MGB: Modern(ish) fuel injection problem.

Well, I'm no wizard with FI of any vintage. That's why I prefer to mess with SU's.

Anyway, my '97 Geo Metro has started and run fine since I bought it a couple of months ago.I had to replace a couple of sections of fuel feed and return lines back at the gas tank due to corrosion. Installed a new fuel filter at the same time. The other day, on a short run, I stopped to pick up the mail and as I stopped, the engine die, as if I'd turned off the key. The starter spun the motor over fine, but there was just 1 momentary sign of ignition, then nothing. Several more attempts, and I arranged to trailer it home. I figured it was probably the fuel pump gone AWOL.

Prepared to drop the tank to get at the pump, but thankfully decided to listen for the pump in the quiet of my driveway. Could hear it's faint hum. Not sure if quieter than usual, never noticed before. Turning to the start position was like yesterday, no "fire in the hole".

Next, I disconnected the hose to the filter. Key on. Pumped fine. Reconnected the line.

Next, disconnected the hose to the throttle body. Key on. Pumped fine (into an OSH bottle). Turned key to "Start" and it actually fired and ran for a few seconds, then died. Rather like a 'B motor does with jammed SU's and a bit of fuel poured into the carb throat! Was also able to blow through the fuel line, so it's not blocked back to the tank.

Figuring it had somehow "cured" itself, I reconnected the pressure line to the TB, jumped in and turned the key, fully expecting it to start. No such luck! As before, no fire!

Next I disconnected the return line at the TB and was able to blow through, so not blocked.

Next, I disconnected the pressure line again and again was to get fire for a couple of seconds. This with NO pressure on the injector.

Anyway, I'm baffled. Anyone got any ideas? This doesn't appear to be a sophisticated FI setup, but over my head at this point!

TIA

PS I do have a manual. Trouble is, it's the fancy 2 volume factory set which has got me suffering from information overload! Wish I had a Chilton and Haynes to refer to as well!

You need to measure pressure and volume of fuel out of the pump to test it. First check for engine codes. Then test fuel pressure and volume then check for spark.

I haven't got a code reader and I don't know if there's another way to extract any stored in memory. The Check Engine light has never been on, other than at startup.

For fuel pressure, I'm baffled as to why/how the engine will run for a couple of seconds with the pressure line disconnected, so effectively zero pressure. For flow, at least into an OSH bottle, there seems to be plenty. Fills a 12oz bottle in 5 or 6 seconds.

Spark? Well, I'm getting it for a few seconds.

Guess it's off to the store to try to borrow or buy a fuel pressure gauge.

Derek,
I don't know if this will help, but I had a similar problem with a 95 Volvo, and like you, I thought it was the fuel pump. Turned out to be an intermittently failing electronic ignition module. I had spark at the plug, and plenty of fuel pressure, but no start. After 3 days of research and head scratching, I discovered that when the module fails, the injectors flood the cylinders. Then when the module works again, no start due to flooded cylinders.

The fix was to remove the plugs, and disconnect the injectors, then with the throttle wide open, crank the engine for 30 seconds at a time (to prevent overheating the starter). After doing this about six times, reinstalling the plugs, and hooking the injectors back up, it started right up.

Even if it starts up several times with no apparent problem, you will need to replace the module, or you will find yourself stranded (guess how I know?).

Another common problem that often gives either reduced power or failure to start is the Air Flow Meter (sometimes known as the Air Mass Sensor). It's the unit mounted on the air intake to the engine and gives the ECU information such as air flow, density, temperature etc.
When my last one failed I took live readings from it and found that it was telling the ECU that the ambient air temp flowing into the engine was 94deg Centigrade!

I would recommend that your first port of call would be a friendly mechanic with a code reader else you could be part swapping forever and a day with modern F.I systems and still not resolve it.

Well, I just got back from Canadian Tire and, as I guessed would happen, discovered that the exact same fuel pressure tester that costs $39.99 at Advance in the US is available to me up here for $59.99. Screwed again. Equus #3640. So I'll probably use this one and buy another when I'm next down South and return it to CT. I hate getting screwed without dinner first.

Anyway, the plan is to start 1st thing in the morning. I'll pull the plugs to make sure the cylinders aren't flooded and then check the fuel pressure. The trouble shooting chart says that if it's below 23 psi to replace the pump, if over 45 psi to replace the pressure regulator. If the pressure's OK, I'll start chasing a problem in the distributor/control module.

I've mentioned to a few people that it runs for a couple of seconds with the pressure line disconnected and they all looked puzzled after I convince them that it was in fact the pressure and not the return line that I disconnected! .:

Good job, keep at it.

I find it very rewarding to troubleshoot something like this. If you guess and throw parts at it you never really know what the problem was, or if it's really fixed. Besides, learning new things keeps life interesting.

have you checked spark while cranking and to make sure you are getting injector pulse.....if you have fuel pressure and it should be pretty high not for sure but should be upwards of 35-45 lbs just right off hand.....but sounds like you might have an intermitten spark due to bad crank sensor or control module. If you have fire and injector pulse while you are cranking then most likely it is low fuel pressure, but you need to check to see if the computer has any stored codes and go from there

OK, back to the fight. I installed the fuel pressure gauge in the pressure line. Turn the key "On" and the gauge jumps to 28psi (to spec., 23-30psi), then when the pump stops (approx. 2 seconds) the pressure drops back to zero. Pinching off the fuel return line and the same thing happens, except the pressure goes to 50 psi.

The test procedure says it should not drop below 13psi after 1 minute. If not, it instructs to check the injector and hoses for leaks. Well, I put a piece of paper towel under the injector while I turned the key and it ended up gas soaked, so it looks like I'm shopping for an injector.

Anyone disagree with the diagnosis? I've never had an injector fail. Do they all of a sudden give up like that? If I put 12v across the 2 terminals of the injector and hear it click, could it still somehow be good?

Advance has them list for between $72 (Python) and $343 (Beck/Arnley). Go figure! I'm guessing I'll have a hard time finding on for under $200 in my area, so I'll probably be hunting in the junkyards, where I'm guessing they'll probably ask $50. Will probably rant later!

Well, rant time!

NAPA - $500
GM Dealer - $437 (non in Canada!)
AutoValue - $341
- $60 Reconditioned (1 week delivery)
Junkyard - $75 for whole Throttle Body

Guess where I'm heading! Might pick up a Control module and MAP sensor, too.

You sure your fuel pump didn't fail the test? Didn't follow the whole thing there, but that didn't sound good. Many of the do have check valves and such, that could have gone south.

Also, check and see if you have a fuel pressure regulator in the line, many do. At least my truck does. That could be the problem.

I would sort more before I go start thinking injectors and such.

As a note, look through your manual carefully, sometimes you can read faults by using a simple volt meter and a jumper wire. Not always true, but often is.

This afternoon I get to pry faults out of my Ford truck....hoping it is simple.

BumbleB74 Wrote:

Derek up North Wrote:

OK, here's the update. I bought and installed a complete throttle body from a junkyard car (75k miles) that I heard running.

Damn! no change! Won't start, but will run for a second or two with the fuel pressure line disconnected, like before. However, here's what I have discovered. While redoing the fuel pressure checks (pumps up to 28psi fine but drops as soon as the pump stops), I've found that the injector is getting battery voltage all the time! So it stays open and fuel dumps into the cylinders.

If I'm not very much mistaken, the injector should be getting pulses to open and close it. So, anyone care to hazard a guess as to why I'm getting a constant signal? As soon as I turn the key to the "Run" position and the pump runs for it's 2 seconds and after it stops, I'm seeing just over 12v. Duff computer (PCM)? Crank position sensor? Cam position sensor?

Don't worry, I'm not planning to throw new parts at it. So far I've only spent $45 at the junkyard for the throttle body and an unknown condition "Ignitor" and MAP sensor, yet to be installed.

Derek
I am no expert but here's my experience. Typically a constant 12v supplied to the injector when the ignition is turned on. The ECM pulses to ground to complete the circuit. The pulse timing and size is based on engine speed and load. There should be no pulses or fuel flow when the engine is not spinning. You can disconnect the wires to the injector, turn on the key to run the fuel pump and check if the injector is still leaking. If it is leaking when the wires are disconnected, the injector is stuck open and needs cleaned. If it stops leaking after you unplug the wires you could use an injector test light or analog voltmeter to verify the ECM is pulsing as it should. If the ECM is keeping the injector open all the time, I would suspect that the ECM is bad or receiving a false signal that the engine is spinning.

With the system leaking down so fast, I would suspect that the injector is stuck open. Since this has happened to two injectors, is it possible that some contaminates broke loose in the fuel lines and is getting into the injectors? Maybe a bad filter that is letting dirt through?

If you need your injector cleaned and do not have a place nearby, you might want to check out http://witchhunter.com/. They do quality work and I have never seen a bad review of them.

OK, with the injector power unplugged, the fuel pressure goes to 28psi and holds there initially, then slowly, slowly drops. The test procedure says it should still be over 13psi after 1 minute, which it is. If I leave the key in the "Run" position and plug the injector back in, I can hear the injector open and the fuel spray into the TB. With the injector unplugged the 12v+ remains constant, at least as displayed by my Fluke 21 multimeter.

Just went and checked again, and the injector has a constant ground, with the key in all positions. 12v+ appears when I turn the key to "Run".

Well, thinking it might be the crank position sensor that might be sending wrong info to the computer, I removed the sensor and it changed nothing (ie, still 12v+ @ the injector). Putting a jumper across the sensor changed nothing.

I've never messed with a crank position sensor before, so I put my multimeter across the 2 terminals. Found around 0.25 ohms (not continuity). Moving a nail past the magnetized sensing end changed nothing. I was rather expecting a magnetic switch to be opening and closing. Anyone disagree?

Sounds like it is an electrical problem for sure. Do your manuals have wiring diagrams? If so, can you find if the ECM supplies the 12v or the ground? You could unplug the ECM to see if your ground or 12v goes away. Whether it provides voltage or ground, it should not be firing the injector when the engine is not spinning. If it provides the ground side like most other ECMs, check to see if the wire from the injector to ECM is shorted somewhere.

GM Crank Triggers usually output a small voltage as you move metal by them but I do not know if Suzuki is the same. I can put a LED across a GM trigger and get it to light as metal moves by it. The faster the pulses, the brighter the light.

So you figure, obviously, that the crank sensor might be supposed to be sending a signal, rather than just opening and closing a switch? Hadn't thought of that. Don't have a diode lying around.

I just went out and pulled the sensor again. With the Fluke across the terminals, I read 0 VDC. Moving a nail past the head gave a small voltage reading. Don't think I saw over 0.020 VDC.

Update on my problem with my F150, not to hijack.

Couldn't get any fault codes out of it, and then it wouldn't even start. AAA towed it to the dealer last night, and so far, they say it runs great for them. What worries me is that they are a darn good shop too.

"You want us to keep it until Monday to play some more".....heck yeah, I drive across a 5-mile bridge twice a day, that has no pull offs, and everybody runs 70mph.

Broke down once on that thing, and just a miracle it was early on a Sat. morning or I would have been splat!

What engine do you have in this car?

Have you had a chance to check the wire from Injector to ECM for a short to ground? Tried unplugging the ECM to see if the continuous ground goes away? Have you checked all the other grounds and connectors? Can you tell from your diagram if the ECM is supplying the 12v or the ground for the injector?

I am just going off my knowledge of how GM systems work. The voltage output is pretty low at low speed. Here is what the GM V6 trigger output looks like on an oscilloscope.

Haven't had a chance to check the ECM connections yet.

The engine is the ground shaking 1.0l 3 cylinder from Suzuki, so I'm not sure how much like a GM it will end up being.

Derek, how's this mystery going? Did you find a fix, or just give it a rest?