kanurys

I'd put the OEM stat in anyways. Stant makes a direct replacement and a cheapo. Search around. Hook the computer up to a scan tool and read the temp. My point is it might be cooling, not getting hotter.

That's pretty much what I thought, thanks. I couldn't even swap the diff chunk?

Would you show us where you put the ECU and relays? Nice work on the swap. I did kind of the same thing for the lower radiator hose. I used a piece of PVC as a sleeve. The upper one was a Gates 20801. It has the right sized ends and I cut about 2" out of the middle and spliced it for my 4" lift, so you'll probably be good without modifying. You might want to splice it in the middle anyways to be able to rotate it as needed. Do you have a number on the lower hoses you picked up? I'd like to ditch my flex hose for the U.

I found two SVX's that are being parted out near me. Would any of the parts be good upgrades for my 87 4x4 wagon? Can I adapt the LSD with different stubs to work with the wagon?

I was going to try to answer this but Gloyale pretty much sums it up. The other question I have is where are you getting your temp reading from? The modified cluster gauge or a scan tool and the ecu? If the EA cluster gauge is hooked up wrong, it might read backwards.

The LED flashes every time the vss pulses. I think the pulse conversion will be important. Maybe I could sell a few of them?

I got the VSS working as well as the cruise, FINALLY. I ended up putting a 5v regulator and a transistor network together to use the reed switch to control pulses to the ecu and ccu completely separately and isolated so they don't talk to each other. It works like a dream and the car drives perfectly. It pretty much takes the one pulsed signal and gives me two outputs that tie to ground every time the reed switch is open (inverted signal). In the future, I will make a little micro controller which will convert the pulses to several so that I can match the ECU speed to the actual vehicle speed (tire size selector and pulse multiplier). A picaxe chip should be great for this since it's really easy code and completely reprogram-able.

Charge light is fine. I just found it easy to throw a resistor in line in my case because I have an aftermatket voltage gauge. Also, the resistor might be good info for other swap applications. I was just intending to throw some options out there.

Thanks, yeah the B/w is the charge light. Its connected from the alternator to the light to 12v switched. Make sure the light is of proper bulb resistance or just throw a resistor in there if you have a voltage gauge.

I seem to recall that yellow had power at all times through the main fuse box under the hood.

Upon a little more testing I found the following with cruise plugged in: With the cruise off the vss signal pulses between .3v and 3v. With cruise on the signal pulses between 3v and 6v. This is enough to satisfy the ECU and it works great. When the cruise is off the ECU sees a signal below trigger voltage and reads 0 mph, causing all sorts of problems, as stated earlier. Somehow the cruise is putting 3v dc on the vss wire. The plan: Try putting a diode between the vss and cruise allowing the signal to move only to the cruise computer from the reed switch. Then, if the vss wire voltage remains consistent no matter what the state of the cruise, I'll boost the signal with a little switching transistor to 5v. Now I just need a little time...

On the drive to work this morning the ECU was going haywire. Fans turning on individually, timing advance stuck at 45°, intermittent cut-outs, and died twice when coming to a stop, AND the ECU speed reading bounces all over from 0 to 200 mph... I think the ECU VSS pin might be floating when the signal goes low from the reed switch. A pull-down resistor between the cap and the ECU should do the trick. Then I have to put a smaller cap in there so it doesn't keep the signal high for too long. Strange how it worked great last night and totally nuts this morning.

Thanks for your help.

Ok, scratch all the vss conversation from before. I had the ECU hooked up to the correct wire all along. Read on: Upon putting a meter on the reed switch signal at the cruise control computer connector (big one on the cuise box just above the hood latch lever) I found that it puts out 4.75v when closed and 0.3v when open (probably inductance). This should be within the acceptable range for the EJ ECU. I found that if the cruise computer was unplugged that I'd get a speed reading from the ECU. Plugged in, the ECU speed went to 0.0. This led me to believe that the cruise computer was somehow interfering with the more sensitive ECU, perhaps drawing current from it or visa-versa. Fist I tried a diode to try and eliminate current flow in one direction or the other. No luck. Then I tried a capacitor to isolate the ECU from the signal current but still letting it see the voltage changes. That was the ticket. It worked with both 100 micro-farad and 220 micro-farad capacitors. I just had these lying around, but probably any coupling cap will work. I soldered the 220 micro-farad in and shrink tubed it. My cruise control and EJ ECU now both get a speed signal and work great.

Nice description. I'll try tapping in elsewhere as you suggest. Then I'll make sure i'm getting a signal from the cluster, then I'll wire in the voltage divider so the ECU gets a happy voltage. The white plug I'm talking about is the round one on the back of the cluster. I can't remember which wire I tapped into by count.

I hooked the EJ Green/Red wire from the ECU to the Yellow/Red wire on the EA white gauge cluster plug. Is this right? I worked out a voltage divider circuit to make the 12v pulse into 5v at the ECU but still drive the Cruise Control. <Reed switch> | | |----------------<CC> | | <70ohm> | | |----------------<ECU VSS> | | <50ohm> | | | | <GND> The ecu might just think the signal is currently out of its voltage range (if its hooked up right). I can display the speed the ECU is reading vs. the EA speedometer in real time with my OBDII reader. This should be 1/4 the speed the gauge reads, according to Gloyale's post.

What's the word? Any progress? I'm battling a few things with the mixed up model years. I have to figure out how to get the vehicle speed sensor signal to the ECU. Phase II computers don't like the 12v pulse, they like 5v. Also, they want to see 4 or 5 pulses (not clear on that) per spedo rotation. So, after that, I'll be in business.

I finally got off my butt and made a center console cover like jackson-rally, except not as awesome. It fits like a glove, though, and accepts the stock shifter boot.

Yep. Lets just say it was a lucky find.

I mean IACV, not TPS. Anyways, all the VW guys are saying that the obd2 subaru engines put out 5 pulses per revolution. I can build a converter or just tap off of my home brew GPS for the speed signal.

And then there's this: http://www.permoveo.ltd.uk/tabid/122/OBD-Error-Code/P0502/Diagnostics/subaru/Default.aspx " OBD Code P0502refers to The vehicle speed sensor outputs a 4 pulse signal for every revolution of the rotor shaft, which is rotated by the transmission output shaft via the driven gear. After this signal is converted into a more precise rectangular waveform by the waveform shaping circuit inside the combination meter, it is then transmitted to the ECM. The ECM determines the vehicle speed based on the frequency of these pulse signals. Symptoms Possible sumptoms of OBD code P0502 - Engine Light ON (or Service Engine Soon Warning Light) - Speedometer may not work properly - Transmission shifting problem Causes Possible causes of OBD code P0502 - Vehicle speed sensor harness or connectors is open or shorted - Vehicle speed sensor - Combination meter - Faulty ECM. The Error code is generally activated on detection of the following conditions: The signal sent by the vehicle speed sensor to ECM is too low." So here's my question for you guys who know: how many pulses does the reed switch send out for every revolution of the spedometer cable?

Here's this writeup, not that it really helps much. http://www.ultimatesubaru.org/forum/showthread.php?t=98358

Actually, yes. I just scanned it and it's pulling up the vss code as well as a throttle position sensor circuit malfunction code. The TPS is being read, though. I can read an accurate value as I press the accelerator. This one I can figure out. The computer reads a VSS value of 0.0 constantly. I know the cluster is generating a signal because the cruise control works. It's an interface issue. How have you guys hooked up an EA82 cluster to a phase II ECU? What kind of signal is the computer expecting? I mean what is it (and no guesses)? What about grounding the neutral position wire from the ECU? Should I leave this un-grounded? It wouldn't effect the speed signal, though. Even without the speed signal the car still drives in closed loop and get decent mileage.

I have an SPFI pump on there, now. I switched to the FI pump wiring harness on the car, so a switch to the MPFI pump should be easy. Then I can keep the SPFI pump in the car for field repairs. The only codes on the ECU are for the (non-existant) fuel tank temp and pressure sensors and the pressure release valve.

Everything is smoothed out. I have a question for you folks who've done a swap: Why does my engine cut out at 4500 rpm just like it has a governor? It runs flawlessly right up to that rpm under load. It will rev above 5000 rpm in neutral. I'm thinking it's fuel delivery. Needs more. What do you guys think? Is the EA82 FI NA fuel pump different from the EA82 Turbo pump? I can probably grab one from the yard.