How do you test the ECU?

[urdrwho]

I'm just wondering if anyone knows how to test the ECU? Aside from swapping for a known good one is there any method to check it?

 

Is there a ground strap for the ECU?

[monkeyposeur]

Kobayashi Maru is the only test I know of.

[urdrwho]

I would have to cheat.

 

Kobayashi Maru is the only test I know of.

[sunny5280]

I'm just wondering if anyone knows how to test the ECU? Aside from swapping for a known good one is there any method to check it?

 

Is there a ground strap for the ECU?

 

While knowing very little about this subject wrt to Subaru's I'm going to throw my opinion out there anyway: It's unlikely there's a way to test the ECU without specialized test equipment designed to test it. Perhaps there's some internal diagnostics test. If there is I have to ask how can you be certain the results are accurate?

 

I say this because a friend of mine owns a 1999 Grand Cherokee which would suddenly stop running. Typically this was after the vehicle had warmed up on hot days. After it had cooled it generally would start right back up. Until one day it didn't.

 

The vehicle has a means to display stored codes by turning the key on and off three times in succession. During the period of time when the vehicle would not run the ECU would not display any codes (or the message stating no codes were stored). This led me to the conclusion "How can I trust the ECU to store codes about a failing ECU?" In the end I ended up buying a new ECU (found one online for about $150) and that solved the problem. Unfortunately swapping in a new ECU, at least on the Grand Cherokee, is not possible as the ECU has to be coded to the vehicle.

[spooln30]

What are you trying to test for? Yes the ECU can be tested via a multi meter or my favorite THE POWER PROBE!

[urdrwho]

I just get weird stuff that happens.

 

Just one example. I've had my P304 code ever since I got the car. Sometimes when I am getting my lumpy idle while sitting in my driveway the code will appear. I immediately turn off the car and attach the scanner. It will show no current, no stored or pending codes. I turn the car back on and the code is gone. So either there was a P304 or there wasn't a P304 but if the CEL was illuminated the scanner should show a current, stored or pending code.

 

This morning the P304 code was set, it was lit for 7 miles and was so until I stopped the car. Went inside for 2 hours, came out started the car and the CEL was not lit. Car was still running as it was when I parked it but no CEL.

 

Odd things like the above.

 

From what I have been told the car sat for 4 years waiting for an engine transplant. They transplanted a 1999 Forester into my 97 Outlook. There are places that the car looks like someone along the way used it to go through streams. There is also an intermittent leak in the windshield directly above the ECU. It sometimes drips on the dash but I wonder if it ever ran down the inside of the windshield, down the inside of the firewall and onto the ECU.

 

 

What are you trying to test for? Yes the ECU can be tested via a multi meter or my favorite THE POWER PROBE!

[Stang70Fastback]

I would have to cheat.

 

Well then you have failed to divine the purpose of the test. The purpose is to experience fear, fear in the face of random fault codes, to accept that fear, and maintain control of oneself and one's vehicle. This is the quality expected in every Subaru owner.

[lockmedic]

In the wild, ECU failures are typically diagnosed by process of elimination. . . or at least, they should be. Eliminate all other possible causes of the fault and if it persists, replace the ECU.

 

Unless you're one of the mechanics around here, in which case you automatically assume that any problem is caused by a faulty ECU, so you swap it in, then call me to "flash it" because now it won't run at all. I load up the proper VIN and calibration files like a good little stooge, marry up the airbag and theft deterrent components, get it started and running, then deal with a pissed off 'technician' when they realize that no, a new ECU will not cure a noisy wheel bearing.

 

---dramatization, but not far from the truth. Also note that Subaru ECU's through abour 2004 do not require 'flashing', etc. . . plug and play.

 

Not sure how they cobbled a phase 2 SOHC engine (which is what the 99 Forester came with) into your car unless they only used the short block.

 

That aside, it sounds like you need to clean your idle air control valve. But hey, change the ECU while you're in there.

[urdrwho]

Idle control valve is a nice shiny probably 98% new one.

 

http://legacygt.com/forums/showthread.php/great-bones-yard-find-217087.html?t=217087

 

It has been determined that I do have a franken-motor.

 

http://legacygt.com/forums/showthread.php/coolant-hoses-need-parts-numbers-216797.html?t=216797

 

"Possibly (if so, it's like the motor that came in the '99 Outback), or you do have a '98 Forester engine (as Osei suggested)."

 

"That motor actually came out of a Forester, and I believe only Forester with a "tow package" came with that setup. I also believe that you have a Phase 2 bottom end there, with a Phase 1 top end. Search '99 to '02 Foresters for that part. "

 

 

 

 

In the wild, ECU failures are typically diagnosed by process of elimination. . . or at least, they should be. Eliminate all other possible causes of the fault and if it persists, replace the ECU.

 

Unless you're one of the mechanics around here, in which case you automatically assume that any problem is caused by a faulty ECU, so you swap it in, then call me to "flash it" because now it won't run at all. I load up the proper VIN and calibration files like a good little stooge, marry up the airbag and theft deterrent components, get it started and running, then deal with a pissed off 'technician' when they realize that no, a new ECU will not cure a noisy wheel bearing.

 

---dramatization, but not far from the truth. Also note that Subaru ECU's through abour 2004 do not require 'flashing', etc. . . plug and play.

 

Not sure how they cobbled a phase 2 SOHC engine (which is what the 99 Forester came with) into your car unless they only used the short block.

 

That aside, it sounds like you need to clean your idle air control valve. But hey, change the ECU while you're in there.

[twisty]

call me to "flash it" because now it won't run at all. I load up the proper VIN and calibration files like a good little stooge, marry up the airbag and theft deterrent components, get it started and running, then deal with a pissed off 'technician' when they realize that no, a new ECU will not cure a noisy wheel bearing.

 

not to go too far off the subject, but ive always been interested in how the ECUs are flashed. is there any info you can toss my way, like a software to look at or something? is there a magical scanport for it or something? what language is it written in? simple ladder diagram?

 

 

as for ECU issues, ive replaced a few, but only when i was having intermittent (or constant) issues with accessing it with a scanner, as well as operational problems with the vehicle itself. older vehicles seem to be much more receptive to ECU replacements.

[lockmedic]

I really just do GM and some Chrysler and only to OEM spec. With GM, I use a Tech 2, although you can use a lot of J2354 devices these days (mongoose cable, Hotwire, Cardaq, etc.). The Tech 2 is easier for me as it's a handheld scan tool, so I can check codes and view live data before and after flashing it to see if it's corrected the issue.

 

I don't actually look at the ECU dumps. . . I just pull the files off the OEM websites and put them on the car.

 

GM has adopted a 'modularity' strategy where one part can work for a lot of different vehicles, but it has to be loaded with the correct software once it's installed. I'm sure it makes things easier on them from a supply standpoint, but it sucks on the repair side as there are some pretty silly things that are CAN-equipped, like power window switches on Trailblazers. They no workee until you program 'em.

 

--thread derailed!

 

back on topic, "clean and shiny" doesn't always mean "good" when it comes to engine parts. If it's aftermarket, there's also a chance it's not correct for your frankenmotor. Short block should be irrelevant, but I know Subaru changed IAC valves from 98-03 like they were getting paid to redesign the things.

[urdrwho]

I cleaned the previous IAC very well, applied a new gasket and I had the same issues.

 

I just find it very, very odd that the car can throw a code like P304 but not store that code.

 

OR when sitting in a parking lot, the code disappears when I stop and immediately restart the engine. When the previous happens, when restarting the car will have a high idle 1200 > 1500.

 

Or that code will be there when I park the car but can disappear when I come back out to start the car.

 

Off topic for a moment ----------------------

 

Speaking of GM. I have a 1995 Caddie Deville Concours. Even though it is an OBDI it is amazing what can be done with the on-board diagnostic system. The built-in diagnostic service mode, press the OFF and WARMER buttons are simultaneously held. Example on one item -- PCM data, you can look at all the various sensor inputs, short-term and long-term fuel trim, injector pulse width, engine rpm, battery voltage and PROM identification. Under PCM INPUTS, you can check switch status (throttle position switch, brake pedal switch, transmission gear switches, etc.). Use the PCM OUTPUTS mode to check the EVAP and EGR systems, torque converter clutch solenoid, A/C compressor relay and transaxle solenoids.

 

You can even disable individual fuel injectors.

 

There is more but you can access a lot of information without a scan tool.

It was kind of nice of GM to do it.

 

I really just do GM and some Chrysler and only to OEM spec. With GM, I use a Tech 2, although you can use a lot of J2354 devices these days (mongoose cable, Hotwire, Cardaq, etc.). The Tech 2 is easier for me as it's a handheld scan tool, so I can check codes and view live data before and after flashing it to see if it's corrected the issue.

 

I don't actually look at the ECU dumps. . . I just pull the files off the OEM websites and put them on the car.

 

GM has adopted a 'modularity' strategy where one part can work for a lot of different vehicles, but it has to be loaded with the correct software once it's installed. I'm sure it makes things easier on them from a supply standpoint, but it sucks on the repair side as there are some pretty silly things that are CAN-equipped, like power window switches on Trailblazers. They no workee until you program 'em.

 

--thread derailed!

 

back on topic, "clean and shiny" doesn't always mean "good" when it comes to engine parts. If it's aftermarket, there's also a chance it's not correct for your frankenmotor. Short block should be irrelevant, but I know Subaru changed IAC valves from 98-03 like they were getting paid to redesign the things.

[sunny5280]

I cleaned the previous IAC very well, applied a new gasket and I had the same issues.

 

I just find it very, very odd that the car can throw a code like P304 but not store that code.

 

OR when sitting in a parking lot, the code disappears when I stop and immediately restart the engine. When the previous happens, when restarting the car will have a high idle 1200 > 1500.

 

Or that code will be there when I park the car but can disappear when I come back out to start the car.

If this were my vehicle I would be suspecting the ECU as well. It's the same kind of behavior I observed with my friends Grand Cherokee. With that said I see you have a 1999 engine in a 1997 car. Have you ruled out any issues with the engine swap?

[urdrwho]

Aside from a bad harness I'm just not sure which issues to check.

 

I know of a nice 99 sitting in the bones yard and wonder how much they want for the ECU?

 

If this were my vehicle I would be suspecting the ECU as well. It's the same kind of behavior I observed with my friends Grand Cherokee. With that said I see you have a 1999 engine in a 1997 car. Have you ruled out any issues with the engine swap?

[spooln30]

While a bad ECU is possible I believe it's another issue. Get a can of carb spray and while running spray where the intake manifold and cylinder heads meet. Your checking for a bad intake manifold gasket doing this. If the engine doesn't want to die while spraying I would check for another common issue. For this you'll have to drop down the exhaust header and look into the exhaust ports on heads. What your looking for is a exhaust valve guide that has dropped down more then the others or a 1/8" down. What happens is the guide becomes loose and will work it's way down until the point it won't let the valve fully close causing a misfire and when the valve hits the guide it will push it back into place. I owned a 96 L and mine did this exact problem. It would misfire sometimes then go away not storing codes all of the time but the CEL would flash. Mine would happen when a heavier load was on the engine. Worth checking out both IMO.

[sunny5280]

While a bad ECU is possible I believe it's another issue. Get a can of carb spray and while running spray where the intake manifold and cylinder heads meet. Your checking for a bad intake manifold gasket doing this. If the engine doesn't want to die while spraying I would check for another common issue. For this you'll have to drop down the exhaust header and look into the exhaust ports on heads. What your looking for is a exhaust valve guide that has dropped down more then the others or a 1/8" down. What happens is the guide becomes loose and will work it's way down until the point it won't let the valve fully close causing a misfire and when the valve hits the guide it will push it back into place. I owned a 96 L and mine did this exact problem. It would misfire sometimes then go away not storing codes all of the time but the CEL would flash. Mine would happen when a heavier load was on the engine. Worth checking out both IMO.

 

It's my understanding the computer stores codes for a period of time if the issue which caused it to illuminate falls back into spec. This allows the problem to be diagnosed dispite the fact the CEL is no longer illuminated. Eventually the code will be purged but it does take some time. Is this not the case with Subaru? If codes are not being retained that would imply a malfunction in the computer.