log1call

Forget about getting things made or modified... way to expensive and unless it's been done before prone to repeated failure while they get it sussed. The best way to do it is to use off the shelf parts out of some other model..

Having the motor hot helps get them out... apart from the burnt hand problem.

I have quite a few morkshop manuals with drawings in. What year and model is it? Are you on broadband? So you can download the whole manual.

Let it cool down then wash it with a bit of petrol followed by a good hosing with water. Then start the motor and run it for about thirty seconds and turn it off and see if you can see where it's coming from. If it isn't leaking after thirty seconds leave it idling and have a good look for leaks, then if you haven't found it rev it a little while you watch, if that still doesn't show what it is go for a SHORT drive, like two minutes worth then pop the bonnet and have a look. Use a light and mirror to look underneath as well.

Yeah Ok. I'll get skype and have a read up. Monday perhaps then, I'll send you a mail and confirm if/when I've swatted up.

That would be cool alright Gazzy, but I fix computers for a sideline and I've tried connection via remote desktop before and if the other person is on broadband and their router has security it is real hard unless you go and sit side by side and work it all out. Well that's what I have found! If you know how to do the remote thing through a router gazzy, I'd love to lear though. I'm on a wireless router without any security if you want to try and connect to my computer first? And before everyone shouts idiot, I'm in the country and my wireless doesn't get even half way to the road.

It doesn't really matter at this stage which ecu model/id you use, all we are trying to do is get the evoscan to connect to the car. When you hit the "start data logger" button, the top right panel which is labeled "recieved data/"event log", should start to produce text and after a few seconds it should change to a green colour(if it's set to do that in the options) and and text should come up with A1100D. If it says it's connected and you get the A1100D then your laptop is connected to the ecu the right way around, the connection is the right speed and you have the correct type of cable selected. If you don't get a connection it is one of those three that's the problem. Once we get it connected then we start to look for te right rom addresses to ask for information. If you go into the evoscn program file there is a folder called data settings, in there you will se a file called "data", if you rename that as "orignal", then download the file called "A1100D data" from... http://cid-4ca3c3459aaa7f7f.skydrive.live.com/browse.aspx/.Public?uc=1&isFromRichUpload=1 and rename it data and put it into the saved data folder you will find some new addresses to try. The other id in there is for my car and if you look you will see I have got all the trouble codes in there both saved and current as well as individual injectors and a few other things that are not in the standard trouble codes. I have other copies of data with most of the subaru models in but this is my experimental version that Gazzy and I were working on.

So when the evoscan connects does it show your id number A1100D?

Hey team, same old problem I see. The way to connect the wires, is the way that got you the A1100D. I presume that was in the vwrx tools? Connect it and confirm that you can get the id(A1100D), then we can figure why the evoscan got nothing but zeros. That id number A1100D, along with a few other A***** ids are not well supported. Gazzy and I were trying to find the right addresses(that's the "request", number shown in that screenshot of Gazzy's). I found a few but there are still a few important ones to be discovered. If your version of evoscan looks exactly the same as that screenshot then I will send you a file with the best addresses we have found so far. Before that though.... Once you have it connected in vwrx start the evo and check that the program connects. If you make a succesfull connection the "recieved data"/"event log", box at the top right should go green and the ecu id number should be shown. If you get that then you are connected and it's just the addresses that are wrong, or in other words the laptop/program is asking the ecu for data out of the wrong places in the ecu. See how that goes and let us know.

So can you log these things as they are running? Should be easy to see. One injector or even one bank shouldn/t cause a surge should it? The injectors fire too fast to cause a surge that you can time in seconds.

Yeah but just running rich will make it chuff and carry on but not really surge. Rhythmic surge will be ICV or air leak surely.

Oh ok, it's the Afm that's ignored, I knew one of them was not efficient at cranking speed and presumed/thought i'd read it was the crank and cam for some reason. Cheers for that ichi. Nice to chat with some one that doesn't get gnarled if you question them about what they said. I like to get the right info and some buggers just have a real bad attitude if you question anything.

That theory sounds doubtful to me. Fuel pressure is regulated. Mixture is controlled by the ecu not the fuel pressure. If there are no trouble codes I'd suspect an air leak, faulty ISC valve(thermo spring perhaps), or a misadjusted butterfly screw.

Bugger. Well, for what it's worth, this is what I'm basing my theory on... From a subaru training pdf called "Fuel injection and engine managment", page fifteen.... The amount of fuel injected or duration is determined by the following: BASIC DURATION + CORRECTION FACTORS + VOLTAGE CORRECTION Basic duration is determined by comparing QA and engine speed. Correction factors include: Start increment This perhaps? Coolant temperature After start Full increment Acceleration Air, fuel coefficient Voltage correction compensates for the injectors time lag affected by battery voltage. PDF availiable here... http://cid-4ca3c3459aaa7f7f.skydrive.live.com/browse.aspx/.Public?uc=1&isFromRichUpload=1

Yup, I've been lurking. Good work Gazzy. If your sure the ecu doesn't have wires for the diagnostics go to a wrecker and ask if you can cut a plug off the wiring. Then don't often sell the wiring so it should be cheap, well my local is good anyway. Cut the plug with long enough wire if you can. There are two little plastic flaps that fold over the back of the plug to hold all the wires securely, release the latch on the flaps and flip them open then use a tiny screwdriver to release two wires and pull them out, lightgreen/red and darkgreen/white are what's usually used for the diagnostics. Open the flaps on the car's plug, push the wires in, close the flaps. As Gazzy says, you won't do any harm if you connect the cable the wrong way so just experiment.

Hi Ichi, thanks for the responce. So what does tell the ecu to fix the timing at start up? I also understood that the start enrichment came from the start input to the ecu... can you throw any light on how that works? I'm talking regular start up whether hot or cold.

Hi Itchi, Don't want to disagree with the master but... The start signal also influences the mixture at startup. My idea about disregarding the cam and crank sensor came from this quote "Signals from the cam and crank angle sensors are received by the ECM. At engine start the ignition timing is fixed at 10 degrees BTDC. After engine start, ignition timing is influenced by the mass air flow meter, coolant temperature, knock sensor and engine load." That comes from page twelve of http://endwrench.com/images/pdfs/Fuel.pdf I might have missunderstood it though. Perhaps the lack of revs causes the fixed timing? Can you confirm or point me to any other reading material? Cheers.

have a look here.... http://www.clubsub.org.nz/forum/index.php/topic,292.0.html

If you take it out you will see.

Hmm... If you are not too good on electrics it would be better to use a test light to check for power. Depending on the model you either have power there all the time and it gets earthed via the relay, or, the relay turns the power on and it earths after the pump. You have to be carefull not to be reading the tiny current running down to the fuel gauge though, which the multimeter will show as twelve volts even though ther is not enough power to run a motor. A bad connection can also allow twelve volts through but it won't be enough to run the pump. A teast light needs a bit of current so it will be dim if you are looking at the fuel gauge wires with it. Probably the easiest and surest test for you is to pull the pump and try it on a battery or, disconnect it from the car's wiring and run twelve volts to it with jumper leads. If you are using the jumpers you MUST know which wires are pump and which are gauge or you will burn out the gauge and possibly burn the car as well(when the petrol ignites).

To test the ecu and trouble codes you could pull the MAf or TPS plugs then give the motor a crank and you should get a code. Sounds like you have found the problem though. Before you buy a new pump it would be a good idea to check it is the pump and not the relay or wiring. If you can hear the pump it wont be the wiring or relay. If the pump is running then it could be buggered, or, you may have been lower on gas that you thought(these fuel gauges are notorious), and the twenty dollars may not have been enough to cover the pump again. You could try putting more fuel in there.

The fuel pump will only pump for about a second when you first turn the key on, so you need to have someone else do that for you while you see how much fuel comes out. The fuel also needs quite a bit of pressure so after you have checked there is fuel coming out check it again with our thumb over the hose and make sure there is pressure there. The alternator does not run the ignition, the alternator charges the battery and the battery supplies power for the ignition. The trouble codes won't detect a fuel problem although it's a wonder you haven't got crank or cam codes after all the cranking. Don't worry about it though, it's probaly all normal. If you want to check the trouble codes are working pull a sensor off and see that it's code gets set after an attempt at starting with it off.

There are a few cars around with clear connectors as Boostin says. The trouble codes are codes that have been saved. Some of the codes need more than one instance of the fault to get saved as a code. Some codes need more than one trip to sabe them as code. Sometimes problems are intermittent and may only happen after several trips so for a mechanic testing his work the codes may not be set during his one or two road tests, so they have a diagnostic mode which makes the codes get saved the first instance they occur. This speeds up the process of checking your work and it makes a very intermittant problem much easier to find. The diagnostic mode also tests all the circuits, so you hear relays clicking, fans running, have to floor the throttle etc. You need to extract the trouble codes firts, then interperate what they could mean and test the sensors etc involved. After repairs you clear the codes and go for a drive or put it in diagnostic mode and go for a drive. Generally just clearing the codes and driving around will confirm the codes are not coming back.

Are you sure the charging and battery are all in good order? Check for codes. Turn the key on and listen for fuel pump noise. Check for fuel flow AND pressure to engine. Pull plugs and look for fuel and spark. Check compressions.

Black for trouble codes. Green for diagnostic mode. Connect them all to clear the trouble codes.