ECU Modding, Fuel Injector sizes and more crap all in one.

[scuzzy]

all in one thread. I've been meaning to do a thread about what I've been up to lately with pictures and what not but I haven't bothered. Most often when I do a thread it gets little attention so I've cared less recently.

So I guess we'll start off with my fuel woes, since this is a common theme. It's important you know about my car for any of this to make sense. I have:

1991 Subaru Legacy L Wagon
Turbo Swap completed minus the EJ22T ECU.
Stock Injectors till this past Tuesday (270cc?)
N/A stock intake manifold and fuel rails.

Not long after the swap I determined that boost was creeping enough to anything after about 5psi and the engine started to det/hesitate bad. real bad. so I did some digging for some quick fixes and found one about the fuel pressure regulator cap. talked to vrg3 about fuel pressure/etc (there's a thread around here about that)

So I took the spare FPR I had and did this:



I later found out via pressure testing (I used the air pressure regulator on my air compressor) that the modified one (crushed in with a c-clamp ever so gently) registered about 7psi higher over the non-modified one at static pressure; which means under 5psi boost it could have been anywhere from 7psi higher to well beyond when compared to the non-modded one on the same amount of pressure; but I didn't have a fuel pressure gauge to check it. I was told that pushing the cap in changes the rising rate ratio as well as the static fuel pressure; no clue if that's true or not.

anyway, long story short that didn't last. This past weekend it turned especially cold out, and I was detting again, just as much as before. I decided I was going to have to give the EJ22T injectors I had laying around another go.

I'd previously tried them, and reported that they just fuel far too much for the engine to even idle; me and doug got in an argument about they don't fit/etc and I was just blowing fuel into the cylinders past the injectors - it was very possible that I had failed to seat them properly. never the less I tried again.

There's a few points to take from this: don't work on a hot engine, don't pull injectors out of the rails when the engine is hot either. fuel in the rails doesn't boil BECAUSE of the pressure it's under, but remove the pressure and it boils away quite rapidly.

Unfortunately I pulled #3 (I think) injector, by the time I figured out that the rails was boiling fuel, pulling it from the tank, and flooding a cylinder the cylinder was already flooded. I disconnected the feed fuel line to stop the suction process and cranked the engine with the coilpack disconnected, spark plug pulled out, and promptly got fuel over everything within a 10 foot radius.

my oil seems to have survived, with not even a hint of a fuel smell to it.

The EJ22T injectors are almost identical to the N/A 1991 injectors.
Proof:



EJ22T on top, EJ22 N/A on bottom; you can clearly see the different size flow nozzles:


Installed:


So I got them installed, didn't bother pulling the battery cable and just cranked the engine. it took about five turnovers for the engine to catch and run on it's own - holy shit it's rich.

You have no idea. I was fogged out of my garage. I had to open the garage door and vent the garage for a half hour.
I took the car out. remember I'm using the N/A ECU, so this thing is trying to idle on injectors 30% larger and it's damn near stalling every now and then.

I drove it to work the following day, I had a quarter tank before I parked the night before and did the injector swap; normally a quarter tank will get me to work and back (~50 miles) and have a good 50 miles left in the tank. No.
I pulled in the fuel stop with the gauge well below the E mark, no low fuel indicator light (it was working the week before) and put 13.5 gallons of 93 octane in the tank. More than I'd ever had to before. it cost me $40


That evening I pulled my modded FPR off and put the stocker back on - it dropped the flow rates in the injectors, but the engine was still too rich at idle. The ecu was trying to correct the rich condition but i take it that the pulse width control is quite grainy: it was cutting them off and nearly stalling the engine (slow enough for the oil pressure light to flash)


I let the ECU sit last night disconnected from the battery, drove it today (as well as yesterday) and it seems to be a little better; it's still quite rich at idle (unburned fuel smells really really bad out the exhaust - lends oneself to make a choking noise)

By tonight I'm down to half way between 1/2 tank and 1/4 tank, I've gone 135 miles; it's about 30 miles short of average but it's not HORRIBLE. I suspect I've burned about 8 gallons at this rate and that works out to about 16mpg.

The Good: It pulls strong. If you've read some of my other replies lately I've had trouble controlling boost, it creeps something awful; well on up to 10psi pretty easily. the injectors flow enough to keep it nice and rich even up top and I don't suffer any hesitation at all.
10psi is faster than a stock Legacy SS, so I imagine that stands for something.


Okay, if you're still with me after that long assed rabble-rabble; I've considered modding my ECU. I want to buy an Autronic, but this fuel problem needs to get solved in a hurry - and I'd rather not spend $$$ on a piggyback I'm going to toss.

So I removed my ECU tonight and took some photos:











In a fit of amateur-ness I managed to lose the pictures of the EPROM at a distance (showing the location of the EPROM) but I can describe it (the ecu is already back in the car)

It's located on the daughter board which is connected by the wire ribbon pairs you see in one of the above pictures. the blue male-type wide socket is to the left of the EPROM, not directly over it.

The photo with the silver label on it is the EPROM, unfortunately I didn't take a further out picture to count the pins now; I'm just a noob - but I'm assuming it's much like the EJ22T EPROM just programmed differently/different data points.

what I need to do is find a relatively cheap EPROM reader/burner and the software to drive it. I'll go get a junk ECU from a parts yard and pull the EPROM off of it (so I don't render my ecu useless) and dump the code from it and work with that (unless someone here can provide me with the EJ22 N/A ECU's code for my model year car.

As I understand it the fuel and timing maps are all located in the first 8KB of the EPROM, so if I can get access to the maps for my ecu I can modify the fueling map at idle and burn a new EPROM, remove the old, install a socket and drop the new EPROM in and see if i can't fry some electronics in the process.

any thoughts?

[Fkyx]

My car's running terribly rich as well. I've love to fix it. Changing the fuel maps sounds like a good idea, but I know nothing of it.

Good luck!

[kleinkid]

I am totally baffled by electronics so can't give you any useful information, but it was informative and interesting to read your thread. Thanks for taking the time to post. You can't use an EJ22T ECU, I'm guessing, because, why?

[AWD_addict]

Good luck to you in being able to rewrite over stock chips.

How much do EPROM reader/burners cost?

[vrg3]

Wow, that's a lot of words.

Crushing the FPR's cap in does not change the rise ratio at all.

Increasing static pressure from 36.3 psi to 43.3 psi would increase fueling by about 9%.

Definitely relieve fuel pressure before undoing any fuel fittings on your car. Boiling fuel is especially dangerous but even at room temperature pressurized fuel is bad. It also makes a mess.

This may be a dumb question, but what made you think increased fuel pressure or bigger injectors would work?

It seems to me that there are two obvious and fairly easy solutions to the fueling problem: switch to a turbo ECU, or install a rising-rate fuel pressure regulator.

The EPROM in the Hitachi ECUs is a 27C256, pretty standard. I actually poked around an ECU like yours a while back, and didn't find any simple way to alter fueling directly. It doesn't appear to use a fuel map from what I can tell. There is an ignition timing map, however.

The weird blue inverse socket thing around the EPROM is actually wired right to the EPROM. There's a pull-down resistor on the ~CE line, so you can make a little daughterboard that plugs into that socket and ties the original ~CE to Vcc and replaces the original ROM.

A fairly cheap EPROM programmer is the Willem programmer. You can find them on eBay for like $50-$75 or something. But you can also build your own that works off the parallel port with a handful of ICs for almost nothing.

[scuzzy]

Wow, that's a lot of words.

Crushing the FPR's cap in does not change the rise ratio at all.

That's good to know

Increasing static pressure from 36.3 psi to 43.3 psi would increase fueling by about 9%.


Definitely relieve fuel pressure before undoing any fuel fittings on your car. Boiling fuel is especially dangerous but even at room temperature pressurized fuel is bad. It also makes a mess.

I had, what I didn't consider is that with the lines left hooked from the rails to the fuel filter, the fuel was sitting in hot rails and boiling; creating pressure again. I tested this theory later by disconnecting the fuel pump, cranking the engine till no cylinder would fire even once, letting the engine sit for just a moment and cranking again (one or two cylinders would catch and fire)

This may be a dumb question, but what made you think increased fuel pressure or bigger injectors would work?

the same principle behind rising rate fuel pressure regulators; to buy one was somewhere around $100 plus the risk of the fuel pump not being able to keep up; it wasn't a short term solution for me. Increased fuel pressure is increased flow.

Alternatively, the ECU will try to pulse bigger injectors the same in open loop mode (WOT) as smaller injectors, thus increased fuel flow; more fuel in the cylinder, richer mix.

I can confirm this works, but you're the second person that has asked me that question and it's puzzling me why intelligent people are asking me why I thought bigger injectors would work. isn't it obvious why it works?

It seems to me that there are two obvious and fairly easy solutions to the fueling problem: switch to a turbo ECU, or install a rising-rate fuel pressure regulator.

I don't have all of the pieces for the turbo ECU, but I'll keep that in mind; I'd need the ECU itself, the map sensor, the MAF sensor and whatever else I'm forgetting + the wiring back to the ECU to hook it all up. safe to say all that could run me $250 pretty easy even used.

The EPROM in the Hitachi ECUs is a 27C256, pretty standard. I actually poked around an ECU like yours a while back, and didn't find any simple way to alter fueling directly. It doesn't appear to use a fuel map from what I can tell. There is an ignition timing map, however.

The weird blue inverse socket thing around the EPROM is actually wired right to the EPROM. There's a pull-down resistor on the ~CE line, so you can make a little daughterboard that plugs into that socket and ties the original ~CE to Vcc and replaces the original ROM.

A fairly cheap EPROM programmer is the Willem programmer. You can find them on eBay for like $50-$75 or something. But you can also build your own that works off the parallel port with a handful of ICs for almost nothing.

I did read all that in some other threads except for the fueling map bit - I didn't know about that. I thought every ECU had a fuel map. as for building a programmer; I've seen those Willem programmers for about $20 at ebay, the software is worth that much I figure.

[vrg3]

I had, what I didn't consider is that with the lines left hooked from the rails to the fuel filter, the fuel was sitting in hot rails and boiling; creating pressure again. I tested this theory later by disconnecting the fuel pump, cranking the engine till no cylinder would fire even once, letting the engine sit for just a moment and cranking again (one or two cylinders would catch and fire)

Right, but to properly relieve pressure you have to keep doing what you just described until you can crank for several seconds without combustion.

I guess there still could be pressurized fuel after that, though... In any case, your advice to not work on a hot engine's fuel rails is good for another reason too -- some parts, like the exhaust manifold, might actually be hot enough to ignite gasoline.

the same principle behind rising rate fuel pressure regulators; to buy one was somewhere around $100 plus the risk of the fuel pump not being able to keep up; it wasn't a short term solution for me. Increased fuel pressure is increased flow.

But increasing flow across the board is wrong, isn't it?

Alternatively, the ECU will try to pulse bigger injectors the same in open loop mode (WOT) as smaller injectors, thus increased fuel flow; more fuel in the cylinder, richer mix

But the air/fuel ratio needs to be negatively correlated with manifold pressure; that is, you need to run richer as boost increases. That won't happen.

Also, it's not clear that long-term fuel trims don't apply in open loop mode. If it discovers that at low loads the pulse widths constantly need to be trimmed back, it may trim them across the board.

I can confirm this works, but you're the second person that has asked me that question and it's puzzling me why intelligent people are asking me why I thought bigger injectors would work. isn't it obvious why it works?

You can confirm it works? It doesn't seem like you've been getting good reliable results. Or am I misunderstanding?

I don't have all of the pieces for the turbo ECU, but I'll keep that in mind; I'd need the ECU itself, the map sensor, the MAF sensor and whatever else I'm forgetting + the wiring back to the ECU to hook it all up. safe to say all that could run me $250 pretty easy even used.

You already have the correct MAF sensor. Off the top of my head, all I think you'd need is the pressure sensor and pressure exchange solenoid. If you can't find a source for them (there are a couple of Legacy Turbos being parted out in the Parts Shed, though, no?) I've actually documented a cheap alternative to the OEM parts:

http://www.graphics.cornell.edu/~v/chrysler_map/

If you don't want to do a lot of wiring, you don't have to; you can mount the sensor and solenoid somewhere under the dash and just run a vacuum hose through the firewall.

You might have to deal with the IAC valve, though; I'm not sure.

I did read all that in some other threads except for the fueling map bit - I didn't know about that. I thought every ECU had a fuel map.

Maybe it does; I might just have not understood what I was looking at. It's been a few years since I've looked at the disassembled code, so I don't remember exactly what led me to the conclusion that it was calculating pulse widths on-the-fly based on the MAF sensor and a few other things, but I remember feeling pretty confident in that. In some ways, the B10 Hitachi ECU is very old-school.

as for building a programmer; I've seen those Willem programmers for about $20 at ebay, the software is worth that much I figure.

Go for it!

[scuzzy]

the same principle behind rising rate fuel pressure regulators; to buy one was somewhere around $100 plus the risk of the fuel pump not being able to keep up; it wasn't a short term solution for me. Increased fuel pressure is increased flow.

But increasing flow across the board is wrong, isn't it?

It's better than the alternative: too lean when under boost - enough to cause constant and repetitive detonation.

Alternatively, the ECU will try to pulse bigger injectors the same in open loop mode (WOT) as smaller injectors, thus increased fuel flow; more fuel in the cylinder, richer mix

But the air/fuel ratio needs to be negatively correlated with manifold pressure; that is, you need to run richer as boost increases. That won't happen.

I can see why you say that won't happen, but I'm only running wastegate boost. Using the N/A computer, wastegate boost, even at WOT at 5PSI I'm likely still pretty rich in the cylinder; but I'm unable to tell because I goofed my connector I built for the laptop to tie to the ecu.

Also, it's not clear that long-term fuel trims don't apply in open loop mode. If it discovers that at low loads the pulse widths constantly need to be trimmed back, it may trim them across the board.

I can confirm this works, but you're the second person that has asked me that question and it's puzzling me why intelligent people are asking me why I thought bigger injectors would work. isn't it obvious why it works?

You can confirm it works? It doesn't seem like you've been getting good reliable results. Or am I misunderstanding?

I've been getting reliable results in that my engine isn't hesitating, it's pulling nice and strong, no hesitation or loss of power. The only drawback is that it's too goddamn rich at idle.

Long term fuel trims don't apply at WOT on my ecu, back when the laptop connection was working I was watching the fuel trim aspect and going WOT; everything up till about 2/3rds throttle was trimming/adding fuel; but past that fuel trim went to 0% and the engine O2 sensor went to 0.00 volts as I wound the engine out.

I don't have all of the pieces for the turbo ECU, but I'll keep that in mind; I'd need the ECU itself, the map sensor, the MAF sensor and whatever else I'm forgetting + the wiring back to the ECU to hook it all up. safe to say all that could run me $250 pretty easy even used.

You already have the correct MAF sensor. Off the top of my head, all I think you'd need is the pressure sensor and pressure exchange solenoid. If you can't find a source for them (there are a couple of Legacy Turbos being parted out in the Parts Shed, though, no?) I've actually documented a cheap alternative to the OEM parts:

http://www.graphics.cornell.edu/~v/chrysler_map/

If you don't want to do a lot of wiring, you don't have to; you can mount the sensor and solenoid somewhere under the dash and just run a vacuum hose through the firewall.

You might have to deal with the IAC valve, though; I'm not sure.

I did read all that in some other threads except for the fueling map bit - I didn't know about that. I thought every ECU had a fuel map.

Maybe it does; I might just have not understood what I was looking at. It's been a few years since I've looked at the disassembled code, so I don't remember exactly what led me to the conclusion that it was calculating pulse widths on-the-fly based on the MAF sensor and a few other things, but I remember feeling pretty confident in that. In some ways, the B10 Hitachi ECU is very old-school.

as for building a programmer; I've seen those Willem programmers for about $20 at ebay, the software is worth that much I figure.

Go for it!

I'll take your word for the ECU not having a map, if this is going to cost me a lot of money to get sorted I might as well just buy a standalone instead of goofing around with quick fixes here and there. A SAFC-II might get me by, but finding one for cheap brings me back to the original problem: money is better spent on a long term solution (a standalone)

I wish I had the 2 grand right now to dump on it, but I don't; and I don't like the idea of taking out a loan to do it.

[scuzzy]

I had to square up that reply as quick as possible; so here's the addendum to that:

I figure I've got a few options as it stands today.

1: Put the smaller injectors back in, stay out of boost as much as possible.
2: Deal with it. I filled up Tuesday Night, and fell below 1/4" left on the way into work this morning. that's about 12 gallons of fuel in 150 miles - about 14-16mpg.
3: Buy an Apex'i SAFC-II or something similar to fool the ecu into lower MAF voltages to lean the mix.
4: Shit $2000+ for an Autronic (or take out a loan.. meh, bad idea)

Thoughts?

[vrg3]

But increasing flow across the board is wrong, isn't it?

It's better than the alternative: too lean when under boost - enough to cause constant and repetitive detonation.

Fair enough.

But the air/fuel ratio needs to be negatively correlated with manifold pressure; that is, you need to run richer as boost increases. That won't happen.

I can see why you say that won't happen, but I'm only running wastegate boost. Using the N/A computer, wastegate boost, even at WOT at 5PSI I'm likely still pretty rich in the cylinder; but I'm unable to tell because I goofed my connector I built for the laptop to tie to the ecu.

How rich? Even when you hook your laptop up you still will won't know if it's rich enough; the stock oxygen sensor just can't give you that information.

I've been getting reliable results in that my engine isn't hesitating, it's pulling nice and strong, no hesitation or loss of power. The only drawback is that it's too goddamn rich at idle.

Assuming that you are fueling in a manner that's safe enough for an engine running with stock NA spark control, does it stay that way even after hundreds of miles of driving?

Long term fuel trims don't apply at WOT on my ecu, back when the laptop connection was working I was watching the fuel trim aspect and going WOT; everything up till about 2/3rds throttle was trimming/adding fuel; but past that fuel trim went to 0% and the engine O2 sensor went to 0.00 volts as I wound the engine out.

The fuel trim reported by my scan tools is just the short-term fuel trim. The only way I know of to monitor long-term fuel trim would be to record and compare injector pulse widths at given load sites.

I hope the oxygen sensor didn't go to 0 volts -- that would mean the engine was running quite lean. You should only ever see it do that when you release the throttle at high RPM.

1: Put the smaller injectors back in, stay out of boost as much as possible.

If you do this you can try tying the wastegate open with baling wire.

2: Deal with it. I filled up Tuesday Night, and fell below 1/4" left on the way into work this morning. that's about 12 gallons of fuel in 150 miles - about 14-16mpg.

In the short time that might not be a big deal, but in the long term you may find it accelerating wear. Also, the cost of all that extra fuel will add up.

3: Buy an Apex'i SAFC-II or something similar to fool the ecu into lower MAF voltages to lean the mix.

Be aware that this will aggravate the ignition timing problem that you already have. Spark advance should be reduced on boost but your ECU isn't doing that, and lowering MAF signals will make it advance timing even more.

4: Shit $2000+ for an Autronic (or take out a loan.. meh, bad idea)

::shrug:: I think that for fuel anyway a rising rate regulator and a high-pressure pump is actually a decent approach. Sure, you're still going to have too much spark advance on boost, but it'll work for mild boost if you keep it rich enough.

Wait a sec -- are you still using your stock old non-turbo fuel pump? If you are, there's no way you could be supplying enough fuel regardless of your injector and FPR situation, isn't it?

So you'll have to put in a Walbro or something anyway. Right? So an RRFPR is probably the cheapest option after #1.

[scuzzy]

no, I didn't change the fuel pump - I didn't think the demand would be that excessive. what are the differences between the N/A pump and the turbo pump, do we know? how much more does the turbo pump flow?

the rails are the same, that I know.

[vrg3]

The NA pump is rated to flow 80 liters per hour at 36.3 psi. The turbo pump is rated to flow 150 liters per hour at 43.4 psi. That's a significant difference. Remember that flow drops off as pressure increases.

The stock non-turbo fuel pump could probably support about 150 horsepower on a naturally aspirated engine. But a turbocharged engine requires more fuel and more pressure.

[scuzzy]

The NA pump is rated to flow 80 liters per hour at 36.3 psi. The turbo pump is rated to flow 150 liters per hour at 43.4 psi. That's a significant difference. Remember that flow drops off as pressure increases.

The stock non-turbo fuel pump could probably support about 150 horsepower on a naturally aspirated engine. But a turbocharged engine requires more fuel and more pressure.

270cc injectors x4 flow 64.8 liters/hr at 100% IDC (55.0 l/hr at 85% IDC)

380cc injectors x4 flow 91.2 l/hr at 100% IDC (77.52 l/hr at 85% IDC)

I'm not refuting your information, just stating this for the record. I'll look at getting a Walbro pump next week, do the injector + etc swap over this weekend.