TheSamba.com :: View topic - No idle control on a Subaru EJ22 conversion. Need a Suby guru.

Sat Sep 10, 2016 10:34 pm

I'm sorting through some running issues with an 1994 Subaru EJ22 professionally installed in an 83' Westy by RMW several years ago. It had several owners after the conversion before me. It appears to have about 5K miles on a rebuilt engine according to receipts. The installation still looks clean and un-mucked with, but I'm learning the setup as I go.

In my my ownership, it has never appeared to have any idle control via the idle air valve. Only manual adjustments to the throttle plate screw (which I know you aren't supposed to touch) will affect idle. There is no RPM enhancement during cold idling, so you have to manually set the screw for ~700 cold and then you get to idle at 1200 hot.

I've pulled apart the entire intake side, replacing vacuum lines and looking for big leaks thinking one big enough would push the idle air valve to one of its control limits. The idle, even at 1000 RPM is also pretty lumpy and the engine bounces around quite a bit, which is another sign of a vacuum leak to me. However, I am now very confident after hose replacements and testing with starting fluid that I have no intake side leaks. My booster also doesn't leak. Also, I can kill the engine by completely closing the throttle screw.

I swapped in a known good idle air valve with no effect.

Compression is about 160 psi in all cylinders. Off idle power seems fine. I've driven one other EJ22 Westy for comparison.

I've adjusted my TPS to make sure the micro switch activates when the throttle closes to tell the ECU to start controlling idle. The TPS also shows correct changing resistance during throttle movement on the other two pins, but I think that's unrelated to idle control.

My coolant temp sensor reads appropriate resistance when hot and cold.

There is no EGR system on this setup. Not sure if it was removed during conversion, or if it never had it, but I know a stuck EGR valve can't be the issue :)

Timing is about 20 deg at idle and seems to appropriately advance during acceleration.

I put fresh spark plugs and wires on with no effect. The old plugs looked great, no fouling.

I probed the signal from the ECU to the idle air valve and the ECU doesn't seem to be modulating the two signal wires going to the valve to facilitate active control. One wire always sits at 12V, one always sits at 0.200 V. No adjusting of the throttle screw, or twisting of the valve itself changes these values. The technical manual says both wires should read fluctuating voltages around 7 or 9 V at 40% duty cycles.

Right now, given the lack of a control signal from the ECU to the idle air valve, I think of the two following possibilities:

1. The ECU is not functioning correctly.

2. Some other sensor or system is not functioning correctly and is telling the ECU not to control the idle for some reason. Based on the tech manual, it seems like those things could be a broken coolant temp switch, a broken TPS microswitch, or some massive vacuum leak, all things that I have checked for.

One other bit of info which may or may not be related. There are a ton of old error codes on the computer. My attempts to erase them via the tech manual procedure have failed. Until I can erase them and see what fresh codes appear, I can't really debug what's on there because I don't know when it was logged.

My inability to erase those codes could be user error, or another sign the ECU is having issues, it's tough to know at this point.

I know this is an esoteric and complex question. I also know there isn't a perfect home for it (Suby forum or VW forum?) I'm just hoping someone else with a converted setup familiar with the control systems might think of something I missed. Otherwise I'm stuck trying to find and swap spare ECU's, which seems like an unlikely solution.

Any help or additional questions would be greatly appreciated!

Sun Sep 11, 2016 7:18 am

Sound like a case of swap and see.

If you can get another manifold maybe see if that can cure things as well. You checked the intake gaskets I assume. It seems that you checked everything

Sun Sep 11, 2016 7:30 am

Search for Subaruvanagon group on yahoo. They have some great knowledge.

Sun Sep 11, 2016 8:24 am

hogan029 wrote: Sound like a case of swap and see.

If you can get another manifold maybe see if that can cure things as well. You checked the intake gaskets I assume. It seems that you checked everything

I hate blindly swapping, but yes it feels like I've gotten to that point.

I extensively tested the intake and throttle body gaskets with starting fluid. I did not break those seals though.

Sun Sep 11, 2016 9:21 am

drj434343 wrote:

...

I probed the signal from the ECU to the idle air valve and the ECU doesn't seem to be modulating the two signal wires going to the valve to facilitate active control. One wire always sits at 12V, one always sits at 0.200 V. No adjusting of the throttle screw, or twisting of the valve itself changes these values. The technical manual says both wires should read fluctuating voltages around 7 or 9 V at 40% duty cycles.

Right now, given the lack of a control signal from the ECU to the idle air valve, I think of the two following possibilities:

1. The ECU is not functioning correctly.

2. Some other sensor or system is not functioning correctly and is telling the ECU not to control the idle for some reason. Based on the tech manual, it seems like those things could be a broken coolant temp switch, a broken TPS microswitch, or some massive vacuum leak, all things that I have checked for.

One other bit of info which may or may not be related. There are a ton of old error codes on the computer. My attempts to erase them via the tech manual procedure have failed. Until I can erase them and see what fresh codes appear, I can't really debug what's on there because I don't know when it was logged.

My inability to erase those codes could be user error, or another sign the ECU is having issues, it's tough to know at this point.

...

^ These are the things that stick out to me from your description.

If it were mine, I'd be sure to test the integrity of the harness and pin connections on the IAC circuit from end to end. High resistance, broken wire, etc. I might even go as far as running all new wires (temporary harness bypass) for this circuit just to be sure.

I'd also go after working on clearing out the codes. Knowing what codes are current/active will certainly help track down the issue. Of course you know this... I'm simply saying I would focus on this point before moving on to swapping known good parts.

-Kevin

Sun Sep 11, 2016 11:13 am

The many codes is because of a lack of proper ground, I have repaired several conversions that have this symptom in my shop. You need a GOOD new ground from the van body to the motor and make sure the engine wiring harness ground is bolted under the intake correctly (this is where the entire system gets its ground, under the intake)... you will be able to sort the codes. You need to put the throttle settings back where they were from the factory as well as put the tps back as it came. You should never be changing these! The idle should only be 5-700rpm. 1000 is way too high. These motors and ecu's are very simple.

You have a ground problem. Keep in mind the ECU gets its grounding from the engine side harness. The best braided ground strap for this application is this:

ENGINE TO BODY BRAIDED GROUND STRAP: NAPA 711699 (10 GAUGE 16"LENGTH)

Install this between the factory ground location on the driver side van firewall and the driver cylinder head, make sure the connections are clean and use dielectric grease.

Sun Sep 11, 2016 4:10 pm

I just cleaned both sides of that ground cable. I am measuring 0.1-0.2 Ohms between any metal point on the engine and the firewall.

Is the ECU to engine ground connector the one under the manifold?

Sun Sep 11, 2016 7:46 pm

drj434343 wrote: I just cleaned both sides of that ground cable. I am measuring 0.1-0.2 Ohms between any metal point on the engine and the firewall.

Is the ECU to engine ground connector the one under the manifold?

It is, and keep in mind the original braided strap will often give a good reading, yet still fail to produce a good enough ground for the ECU. The correct replacement ground strap I listed is like 7 dollars...why mess around? You already did a bunch of work in vain trying to solve the problem. The motor will never run correctly until you completely restore the factory settings on the items I mentioned above, even if you solve what the original problem was.

Sun Sep 11, 2016 8:02 pm

Deleted. Weird double post problem.

Sun Sep 11, 2016 8:03 pm

I'm not saying this is your issue, but I had idle problems. Coming down from highway speed, it would stall. Sometimes, upon startup, it would rev high, like it was about to blow apart. The IAC on a Subaru is a very expensive part. New, from the dealer, half the cost of a used engine. So, troubleshooting with a new one is a gamble. What solved my issue was a new, higher amperage alternator (see alternator thread). Again, I'm not saying this is "the fix", but it worked for me.

http://www.thesamba.com/vw/forum/viewtopic.php?t=601176&highlight=subaru+alternator

I admit my solution may have been just enough to overcome some other issue, but it directly coincided with the other problem going away.

Mon Sep 12, 2016 4:10 am

With multiple codes that won't go away,are you running it with the two green test connectors together? These should be connected for test purposes only.

Mon Sep 12, 2016 7:44 am

I've learned a few more things that is driving me closer to a solution.

First, no I am not running it in test mode. When I want to read the codes, I put it in read mode by connecting the two black connectors and seeing what the memory contains. I just recently learned about how to use the test mode with the green connectors to read current codes, but haven't tried it yet.

I swapped in another ECU and was able to read codes back that made sense. They were 35 (EVAP system) and 24 (IAC circuit).

After examining my IAC more closely, I discovered it came from the PO with the wrong 3 pin unit meant for a AT car. After Googling my ECU number, it appears I actually have a 90/91 system. So one order of business is swapping in the correct 2 pin IAC for a 90/91 MT car.

The only other issue I haven't figured out yet is one ground connector issue. On the 22 pin ECU connector, there are two pins labeled control system GND, pins 11 and 22. Pin 11 reads 0.2 ohms to GND, but pin 22 reads 2.2 Ohms to GND no matter how much I clean the harness ground and the engine ground. I haven't found another contact to clean yet, but this delta in resistance seems like a bad thing.

On a side note, I'm hobbling along with a 92 service manual. If anyone knows where I can get a 90/91 service manual, that would be a big help!

Mon Sep 12, 2016 9:07 am

The IAC for the automatic trans in 90-91 is a 3 pin, the IAC in a 90-91 manual trans is a 2 pin. I saw your post over at the yahoo groups and if you have an ECU that ends with 381 382 383 383F then you have an auto trans ecu, which requires a 3 pin IAC valve. The transmission is the decider in the original car. They are not interchangeable.

You still have a ground problem. There is also a great possibility that you have a couple of bad IAC valves. That is totally common, especially if they have been sitting around. I see this often. Make very sure the IAC valve is set back exactly to where it was from the factory. If the ECU says you have an IAC valve code, then the IAC valve is not functioning correctly...no question.

Pin 11 and 22 should be connected together inside the harness(ground).

Mon Sep 12, 2016 9:12 am

(ECU plug position - wire color - description - destination) (if you have three engine plugs, not two, the pin # is in parenthesis)

A1 - Empty
A2 - Yellow/Red - Ignition on
A3 - Empty
A4 - Empty
A5 - Empty
A6 - White - O2 sensor - O2 sensor plug 4
A7 - Black/Yellow - Water temperature - F26 pin 5 (pin 2 if three plugs)
A8 - Red - Air flow meter - Air flow meter plug pin 1
A9 - Black - Air flow meter - Air flow meter plug pin 4
A10 - White - Air flow meter - Air flow meter plug pin 2
A11 - Black red - Ground
A12 - Empty
A13 - Yellow/Red - Ignition on
A14 - Empty
A15 - Yellow/red - Hot from battery
A16 - Red/Yellow
A17 - Black - Shield for O2 sensor - connects to shield on grey sensor wire
A18 - Empty
A19 - Gray - Air flow meter shield - connects to shield on grey air flow meter wire
A20 - AT/MT selector
A21 - Red/green - Water temperature - Connects to F26 plug pin 6 (pin 3 for three plugs)
A22 - Black/red - Ground

B1 - White - Throttle position - to F26 pin 2 (pin 7 on three plugs)
B2 - Black - Throttle position - to f26 pin 4 (pin 5 on three plugs)
B3 - Red - Throttle position - to F26 pin 3 (pin 4)
B4 - White - Cam position - to cam plug pin 1
B5 - Black - Cam position - to cam plug pin 2
B6 - Grey - Cam shield - to cam plug pin 3
B7 - White - Throttle shield - to throttle position shield
B8 - Blue - Air flow meter - to air flow meter plug pin 5
B9 - Blue/Pink - Park switch
B10 - Light Green - Neutral switch
B11 - Green/Black - Speed sensor
B12 - Yellow - Ignition relay

C1 - Black - Crank angle sens - to crank plug pin 1
C2 - White - Crank angle sens - to crank plug pin 2
C3 - Shield - Crank angle sens - through shield to crank plug pin 3
C4 - Shield - Knock sensor - to shield for knock sensor
C5 - White - Knock sensor - to knock sensor plug
C6 - Brown - Throttle position - to f26 pin 1 (pin 6)
C7 - Light Green/Red - Select monitor
C8 - Green/White - Select monitor
C9 - Blue/Red - A/C compressor - (connect to positive for high idle)
C10 - Red/Yellow - Connect to starter selenoid (hot when starting only)
C11 - Green/Black - California mode switch - goes to f26 pin 7 (pin 12) (grounded for CA mode, determined by engine wiring)
C12 - Red/Black - Read memory - to black "read memory" plug
C13 - Orange - Test mode - to green "test mode" plug
C14 - Empty
C15 - Blue/yellow - Select monitor
C16 - Black/blue - Tachometer

D1 - Black - Idle air control - to f26 pin 12 (pin 14)
D2 - White - Idle air control - to f26 pin 11 (pin 13)
D3 - Empty
D4 - Empty
D5 - Light Green - Ignition relay
D6 - White/Blue - CPC solenoid - to f25 pin 2 (pin 7)
D7 - Empty
D8 - Empty
D9 - Green/white - Ignitor - to ignitor plug pin 2
D10 - Yellow/blue - Ignitor - to ignitor plug pin 1
D11 - Blue/white - Injector #3 - to f25 pin 11 (pin 15)
D12 - Light green - Injector #2 - to f25 pin 10 (pin 14)
D13 - Brown - Injector #1 - to f25 pin 9 (pin 13)
D14 - Black/White - Ground
D15 - Black - Ground
D16 - Empty
D17 - Red/Blue - Fan control
D18 - White - Select monitor
D19 - Red/Yellow - Check engine light
D20 - Empty
D21 - Empty
D22 - Blue - A/C cutout relay
D23 - Light Green/Black - Fuel pump relay
D24 - Black/Yellow - Ground
D25 - Black/Yellow - Ground
D26 - Blue - Injector #4 - to f25 pin 12 (pin 16)

Mon Sep 12, 2016 9:21 am

Also, make sure the pin A20 is grounded, this is the AT/MT identifier and the idle will not stop loping if it is not grounded. (assuming you have a 380/381/382/383/383F ECU)

Mon Sep 12, 2016 9:57 am

My ECU is AA563, which comes up as a 91 Legacy MT. I have 3 engine harness connectors.

Someone clearly had installed the wrong IAC before my ownership, and I've been told the MT ECU will not drive the 3 pin IAC at all. I have a 2 pin IAC in the mail as of this morning.

I will check that A20 is grounded, but can you confirm that a 563 ECU needs this?

Yes, there is still some GND issue on pins 11 and 22, I'll need to spend more time tracking that down.

Mon Sep 12, 2016 10:02 am

You have the right valve. Do not change it. The idle air control valve from a 90/91 automatic is a 3 pin valve. Code 24 is rarely due to a bad IAC. There are countless stories of people replacing the IAC to no avail. Take the IAC apart and clean it. It can get gunked up. Most likely, the ECU thinks the IAC is bad because it can't get the idle within spec for some other reason. Look for something that the ECU does not monitor. For example, your fuel pressure regulator could be bad or its vacuum unplugged making the idle high. Or an intake leak or worn throttle. ETC, etc.

Mon Sep 12, 2016 10:11 am

drj434343 wrote: My ECU is AA563, which comes up as a 91 Legacy MT. I have 3 engine harness connectors.

Someone clearly had installed the wrong IAC before my ownership, and I've been told the MT ECU will not drive the 3 pin IAC at all. I have a 2 pin IAC in the mail as of this morning.

I will check that A20 is grounded, but can you confirm that a 563 ECU needs this?

Yes, there is still some GND issue on pins 11 and 22, I'll need to spend more time tracking that down.

Disregard my above post. I had thought you had an automatic ECU. The PO must have compiled the conversion from several different donor cars which is definitely frowned upon. Let us know how it turns out once you get the right IAC in there.

Mon Sep 12, 2016 10:52 am

2wdvanagon wrote: You have the right valve. Do not change it. The idle air control valve from a 90/91 automatic is a 3 pin valve. Code 24 is rarely due to a bad IAC. There are countless stories of people replacing the IAC to no avail. Take the IAC apart and clean it. It can get gunked up. Most likely, the ECU thinks the IAC is bad because it can't get the idle within spec for some other reason. Look for something that the ECU does not monitor. For example, your fuel pressure regulator could be bad or its vacuum unplugged making the idle high. Or an intake leak or worn throttle. ETC, etc.

Im sorry to disagree, but this is incorrect on OBD1 cars. The code is specifically thrown when the resistance value is not correct on the IAC itself. Ive been working on these cars specifically for 18 years, and I have never seen an OBD1 code that was anything but the component electrically not falling within range. On 1995 and newer Subaru, you would be correct in your assumption, but 94 and older it is either the component itself or the wiring circuit of the component, quite a simple diagnostic code symptom.

Now, he may actually have a problem with that portion of the harness instead of the IAC, and he should check that for sure.

Mon Sep 12, 2016 10:52 am

Is the engine a '94, & the ECU a '91?

I had an issue on a swap before (not on a vanagon).

I installed a later engine in my wife's car (same engine displacement / specs), but I used the original ECU. I had idle problems. Long story short, the later engine had some slight engine changes, even though the HP & TQ were the same. The head passages were opened up a little. They installed larger injectors because of it. Because of the larger injectors on the late engine, they had to modify the injector pulse width in the ECU (shorter duration). My original ECU used the smaller injectors & longer duration pulse width. So even though the engine ran great, the idle was always surging / revving.

I would confirm your injectors match your ECU