Cougar

Your a true fighter Matt. You have been through the ringer and are still fighting. Way to go. I wish there was a way we could test the spare ECU. It may have the same problem as the original one but it would be a pretty remote chance I would venture. Hopefully you will come up with something with the voltage readings. If they are ok then I would try replacing the cam sensor since that is the only thing left that hasn't been changed. I can think of one other test we could do on the ECU if you want to do that. We haven't checked the resistances looking back into the ECU on pins 13 and 14 and compared them. It may tell us if the ECU's are bad. Remove the plug to the ECU and with a digital meter check the resistances between pins 13 and 45(ground pin) and 14 and 45. You may have to use the diode setting to get a reading.

If I have read all the posts correctly, here is what we have up to now. First let me say that the problem we are trying to solve right now is why the ignition is not working (no spark) on cylinders 3 and 4. Other than the cam and crank sensors, there are no other engine sensors that need to be looked at since they are the only ones that can cause this kind of problem. From previous tests done we have proved that the circuity is ok from the input of the ignitor to the plugs, including the two wires between the ECU and the ignitor. This leaves the ECU, the two sensors, and the cogs as the possibilities for the problem. We have also tried replacing the ECU with one of unknown condition that gave the same results. PacMan suggested looking at the power to the ECU and I agree it is good to verify that it is ok. My reference gives me some different points to check for power though than from what he stated. My data shows that pins 15, 16, and 42 should be tied to power. It is interesting that power goes to pins 15 and 16. These are right next to the ignitor pins 13 and 14 and it appears 13 is the faulty pin connection that is causing the trouble since the voltage is 10 times lower than pin 14. Since these are mirrored circuits they should be the same. Matt, if your still with us, check the power to those 3 leads using a meter. Also check the connector for a problem, like a dirty contact, on those pins (13 thru 16). Matt (I hope your still with us), you stated that you changed the crank sensor with no success. If you haven't changed out the cam sensor yet I would try that. Also check the polarity of the connections to the sensors. I read from another source that it makes a difference. To 9pec: Great post about the ignition.

An IACV is a idle air control valve. They can be a problem but not in this case. My best guess is the CAS sensor is not working like it should and replacing it hopefully will solve the problem. Nothing else will cause the ignition to do what it is doing except for what I mentioned previously.

Well it looks like a lot has gone on since I've been gone. After looking at the new posts here is what I think. Mjez, your idea about testing the ignitor was good and I also wanted to try something similar to that (I would have added a series current limiting resistor of maybe around 4.7k ohms) but like Nipper said, you can't just touch things like that to full battery voltage and expect it not to damage the circuit. It looks like you got away with it on this one though and have also proved that the ignition is good after the ECU pin connections 13 and 14. This is great to know. This means that the problem is either with the ECU, one the sensors, or one of the cogs for them. All of the other things like the TPS, coolant sensor, MAF sensor, or anything else mentioned, have nothing to do with the ignition not working on two of the cylinders. If you can get another CAS sensor cheaply you might try replacing yours with it. The sensors can get flakey.

Very good find. Thanks, I will use it.

Hi Nipper, I see you are online. Any comments on this trouble. You always have some good input.

Ok. Now we are getting somewhere. Pins 13 and 14 of the ECU control the ignitor for each side of the engine. The voltage readings should be the same for each side relative to ground. It appears there is a significant difference here. One side is firing and one side isn't. Since you have already changed out the ECU and it made no change then it looks like the input to the ECU is the problem. Again, you need to physically look at the cam and crank cogs to see if they are damaged. My bet is some of the tabs are broken. Refer to the earlier link to the picture, for reference to the crank cog at least.

Ok. If the new ECU doesn't change anything and the new ignitor doesn't change the results then the crank and cam cog tabs need to be checked.

Please tell me what the readings are on those pins I told you about. Tell me what the replacement ignitor pins read also. It may help also to take voltage readings with the ignition on. Check pins 1 and 3 and then 2 and 3. If the voltages are significantly different then this may mean the ignitor input is bad. I don't think plugging in the other ECU will hurt anything as long as the plugs match up ok.

The grounding is not the problem since the ignition ground is common to both sides of the coil and half of it is working. The grounding of the engine is relative to the negative side of the battery, which is tied to the engine.

I left out one thing in my last post. The problem may be with the ignitor module also. Check the wiring between the ignitor and the coil for a problem. EDIT: Check to make sure that the following wires have good continunity since they control the ignitor. On the ECU, pin 13 (yel/vio) should tie to pin 1 of the ignitor and ECU pin 14 (yel/blu) should tie to pin 2. You could also check the AC voltages on ignitor pins 1 and 2 with reference to ground while the engine is turning over. If you read the same voltages then it would appear the ignitor is the problem since the input to it would seem ok. With the car off, remove the connector to the ignitor and check the resistances between ignitor pins 1 and 3(ground) and then 2 and 3. If the readings are close to the same the ignitor may be ok but I can't say that for sure.

Here is a link to a picture of the crank position sensor. Look at the tabs on the outside of the gear. It is these tabs that get broken off and cause this kind of trouble. Using a meter on the sensor itself will not tell you if all the tabs are really ok. You need to check them physically. http://www.main.experiencetherave.com:8080/subaru/images/timingbelt/DCP_2703.JPG Thanks to Legacy777 for the picture. As far as checking a relay that was mentioned in the last post, I have no clue what relay could cause this kind of trouble. To me, the trouble has to be either with one of the sensor gears or in the ECU.

Using a meter to test the sensor will not show up broken tabs. You will still see an output from the good ones. A scope with the proper triggering is the the only sure way I can think you could tell if the output was good or not. You should visually inspect the tabs on the gear to see if they are ok or not.

I suspect that the problem may be due to some broken taps on the CPS sensor. I would check that out.

You may want to double check the CPS sensor to see if it is damaged. A couple of the tips may be broken off and causing the problem.

Using a tap makes it so easy to add a extra fused circuit. A good auto parts store or marine parts outlet should have them.

You should have OBD2. The socket should be down near the steering column. Here is a link that shows what the outline of the socket looks like. http://pinouts.ru/connector/16_pin_car_OBD2_special_connector.shtml

If I was doing this I would use a fuse tap in the fuse panel to provide the connection. Looking at my reference it shows SBF-1 is a 30A fuse that supplies power to the sun roof. It also supplies power to fuse #11 in the fuse panel which is a 20A fuse to supply power for the door locks. By putting a tap on the hot side of fuse #11 it will be taking power from SBF-1. You can purchase fuse taps that replace the original fuse and have a slot to put the fuse back in along with another fuse to power your addition. By using a tap you don't alter the wiring.

It sounds like the fan switch may have an intermittent short in it. You may have to replace it later. Glad your other problems have corrected themselves. Since the two gauges were on the same power lead, as are some other things with them, it seemed likely that the connector may be the problem. Thanks for the update. Drive On.

So you know that it isn't a fuel problem then. That would leave the ignition to check and then the MAF sensor, if the ignition is ok. Check to see if you have spark immediately after it quits.

I think the normal thermostat temperature is set around 190 degrees. You could try replacing the thermostat with a new OEM unit to see if that will fix this problem. The radiator may be partially plugged up so replacing it may need to be done. You can also have a block check test done to see if exhaust gases are getting into the coolant.

It sounds like you have swapped a plug into the wrong socket or have not plugged it in correctly to the proper connection. Check the plug wiring from the fan switch.

You can do it. The numbers should be marked on the plug and socket. You will need to take the voltage readings while the connection is together, not seperated. The problem may just be a bad connection at the socket and you may fix it just by disconnecting and reconnecting the connection. If you have an analog meter, set it up to read DC volts greater than 12 volts. Perhaps you have a 25 volt scale you can set it to.

The gauge meter itself is electrical. The sensor in the tank is electro-mechanical. As the float in the tank changes in level there is a change in the resistance of the sensor circuit. That changes the voltage at the gauge and makes the meter move.

The fuel pump may be shutting down on the car. Try spraying some starter fluid into the intake when it starts to die on you. If the engine trys to keep going then you know you have a fuel problem.