.5 volts under boost. WOT to the ECU is, quite literally, WOT. Call me crazy.
</TD></TR></TABLE>

Yeah, I could build that. What sort of switching method do you use to swap in the faked TPS signal?

I also read the article on the MAPhack. Good article. I'm guessing if I cut reference voltage ½, I would then be calibrated for injectors twice the size of stock?

Tactic$

bump... i love seeing what these two have to say

Archidictus

J. Davis is my hero.

Gotballcancer

I like stuff!

Joseph Davis

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by beepy &raquo;</TD></TR><TR><TD CLASS="quote">Yeah, I could build that. What sort of switching method do you use to swap in the faked TPS signal?</TD></TR></TABLE>

Opamp to cheapass FET is my recommendation. For other opinions, http://www.electro-tech-online.com

Try to figure out which poster I am there. Still only a small fish in that world...

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by beepy &raquo;</TD></TR><TR><TD CLASS="quote">I also read the article on the MAPhack. Good article. I'm guessing if I cut reference voltage ½, I would then be calibrated for injectors twice the size of stock?</TD></TR></TABLE>

Yes and no. Or, maybe just no.

MAF outputs are curved where MAP outputs (for the most part) are linear. You have a little more leeway when cutting signal on a MAP setup than MAF because of this.

Given how injector deadtime is a large factor at idle compared to cruise/WOT, the more you cut signal to the ECU the more exaggerated the discrepancy of idle AFRs to cruise/WOT AFRs. For example... say it takes 1/6th of the injector pulse at idle to actually open the injector so flow can start - aka deadtime or nulltime. With a reduced MAF/MAP signal, you end up with that deadtime eating a larger slice of the available pulsewidth at idle - to the point you may not idle, or idle well, at all.

There are a couple ways around this, such as another opamp that reads unmolested MAF/MAP output and determines the desired cut for idle/cruise conditions versus boost events. Or you could just buy a damn Megasquirt, which is what I think you should have done since day one. The electronics hacks will help that fine mind of yours grow... but I think an actual standalone will get you further, mentally as well as performance wise.

JDogg

the only obd2 stuff i have tried to tune w/piggybacks is an s2000 with a turbo and some 550 cc injectors with a e-manage... that ecu would apply short term fuel correction to part throttle/idle as well at wot. usually in less than 2 minutes

i could adjust the fuel while driving.. do a wot pull, it woudl be great.. the next pull it would run like ***.. rich as ****.. basicly undoing what we had done w/the e-manage.. it would correct for the idle in about 30 sec.. lean it out and it would idle great, 30 sec later.. it was back to idleing like it had 550cc injectors on a stock ecu... lumpy as hell ~10:1 afr.

obd2 sucks.. do it right and get rid of that ****

Joseph Davis

Seeing as how the S2000 in question should be correcting via O2 feedback - factory wideband for the primary sensor, yes? - then we can label your experiences as operator error, JDogg.

hehe

JDogg

i have no idea what kinda o2 the s2000 uses from the factory..

Joseph Davis

Regular 4 wire O2, non-wideband, still tunes the idle for stoich on everything OBD1+.

Then again, I refuse to "tune" QAFCs and Q-manages. That might be your problem.

Blitzkrieg

Typically OBDII gets it's LTFT and STFT during idle and partial throttle. You have about + or - 12% before it starts to compensate. What we do is figure out the percentage difference between the OEM injectors and the larger ones. Then we drop the fuel pressure on a 1:1 raising rate regulator to bring the larger fuel injectors with in that 12% range. From there with the boost you will pick up the pressure at WOTs and have plenty of fuel. Of course the minimum fuel pressure restricts the size of the largest injector you can use and still stay with in that 12%.

Joseph Davis

That won't work except for the NA boys who only need small changes in injector size... you drop the fuel pressure too much and your spray pattern + atomization goes all to hell.

beepy

J. Davis is a genius!

There, I said it.

I decided not to try to fiddle with the O2 signal, and left the MAF basically alone (still the SAFC), but I followed your advice and built a TPS hack.

It was simple. I bought a boost-switch (mechanical type) that has dual inputs and switches at about zero boost. Then wired it as such:

"common" connects to the TPS signal wire going to the ECU.
"normal on" connects to the TPS signal from the sensor.
"normal off" connects to a variable voltage regulator circuit.
And of course the vaccum hose to my common vaccum source.

The voltage regulator is then connected to a switched power source and ground, and I adjusted its output to (almost) exactly match the TPS signal voltage at WOT.

Now, as soon as the engine tips into boost, the ECU reads WOT, gives my a nice accelleration enrichment, and goes O2-sensor stupid. With 19# injectors I got 11:1 AFR at any RPM I could make boost at. Of course, I can adjust that via the SAFC, so that I have a "normal correction" and a "boost correction".


Thanks J. Davis! Running rich in boost is so much more relaxing than running stoich.

beepy

J. Davis is a genius!

There, I said it.

I decided not to try to fiddle with the O2 signal, and left the MAF basically alone (still the SAFC), but I followed your advice and built a TPS hack.

It was simple. I bought a boost-switch (mechanical type) that has dual inputs and switches at about zero boost. Then wired it as such:

"common" connects to the TPS signal wire going to the ECU.
"normal on" connects to the TPS signal from the sensor.
"normal off" connects to a variable voltage regulator circuit.
And of course the vaccum hose to my common vaccum source.

The voltage regulator is then connected to a switched power source and ground, and I adjusted its output to (almost) exactly match the TPS signal voltage at WOT.

Now, as soon as the engine tips into boost, the ECU reads WOT, gives my a nice accelleration enrichment, and goes O2-sensor stupid. With 19# injectors I got 11:1 AFR at any RPM I could make boost at. Of course, I can adjust that via the SAFC, so that I have a "normal correction" and a "boost correction".


Thanks J. Davis! Running rich in boost is so much more relaxing than running stoich.