A new ignition switch is around $20, the one you have in the picture shows some wear. On my car, if the battery voltage fell below 12.3 volts with the worn ignition switch, it wouldn't even fire up the starter relay. Replaced the ignition switch, and I could start the car even if the battery was down to 12.0 volts, I didn't test any lower as to not damage my battery. Anyhow, that's my two bits.

 

You've done an excellent job of tracing out the powerflow of the system, locating all the components, and determining how it all works. What does strike me through is the testing methods seem to be pretty intrusive, and shy on data gathered. Seems to me you'd get a much better picture much faster of what's going on simply employing a test lamp and volt meter. Confirm with actual voltage readings that the ECU is providing a strong ground path for the control side of the relay, and measure the voltage at the power side. If getting a drop, check different sections of the circuit to hone in on the high resistance. Using those methods will take out the guesswork and confirm without doubt you've found your problem, and confirm your fix.

 

That guy ^^^^ knows his way around our cars "attic".
Make that THOSE two guys...

I know throwing parts at a car is NOT the right way to fix it, but his suggestion makes sense and his reputation alone is worth the $20.
And I think you can handle what Larry has suggested...

Good Luck Charlie (and Happy New Year [since I haven't spoken with you lately])

Bob.

 

After reading your description of extensive troubleshooting and about what you’ve tried, I find some unusual inconsistencies concerning the circuit involved. This makes me wonder if you’ve confused some of the results or if an intermittent problem might be in play as well.

If power at pin 7 of MFI is momentarily interrupted and all else correct, the relay would simply blip open and closed so fast that I question if the blip in pressure from pump would be long enough to stall engine.

On the other hand if ground from pin 5 (White/red wire) was lost relay would open and stay that way until ground was restored. I know you think it’s a power delivery problem but I’m not sure especially since you couldn’t detect voltage drop with multi-meter when problem occurred.

Might try running a temporary ground wire from pin 5 (White/red wire) to any good chassis ground point to see if that corrects issue, if it does then might need to troubleshoot why ECU is dropping that ground.

 

Thanks for the replies... Any kind of advice is appreciated. I've been dealing with this for 3 months. Its winter now in Vermont so I have about 3 more months to figure it out.

I will order 2 OEM ignition switches Monday just because. Will be nice to rule that out 100% and to have a spare.

I did test power at the ECU pins 12 and 25. I also tested continuity to ground from pins 91, 13, and 26 all the way back to the negative post on the battery. I had measured voltage at pin 7. It would be 14.x volts and then when the car died it would drop to 12.x volts. I also cleaned the ground that bolts to the firewall near the battery.

I didn't test anything with pin 5 because when I had jumped pin 7 to 12v it didn't die on me. I have about 5 warmups (2 from the relay jumped directly to the harness on the ignition and 3 jumped from the connector on back of fuse box) where it didn't die. I just went and pulled pin 5 from the connector though and I ran a ground wire to the relay connector. AEM would not behave. My fuel pump would prime forever (instead of just 1 second) and it didn't want to stay running at all. I wonder if AEM somehow reads feedback from that pin?

I made a video of my multimeter monitoring pin 5:

And here is a video of it monitoring pin 7. I watched it in slow motion and I did see it dip to 11.27v but I can't tell if that is after car shuts off or before. Thinking about it maybe I should repeat the video with just turning the key off:

 

The monitoring of pin 5 appears to show a problem to me. With key on showing 11.86v is normal but when key is turned to start and after engine running that voltage should drop to ~ 3v or so and stay there until key is turned off, or until ECU drops ground which should cause it to return to about the 12v range. It appears when engine starts yours looks like it’s going to ~ 128v, which makes no sense to me. I don’t know anything about AEM units, maybe someone that knows those units can offer an explanation.

Pin 7 video looks normal, at least what I’d expect to see.

 

I've had a hard time telling what's going on on my screen, but it looked to me like it was switching to MV scale.

 

Where the hell is @Steve '93ES ?

You have the best helpers...

You got this...

Bob.

 

Yeah that is millivolts (mv)

 

Chances are it's your AEM ECU that's faulty

As has already been said, you should be using a voltmeter, waste of time swapping anything until you know where the fault is, that's what the meter is for.
The AEM uses a switched earth to operate the MFI relay, that's where you should start. Use clipleads so that the meter is permanently attached so you can monitor the various wires one at a time.

 

I'm here. It takes me a minute to catch my breath after carrying Don's tool box...
I'm a bit confused though. Like Don I know nothing about the AEM, but if jumping MFI pin 7 to power fixes the problem (even temporally) why are we looking at pin 5?
Grounding pin 5 BTW will force the fuel pump on any time the key is on (pump would prime forever (instead of just 1 second) I'm not sure why it wouldn't run like that unless the AEM does monitor that pin...

If this were a stock system I would be wondering if an intermittent crank sensor was causing the ECU to drop the MFI signal..
You say the AEM logs that though? Is crank signal something we can rule out?

I cant quite see the buttons on your meter but if you can select a range instead of letting it "auto range" you might clean up some of those funny readings..

 

I'm here. It takes me a minute to catch my breath after carrying Don's tool box...

You need to start carrying my brain; it’s a lot lighter these days.

Like Don I know nothing about the AEM, but if jumping MFI pin 7 to power fixes the problem (even temporally) why are we looking at pin 5?

I mentioned that because he indicted relay was opening without any significant drop seen on multi-meter (assuming taken at pin 7 of relay) and I knew that couldn’t cause relay to open, so I questioned dropping of the ground.

If this were a stock system I would be wondering if an intermittent crank sensor was causing the ECU to drop the MFI signal..

Since you already brought up crank sensor while I was typing, I’ll show it below any way.

Since pin 7 and assuming pin 5 test are in order, it might be something else causing engine to die, which would cause the ECU to drop fuel pump side of relay after few seconds. Just for giggles, take a look at wire coming from crank sensor up front of timing belt cover, to make sure the accessory belt hasn’t cut into it.

 

Not to ask the obvious, but what are your main relay input and output settings in AEM? If the ignition switch is suspect, then you can perform a "test" by trial and temporarily raise your "stop engine" duration in the options menu (with the rest adjusted accordingly...)

-sent from my Galaxy Note 9

 

perform a "test" by trial and temporarily raise your "stop engine" duration in the options menu (with the rest adjusted accordingly...)

Hmm, I haven't messed with that. It's set to 0.5 seconds (what AEM delivers it as) and it's been that way for 10 years. You are saying if I make it higher maybe that will give me some time to see something else happening? Here is a pic.

take a look at wire coming from crank sensor up front of timing belt cover, to make sure the accessory belt hasn’t cut into it.

I'll double check that, but I know the zip tie push pin was in tact. I also went through the factory service tests for the cam and crank sensor before sending my ECU in.


Tomorrow I'll see if there is a way to log the sensor input with AEM. Here is a picture to show why I didn't think it was AEM calling for ECU to turn off. You can see where the RPM drops and the MFI Main Relay and MFI Fuel Relay are still on. I know its the fuel pump relay turning off because I labelled each side of the relay and I can see it physically opening back up.

Its worth noting that if I rev the car to 3-4k rpm this can still happen. When it does the car will act like fuel cut but come back to life as the engine is still spinning fast enough. You can hear the relay click. I put my calibration file and 2 log files in a zip file if anyone wants to take a look. In one of the log files the car shuts off idling. The other I'm trying to hold the rpm steady at 3-4k and it drops.

https://802projects.com/images/random/MFI/aem.zip

 

@ fastfalcon94 I got to tell you that with the below list you can’t get any better information.

BlackStealth
RealMcCoy
OhioSpyderman
Steve ‘93ES
bboyalan

That’s a hell of a knowledge list of minds about these cars.

 

 

I just went out and logged the cam/crank sensors. Before doing so I set the Main Relay shut off time to 3.0 Seconds instead of 0.5. From what I can tell they are logging correctly. If anyone wants to see the log it is available here: https://802projects.com/images/random/MFI/aemlog1.zip

Now I couldn't go to the Advance Pickup Tab and Log. I forgot to mention that. If I "start PC logging" and the car dies, when I go up to the logging menu "Stop and Save" is greyed out. It loses the log. Maybe the ECU resets or loses power too as that same power wire also goes to pin 82 of the ECU. So in order to log the pickups through a shut off event I need to setup the channels using the built in logger. So thats what I did prior.

One thing I did notice about this log as the first event I really see that looks abnormal is the ignition timing. Before the rpms drop the ignition timing goes from 25 to 10 degrees. 10 degrees is what my "Crank Advance" timing is set to so maybe some of the reporting bug in AEM. The rpms drop more or less at the same time and maybe its just seeing the rpms drop below my Crank Exit RPM and showing 10 degrees.

Also I ground pin 5 of the MFI relay again. The fuel pump stays on as expected. This time I was able to start the car. I think last time it just shut off immediately and I didn't give it another try. I reproduced the issue 4 times with PIN 5 grounded. The only difference is you don't hear the relay click before the car shuts off because the ground cable is forcing it to be latched all the time. Video here of me doing it twice.

I'm still trying to figure out what this means. The same power wire that goes to Pin 7 does drive Switch 1 of AEM. I am not logging that in the latest log file, but I was in the two prior ones at the fastest sampling rate. It never changes anywhere near the shut off.

 

I'm still trying to figure out what this means. The same power wire that goes to Pin 7 does drive Switch 1 of AEM. I am not logging that in the latest log file, but I was in the two prior ones at the fastest sampling rate. It never changes anywhere near the shut off.

Watching that video makes me wonder what pin you where reading voltage on, pin 7 maybe (can’t tell for sure). But if so that would be correct and as expected, the voltage drop from ~ 14v to about 12v after engine quit is simply alternator supplied voltage (~14v) to battery supplied voltage (~12v). Since engine died with pin 5 still grounded confirms to me that something else is causing engine failure other than MFI related, something I suspected early on and the more knowledgeable RealMcCoy referred to in his post #16.

Not knowing how to read AEM data chart, I thought something looked suspicious about cam sensor counts, I see bboyalan made comment about that and being curious I’ll bring that up with him so I might learn something as well as you.

 

IMO, your cam tooth channel looks suspect, but I would not say it is definitive yet since something else could be at play i.e. affecting it (edit: also see that you are zoomed in). I just diagnosed a bad crank sensor on a Stealth three weeks ago to justify replacing it - been purring smooth since then. Part of the troubleshooting included the cam sensor, of course. You can easily rule either one out from just the startup event:

-sent from my Galaxy Note 9

 

IMO, your cam tooth channel looks suspect,

Thanks for your knowledgeable input, not knowing anything about AEM system I too thought cam sensor count didn’t seem right as compared to crank sensor count.

Does the AEM actually read each flag hump of reluctor plates as (1) count? Like 4 counts of cam plate per revolution and 3 counts of crank plate per revolution?

 

Not sure how I got added to that list in one of the above posts, but I have to tell you....YOU GUYS ROCK!!!!

It's EXTREMELY nice to know we have members here like @bboyalan @RealMcCoy @white93gt @Steve '93ES and others....

Thank you for being here....

Bob.

 

You might, maybe, have better luck monitoring voltage by selecting the range instead of "auto ranging" .
On most meters that will significantly improve it's response..

 

I'll give that a try. I just did another test, this time purposely stalling the car in 6th gear with the laptop datalogging running. I was able to "stop and save" afterwards. This leads me to believe that pin 82 of ECU is seeing a voltage drop as the real issue. The wiring goes to both the ECU and the MFI relay, so same wire I've been dealing with. Maybe the problem could be in the section that goes between ECU and wiring leading to MFI relay though.

 

Kind of makes me curious what's happening on the other side of the MFI, pin 6.
If pin 6 is also shutting off ( losing ground, same as pin 5) that would kill the ignition.
In other words if the AEM were trying to shut down wouldn't it kill both relays?

 

The one thing that kinda sucks about AEM is the lack of failure diagnostics... There is no such thing as a trouble code with that platform. They do somewhat make up for it with very complete and powerful datalogging, so you just have to know what you're looking for, and looking at...

At this point I'd want to get a scope on that cam sensor... Alan and I have discussed the matter in the background a bit, and I'm convinced that given the same time base and testing window, the crank and cam sensor counts should always maintain a 1.5:1 ratio. (in a a 720 degree firing cycle, there are 6 crank pulses to 4 cam pulses.) If that is the case, you're missing 6-7 cam pulses in the test window under the cursor on that log...

 

I may have to wait until my oscilloscope comes to find out more. I messed with the range on my multimeter but I don't think there is anything I can do to make it more accurate or sample better. On youtube you can click settings and change the playback speed. If you watch some of my videos at 0.25 speed (turn volume down), the last video actually dips to 9.9V. I don't know if this is normal because car is stalling, or if it is the cause for the stall.


The other side of the MFI relay remains active when the car dies. I have the cover off and physically watched it. Also now that I'm grounding pin 5 I no longer hear any clicks. I only hear the MFI side click after I turn the key off and wait a few seconds. When I have a better measuring tool I'll try to grab the wire that runs to the fuel pump. That controls another relay for my fuel pump hotwire so I don't think it would cause much of a draw.




I'm glad multiple people are trying are working on this in the background I've been trying to wrap my head around what you just said... Working it through I think the numbers come out right but maybe I'm not looking at the right thing.

• 4 stroke cycle = crank spins twice as fast as cams
• plate on crank gear has 3 tabs on it
• plate on rear cam gear has 4 tabs on it
• AEM Crank is set to trigger on rising and falling edge
• AEM Cam is set to trigger off of falling edge only
• One 720 degree cycle = crank gear spinning twice = 6 tabs, would see 12 edges
• Camshafts spin once = 4 tabs, would see 4 edges though as it only counts falling edge

AEM Cam Count - "Total count of cam signal significant edges from the Cam (T2) input. Resets when value reaches 255"
AEM Crank Count - "Total count of crank signal significant edges from the Crank (T1) input. Resets when value reaches 255"
AEM Cam Tooth - "(Unit=Teeth) Displays the number of teeth found before the cam tooth is received"

If I place the cursor after the cam tooth resets to 0 and look at the crank/cam count between that one and the next one I get a Crank count of 12 significant edges and a Cam count of 4 significant edges. That matches what I would expect. If the cam saw rising/falling edge then it would be 12:8 which is how I think you got to 6:4 then 3:2 then 1.5:1?

- Charlie

 

Just out of curiosity what kind of oscilloscope did you get? I'm hankering for one myself, but those $50 Tektronix e-bay finds have been eluding me...
As far as setting your meter's range, it's not immediately noticeable. What it does is stop it from dancing around and displaying weird things like "-28v ac " out of nowhere. In auto range it has to re- calibrate to any major voltage change. With the range selected it just displays it. That way If the voltage dips, instead of going through it's routine it just shows it. Trust me, when looking for dips and spikes you don't want your meter hunting around looking for range..

 

Nice!
Mr. Cooper is right, floating the scope is a bad idea. I apologize for posting it.

 

The only way to isloate a scope is to use an isolating transformer.
For a car isolation isn't required, you just earth the probe to the body of the car.
In most countries it's illegal to remove the earth from the power lead, someone could get killed.

 

This? Do you know exactly what his fuse problem was? I'm not blowing fuses or anything.

"After reading this post again, maybe you should monitor voltage at end of this bypass connection. I remember OhioSpyderman having an explainable fuse problem because of this bypass connection. "



So originally I had the fuel pump control relay that is below the engine bay fuse box in tact. If you look at the 2nd picture on the first page in this thread you'll see that ECU pin 31 controls when the relay is triggered or not. When its active it would use the built in resistor box, when its not active it wouldn't. Pin 31 is not used by AEM so the underhood resistor box is never used. I had it this way for 10 years. After this issue started I decided to remove the relay (bypass it by connecting pin 2 to pin 5). I also unplugged the resistor box and removed it.

So pin 1 of the MFI relay goes out to the engine bay, through the bypass, all the way to the back of the car where it triggers a relay for my fuel pump hotwire kit. I did test that relay and also tried a new relay before I really started digging into the wiring diagrams.

Also looks like oscilloscope will be here on Friday.

 

So pin 1 of the MFI relay goes out to the engine bay, through the bypass, all the way to the back of the car where it triggers a relay for my fuel pump hotwire kit.

That’s what I originally thought when I read your first post and looked at diagram. What thru me off was in your post #35 where you said, “When I have a better measuring tool I'll try to grab the wire that runs to the fuel pump. That controls another relay for my fuel pump hotwire so I don't think it would cause much of a draw.” by saying controls another relay lead me to think the fuel pump relay was still in place.

I understand now that was as I expected to look like, just thrown off by wording.

 

My fuse problem was with fuse #12.
I did the relay bypass and after several years (for whatever reason) there was intermittent arcing between two terminals.
I bought an auto-reset fuse that allowed me to troubleshoot without going through dozens of fuses.
I solved the problem by generously applying di-electric grease to all the terminals in that relay holder.
Damnedest thing I ever saw....

Bob.

EDIT: When I saw you post about fuse #12, I realized I put the wrong fuse number here.
In fact, there is NO fuse #15.
I need a vacation....

 

Interesting. My next test will be to pull pin 1 from he MFI relay and jumper my hotwire relay in the hatch so the fuel pump runs constantly. That should eliminate all the wiring that goes in/out of the engine bay including the factory fuel pump resistor bypass relay socket. We'll see if that helps. None of my fuses have blown. Tested with multimeter and I've swapped them out too as a precaution.