tahuyahick

Well, after a nice tall glass of home-made triple shot coffee & cocoa-mix... I decided to ignore the lack of sensor-safe sealer and electrical-taped the AFM cover back into place and just threw that back into the car. Ignored the lack of a clip to hold the harness connector to the AFM, more tape please! Funny, when I try turning the key on, I see a code 11, 12 & 14 on the first cycle, then it's just codes 11 & 12. Maybe I did something right inside of that AFM! I wasn't really sure about that part:grin:! And I decided to unplug a mysterious green connector next to the computer. Hey! With the key on but not started, the pump stays off. So I arrange the fuel pressure gauge where I can see it and try cranking the engine to see if the pump kicks on, oh and the cars starts! Woo-Hoo! And the pump does indeed kick on and stay running. And the engine is running kinda normal! So I drop some tools, run around up front and try bumping the gas by pushing on the throttle body end and the throttle immediately sticks wide open and whadda ya' know, it revs up all crazy-like, and it starts cutting of momentarily like there's some kind of rev limiter and something under the dash starts beeping. So I shut it off. With a quickness! Yeah, the throttle cable got mis-routed at the throttle body. A restart using the gas pedal this time, and it's almost normal. So I check the fuel pressure and it's back to 60 plus. Darn! Ok, one problem is I keep forgetting the vacuum hose to the regulator. That's one point for the leaks, home team still struggling. Pressure is holding at 60 with the engine off, so I remove the return line to check if it's a blockage on that. But nope, just a slow drip coming from the regulator. So I must've not quite fixed that thing, or maybe there's some debris in the lines that's sticking the regulator closed. Yeah, right. Debris sticking valve issues usually mean the valve is unable to close and the pressure would be too low. I probably need to re-think the regulator fix. Maybe an aftermarket unit is in the works. But I can drive the car as it is now. That's progress.!!!

Well if I have to shim the regulator again, it's not really that big a deal. At least it isn't 60 plus! Hey! I just thought of something, if the starting enrichment wasn't kicking-in but the pressure was at 60, then it would be enriched anyways, but would stay that way after cranking. Might be why it started to begin with. It's too bad I didn't look at the codes from the beginning. But I bet the code 12 was an issue before I got the car. Well, I haven't done a thing yet to the car. It was nice enough outside I had to take my dirtbike out in the woods. But the next two days are my weekend, so maybe I'll get something done around the house...

Uhm, fuel-pressure above manifold? Why did that give me a troubling feeling?:-\ I'm reading pressure from a standard 100 psi pressure gauge connected via a t-fitting inserted between the fuel filter and the engine. The book says 30 to 36 with the vacuum line disconnected & plugged, holding at 30 when the pump is off. The book also specifies that the regulator should be holding 30 psi after the engine is turned off for five minutes. According to the book I have, it should not be 36 above the manifold pressure level while running. But it will be that way when the engine is off! A lack of starting enrichment...darn pesky computer! Why can't it just understand what I'm thinking?

But after looking further into this, it's not actually a curve that's being outputted. If I go this route, I'd be using a hotwire sensor and it's a moving vane that creates the curve, but the resistor network is very carefully designed & laser-tuned to correct the value and output a linear signal. This era of computer is setup to expect a linear input to designate the volume of airflow, which is absolutely perfect! The newer computers had their software changed to use a cheaper AFM that outputs a log signal which the computer must then convert into the correct value. So, because the earlier AFM was designed to output an actual airflow value, it would be very easy to use a simple hotwire controller that calculates an airflow value and only make sure the voltage range is matched correctly. I'm supposed to sell this car, so I might not actually do this. But if anyone else would be interested in trying this out, I'd be very curious in what results!

Well, I just connected an ATX power supply across pins 4 & 5. The loaded supply voltage is 11.33 Setting my DMM to a manual range selection and connected to pin 2, I see 0.94 at the wipers rest position, moving the vane slowly results in a very smooth increase in output voltage ending in 7.23 volts at the highest travel. There were no reverses in voltage rise, or any kind of troublesome spots in the output signal, so I'm gluing this shut with some dessicant and move on to testing the harness wiring for the AFM.

http://the944.com/afm.htm Doh! - "You can not use an ohmmeter to test an AFM's potentiometer. It will jump about meaninglessly. You must test it as it is being used. You have to force a voltage UV and measure the output voltage UP. This can be done with the AFM in the car and the ignition on." Uh, you mean like what I've been seeing all this time?

Ok, I understand the intent. I'll have to ponder the ramifications of this circuit, it looks very useful for other applications. So if I understand that correctly, pin 4 is getting the regulated 5-volt supply positive. Pins 3 & 2 will show as a ratio of each other which will stay the same ( for a particular vane position) no matter what happens to the voltage level at pin 4. I haven't checked the links yet. I'll do that when I re-arrange my note taking. Thanks! I've got a lot of testing to do before I try cranking the engine over again. ps - did that formula give the definition on those variables? U-Fix, U-S & U-B are in volts, right? What's Q? CFM? And A? Is that a conversion factor?

I've got one of those simplified diagrams that tend to really annoy me. But if I think of it as a stand-alone harness, then everything is shown. Just not in the locations. And a real miracle - my diagram has numbers on the EGI pins! More notes to write-up here, thanks again. You really get the homework done! Awesome! I should've tried this already, I can connect a testlight to pins 3 & 4 and then use my DMM to measure voltages. This is showing what I measured on my AFM circuit board: It sure doesn't look very complicated when you see it like this.

Ahhh! My head thinks on it's own! - ok seeing as I measured the resistance between each point on the circuit board to find combined values for the resistor network, and having noted that the values are in a staggered decreasing value starting at 69.2 ohms/network section at zero airflow and ending with 1 ohm at the last network section. Well it's likely a series of straight lines, that when connected together form that curve. I'm actually just learning to program these controllers so I think I would prefer to just step through a series of linear stages to map an output signal. Maybe with that approach I could use a rotary encoder, or better yet switch to a hot-wire sensor and gain back that 7.3% torque loss from the vane! Of course I would rather do this with a car that runs perfect.:-\

Darn, it's on a curve! That might mean programming a $20 logic controller for a sensor interface. I'll just shelve that idea for a more desperate rainy day!

I'm just being funny. Makes sense that Bosch is involved. That's just Subaru making a sensible design decision to get their fuel injection up to speed. Nice techie page, thanks. I have some questions on the EGI pinout, it's a lot of answering. Is there a link to a description for each pin? For example, what's a Ti monitor?, what's the D-check jumper?, is throttle switch idle/full/ground at the gas pedal or on the throttle-body?, what's the pressure switch monitoring?, why is there an altitude signal when clearly there's a barometric sensor inside the EGI as a daughterboard?, is the power source on a key-switched circuit?, what's the F/ICCD?, is the line end code output going to a test connector? what does the 'distinction between california and 49 states' pin do? And many thanks so far! If you were ever to scan that FSM to a CD or DVD - I've got cold hard cash I could Paypal you!

I just remembered, what's the purpose of pins 3 & 4? One should be a voltage supply. I don't understand the function of having another pin with a resistor between 3 & 4. It's like a ballast getting bypassed during starting. Anything like that going on?

On a BMW? Did somebody 'borrow' somebodys design? I thought you meant he went at the carbon track with a chunk of 220 or something. I used a paper towel soaked in Deoxit D5 for the track, rubbing carefully. Then a piece of wetted 3M 600 to clean the contact surface on the wipers. I also shifted the board to get into a new swipe-zone. Didn't make enough of a change on the ohm meter, and I think I did about the same as the bimmer guy. But I hadn't thought of the dessicant yet, or the sensor-safe RTV silicone sealer. Good points there. So, does the voltage on the signal output ramp-up in a linear mode? I could get really serious and convert the whole deal over to a different kind of sensor. I'd like to see a chart on vane position/voltage output.

Awesome! Thanks!

Golly, I need to write-up a check-list! I can try the eraser trick, but I wouldn't want to really create some more dead-spots! I still don't know exactly what the test connectors look like, there's a bunch of stuff under the dash. Maybe there's a reson I don't see two sets of unplugged connectors? I only see one pair that look like the thing. Ok, I'll check on that Coolant temp sensor, and if it wasn't just that wire I had disconnected on the starter, another check there.. I took out the board in the ECU and looked it over. No visible signs of solder-joint breaks. Looks very clean too. I like that live voltage check idea on the MAF. I was just thinking about hooking power up to it on the bench and wiring the signal to an experimental analog gauge I built. It reads from 2 volts up to 6 volts in it's current configuration. I don't trust my digital meter to correctly show a live variable, always a nice thing about an analog meter. My analog can go from zero to full scale in just over a second, which is a sweep of 270 degrees. I'm trying to get it to move faster but that stuff is 'top secret squirrel' only!