Wide Open Throttle Solution

M

mp3rican007

Member

• Sep 26, 2002

• #1

Well folks I think I may have the solution to the wide open throttle problem on the protege ecu. Well this isnt my idea but my shops idea. Ecus run 0-5 volts or 5-0 volts assumming. Well with a tester we have to figure out what volt wide open throttle kicks in at. It might be 3.5 3.7 4.4 nobody knows, but when we find out what you do is get a resistor to clamp voltage. Say wide open throttle kicks in at 3.8 volts then you clamp it at 3.7 volts, so the ecu will never see wot. Some hondas and mitsus use this trick. Tell me what you think I am installing my emanage sometime next week once i figure out what the injector ground wire is.

Oh yeah you should be able to buy a resistor that we would need at radio shack or some shop like that
Omar

 

M

MazdaT

Member

• Sep 27, 2002

• #2

You cannot make a voltage clamp circuit with a resistor!

 

K

Keith@FP

• Sep 27, 2002

• #3

Very true.

The Protege ECU is a very complex piece. It even has short-term learning behaviour. We're doing a lot of experimentation at the moment.

What exactly is the "WOT problem" that you're trying to solve? The TPS on the Protege is not a binary switch, but is sensitive throughout the throttle travel. The computer may go open loop at 4000 rpm at full throttle or 4700 rpm at 60% throttle. WOT is not the only factor.

Keith

 

2

20ESGUY

Member

• Sep 27, 2002

• #4

I think what mp3rican was getting at is there has been disscusions on the forum as to the ECU being "hard" programed at WOT and ignoring all but a few inputs and using one preprogramed fuel and igntion map, instead of being able to use an adaptive map. I guess thats the best way to explain it. Thats also a reason nobody's put the Apexi S-AFC into use on the protege. If it is programed as mentioned above it wouldn't have a good use for anything except below WOT.

 

K

Keith@FP

• Sep 27, 2002

• #5

Here's the science behind the rumour.

ECUs have two basic modes: open loop and closed loop.

Closed loop means that the ECU looks at inputs such as the front O2 sensor and adjusts the fuel for the ideal stoichiometric ratio for maximum efficiency. The ECU makes a change, looks at the O2 readout, makes a change - looping around.

Open loop means that the ECU will ignore most of the sensors and run off the pre-programmed fixed fuel and timing maps. These are very rich, partially for safety.

Now, what does this mean to those who modify their cars?

First, what prompts a car to go open loop? On older cars, it could be as simple as going over 4000 rpm. However, the Mazda engineers want the Protege to run closed loop as often as possible for maximum power and efficiency. So the stock ECU will only go open loop under certain conditions.

Here's where the Protege gets tricky. It'll go open at WOT over 4000 rpm, at partial throttle openings and higher rpms. Fair enough. But it will also go open loop and stay open loop on a sustained run through the gears. If it has to make a lot of fuel changes, it will go open loop on a small provocation until things settle down for 30 seconds or so, showing a short-term learning ability.

Now, so what if the car runs closed loop? Well, the Protege ECU has a huge range of authority to alter the fuel delivery. A boosted car wants to run 12:1 A/F under boost to stay happy. But the ECU wants to see 14.7:1 and can pull at least 30% of its fuel away to do so. So your extra injector controller, your higher fuel pressure or whatever is not going to do you any good at all. You can dump in a huge amount of fuel and the stock ECU will pull its own fuel delivery back until it gets what it wants.

It's not simple. We're working on a solution for our turbo kits, and it includes a lot of time driving around with an OBD-II scanner, a wideband O2 sensor and various electronic devices all hooked up to see what's going on and how we can alter it. There's a photo of our poor car hooked up to all this gear on our website: FM Projects in the shop

Keith

 

Last edited: Sep 27, 2002

N

njaremka

Wiggity-Wack-O-Tack

Contributor

:

2012 Mini R55

• Sep 27, 2002

• #6

keith,

anyone ever tell you that you guys rock! get things straightened out so i can get boosted!

 

2

20ESGUY

Member

• Sep 27, 2002

• #7

I gotta agree with njaremka, you guys definitely do your R&D. Thanks for the good explination of the ECU function too. Seems like mazda has thrown in lots of little road blocks for manipulating the ECU's maps. Now only if I knew more about how to actually program the ECU's rather than just flashing them it'd be all good.

 

L

LinuxRacr

Contributor

:

2001 Mazda MP3

• Sep 27, 2002

• #8

Damn...Some hard facts that get the issue straight!

 

K

Kooldino

MOTM Winner Jan '09

Authorized Vendor

Contributor

:

MP3

• Sep 27, 2002

• #9

Awesome, thanks for the explantions and everything.

You guys are doing the right thing by being an active member on the forums, and having us here step by step. It will pay off in the end.

 

K

katsmp3

Member

• Sep 27, 2002

• #10

we should direct enry's attention to this thread. he has a degree in electical engineering.

<joke> heard he minored in ballet though </joke>

kat

 

L

LinuxRacr

Contributor

:

2001 Mazda MP3

• Sep 28, 2002

• #11

katsmp3 said:

we should direct enry's attention to this thread. he has a degree in electical engineering.

kat

Click to expand...

So do I, but if you don't use it, you loose it.

 

B

big_ben

Member

Contributor

• Sep 28, 2002

• #12

I've learned a lot from this thread. Hey, if you guys at Flyin Miata ever need a river modeled, or any type of structural of civil engineering work done I'm your man, But when it dcomes to any type of electronics, I think I'll leave it to you.

 

Antoine

Administrator

• Sep 28, 2002

• #13

Sheesh...hope you guys aren't taking any dates in that car

Great thread...great info...intense R&D...that is why I'm honored to have FM on board. (thumb)

 

S

Stou

Member

• Sep 28, 2002

• #14

Keith, thats some very good information. I can't wait to see your solution for turbo Proteg! (thumb)

 

7

70

• Sep 28, 2002

• #15

LinuxRacr said:

So do I, but if you don't use it, you loose it.

Click to expand...

I've heard that ballet is like riding a bike...You never forget, you just need a little warm up.

But, on topic, FM is cool. Do you guys watch DVD's too with all that gear in there??

 

T

Traveler

Member

• Sep 28, 2002

• #16

Hey Keith, I've had my car on the dyno three times now, and the dyno guys and I came up with an idea. If you pull some of the pressure out of the stock fuel system (adjustable regulator) say 2-3 psi, then when the car goes to open loop it should theoretically run leaner. We tracked the fuel under WOT, and it started going rich (stock) at around 3500 rpm. With my header on it, it doesn't go all the way rich until 4500, but it's still rich. That's under 100% throttle by the way.

I know that the ECU can adjust the fuel pressure within a limited range, but if you can take out 2-3 psi throughout the rpm range then it should lean up WOT. The ECU will just compensate for the slightly lower pressure during closed loop by increasing the duty cycle of the injector until it gets the A/F ratio it wants.

Does this sound like a workable solution? Or is the ECU able to boost the fuel pressure right back to what it wants to see?

 

L

LinuxRacr

Contributor

:

2001 Mazda MP3

• Sep 29, 2002

• #17

StuttersC said:

I've heard that ballet is like riding a bike...You never forget, you just need a little warm up.

But, on topic, FM is cool. Do you guys watch DVD's too with all that gear in there??

Click to expand...

You're right about that! It would all come back if I were to start using it again, with the quickness!

 

S

spoolinmp3

Member

Contributor

• Sep 29, 2002

• #18

Damn this is a good thread
Keith,
in order for the ecu to maintain a leaner a/f ratio does it always instantanously pull out 30% of fuel or does it vary in percentages?
Lets say for example the ecu pulls out 10% in a ceratain condition and 25% for another, this is a bad thing since we cannot consistently have a predictable baseline a/f ratio.

I have noticed during dyno runs while using the wideband I get a consistent flat line on my air fuel ratios right under 12:1, where I want them to be, but on a few occasions I have noticed it will intially start off to at 11.5:1 and then spike up a bit to 13:1 showing some inconsistensies. Then I will slightly fatten up the mixture at that rpm and the subsequent run the a/f will drop to like 10:1.

But then on previous dyno runs we were not getting any problems. Regardless I am interested to see how extensivley the ecu has authority over a/f
Thanks
Terry

 

Last edited: Sep 29, 2002

K

Keith@FP

• Sep 29, 2002

• #19

Traveler: You're trying to lean out the open loop operation? What you suggest might work. I'm not sure what it would do to power output, but it would probably help. You're taking some of the safety margin out of the car, but that's what a lot of hot-rodding is...

Spoolin', it'll pull out as much as it needs. It seems the limit is +25% or -30.8%. At least, that's the most I've seen. Increments seem to be around 0.8% in both directions. It will pull out (or add) just what it needs to hit 14.7:1. To compensate for our AFPR, it was running about -10% or so. When we disconnected the signal from O21 (the front O2 sensor), it defaulted to 0.450v and started pulling fuel like a maniac.

Our car was open loop almost all the time on the dyno. In the real world, it had very, VERY different behaviour. The amount of fuel being pulled was varying all over the place depending on what the O2 sensor was reporting. I think it was also getting information from the MAF meter and taking that into consideration, although I need to do some more measurements and calculations to see what the airflow of a non-turbocharged engine would be in certain conditions.

So the end result is - there's no consistent baseline. If you tune your car to run well on the dyno (ie, open loop) it will run at stoich when you're on the street. If you manage to dump in enough fuel to overcome the stock ECU's authority when it's closed loop, you'll be running about 4:1 as soon as the car switches to open loop. We're going to have to find some way to deal with the computer, as simply adding more fuel isn't going to help.

Terry, I'd be careful with your car on the street.

Keith

 

S

spoolinmp3

Member

Contributor

• Sep 30, 2002

• #20

Interesting Keith,
I drive around sometimes with the wideband on the street, I agree the dyno and street are completely different. I prefer to tune the car to real world conditions then make minor adjustments on the dyno. I have been running acceptable a/f levels and my egt's are generaly on the low end. I get about 770c at WOT. My probe is mounted on the leanest cylinder so that translates to very acceptable egt's. I also monitor my gauges like a mad man

Buton few occasions things "freak" out and the a/f ratio will either go really low or peak a bit (12.5:1).

My question is this(This may sound stupid) How about if you account for maximum ecu fuel compensation -30.8% and set fuel according to that for a fail safe? This obiviously will not be set for maximum performance (when the ecu is not taking out as much fuel or adding fuel) but for reliabilty sake??

What is your opinion on this?
And do you plan on programming or tricking the ecu to resolve the problem? Or a full link replacement ecu like the M1 miatas?(thought)

 

Last edited: Sep 30, 2002