Good to now fuel has been ruled out.

I just bought a new ballast resistor for wife's '58 ****** with Ford 260 V8. Your problem reminds me of how old engines would start enjoying full voltage from the starter bypassing the ballast resistor. If the ballast resistor fried open, the engine would start up off the starter then die. Is there a clue there? Yes. Jag has a similar starting circuit! More later.

There are four wires to the coils, a common ground, primary timing to ECU, misfire feedback and power source from ignition relay. Chances are good the ignition relay is going open for some reason, cutting out power. The relay gets its signal from .White/Blue wire #61 that runs to the Engine Management Fuse Box 10A fuse #14. All engine wires, #53-57, are powered via 4 different fuses, from the Engine Management Power Bus in the Engine Management Fuse Box. They're all designated inside a circle with a capital E. The means these are powered when the ignition switch is in the RUN position II and in STAR position II, or at least they're supposed to be. What are the chances wires #53-57 are all hot in the START but dead in the RUN? I'll wager the chances are pretty good. EMS bus wire 5E gets its juice through EMS CONTROL RELAY which is probably OK since the engine fires up. IF you can isolate the relay, see if it clicks when someone turns the key to run. Blue wire ECM20-6 dictates when the relay is energized, and that's probably not the culprit. This is where the trail gets cold for the Ignition Switched Power Supply +.

The ignition switch grounds power. There are sure a lot of blue white wires all of which are designated in a hexagon. Ignition switch position III is a blue/white critter that runs runs right to the BPM and we can conclude that it has been doing its job rolling the starter, pumping fuel and sparking plugs. Blue/white wire wire FC4-3 runs from the ignition switch II terminal a junction splitting off out to 6 blue/white wires. One of those splits off to the inertia switch that has a green/white wire and red/green wire. The red/green eventually joins up with 5 others. So ignition switch position 2 reaches out to Switched Ground wires 2 through 19.

It's past my bedtime, my wife isn't happy with me ignoring her and Acrobat is having trouble with the electrical diagrams, so I don't have the liberty of tracking down all those leads, but I will say, wires blue/white #5 & red/green #8 are the most suspicious. Blue/white #5 runs from the BPM FC14-15. Quick test- disconnect battery, switch ignition to run and check FC14-15 to ground. Should be grounded. #8 runs from the Normally Open contacts of the Inertia Switch to BPM FC14-33. See if this is grounded. It should be open.

Nothing runs from the ignition switch or inertia switch to the ECM so I'm willing to bet, the ECM is not getting a run command. So why would the engine start if there's a communication problem? Because the signal to hit the starter, fuel pump and coils does not come directly from the BPM, KTM, SLCM, transmission, ignition switch..... but guess what, it comes from the wire to the starting solenoid similar to vintage cars! The ECM "Engine Start" is tied in with the BPM wire to the starter solenoid!

 

You have done a brilliant job dissecting the circuit diagram and giving a test procedure.
The similarity to a broken ballast resistor was pointed out in post 47.
Great work

 

It was TLDR...

Lots of modern cars will spin the engine regardless of key & security but won't inject fuel if those fail, thus a no start.

 

It is possible that the EMS control relay is dropping out, but that should be easy enough to check for B+ on any of the engine management outputs when the car is in the stalled state.

It has still not been established what happens at the point of stall, as Don notes. It could be loss of ignition or injectors. Either (or both) would cause a sudden stall.

It's also worth recounting that the current symptoms are totally different to those which prompted the ECM replacement.

 

Just thinking this through without reviewing the documentation, it seems that the ECM must be directly involved in engine startup because it controls ignition via coil timing/firing and fueling via injector pulse timing/duration. So jrnsr's theory that the ECM is getting a Start command but not a Run command is worth investigation.

Also, Michael's point about the injectors possibly being shut off is worth investigation.

 

I have finally figured out how to upload a video clip to this forum. It might help with the understanding.
There was a mention earlier of perhaps a low voltage condition due to the multiple starts but I put a new battery in it last ear and it lives on a C-Tek so there is always good voltage.

 

The BPM issues a 'request to crank' to the ECM after its security dialogue with the involved modules. The ECM fires the starter relay in response to complete the action if the shift is in park or neutral.

I imagine the ECM derives information from the crank & cam sensors to determine that the motor is running.

 

If the Engine Management Power Bus relay does not click when the key moves to RUN position, a quick switch of the common relay can rule out bad relay of course. Then wires to the relay can be tracked down. The red wire EM70 at that fuse box should be hot at all times to energize the relay after the ECM command. The Engine Management Power Bus relay is energized by white wire EM81-03 from ECM from the Engine Management Fuse Box EM20-6. At least these wires and any harness connectors can be checked for continuity, 0 Ohms from EM81-03 to EM20-6. After that, the reason for an unsatisfied ECM can be investigated. Broken Glass signal from BPM? No Security Acknowledge from BPM? No Ready To Start from KTM? Neutral switch signal failure? The planets are not aligned right?

"I imagine the ECM derives information from the crank & cam sensors to determine that the motor is running." Excellent point. These engine management functions have to be working when it cranks & fires. If the relay drops out after key backs down to position II all these engine functions disappear.

If the key is held on even after it fires, will the engine keep running longer than the 2 seconds? I just tried starting my engine and continued holding the key on and the starter was still spinning with the engine running, so the key does command the starter through the BPM.

 

"The similarity to a broken ballast resistor was pointed out in post 47.
Great work"

Great minds think alike?

That similarity struck me the first time I read this thread. Buying a new ballast resistor to replace a rusted out unit nudged me in that direction again. Hope we narrowed things down.

 

It may be my poor laptop speakers, but it doesn't sound like the engine is actually running for 2 seconds, it sounds like it is "trying" to start but doesn't quite make it, or just barely makes it then dies. Is the engine really running for a full two-second count?

Regarding voltage, what matters to the ECM is not the voltage at the battery, but the voltage at the ECM. There are many reasons you could have a healthy battery but low voltage at the ECM, especially if your starter is drawing excessive current and pulling the cranking voltage low.

If you remove the steering column lower clamshell/cowl, you can access the connector for the ignition switch and test the switch at each position. The Electrical Guide shows which contacts should connect in each key position.

 

What we are saying is that there are two supply circuits for the igniton system . one when the key is in the start position and one in the run position. the first works the second does not..
So if we remove the starter relay and connect a switch to bridge its contacts then put key in start position and push the new switch...I assume car starts ..then release the new switch but not the key ...does the engine continue to run... If so you are half way to a solution or work around...In 10 min you will have an answer

 

I assume the engine is running because the tachometer makes it up to 1100 RPM every time before shutting down.
I currently have the cluster out and the steering column cowling off trying to get at the KTM. When I put it all back together, I will be certain to use contact cleaner on every connection and will check the ignition switch contacts as you described once th battery is back in.

 

It does indeed sound like the engine fires for about more like two tenths of a second rather than 2 seconds or like the length of time it takes to say "one thousand one, one thousand two." Nontheless, the security system had been satisfied, fuel pump kicked on, the engine rolled over to read the cam and crank position, and then spark to fire it up. I'm curious if the engine shuts down due to the ECM or is the key released in 0.2 seconds. IF the key is held on after the engine fires and the engine continues to run, we can assume the key is holding it in position III which is keeping the Engine Management System operating. If the engine just shuts down at .2 seconds as usual, then the ECM, BPM, SLCM, instrument cluster... one of them has changed its mind for some reason.

I have '97 '99 and '06 Xk100s and that means three different security systems and two completely different key chips. Details for are scant for all of it They don't tell us when the security system checks the key, whether is is on insertion, turn to accessories, turn to run position or turn to start. The best I can interpret is maybe at accessories or run, it rattles off the security sequence so the starter will be locked out when attempting to start if it has a security issue. It'd be nice if there was a display message identifying a bogus key. We know when the key is turned to run, the throttle body rattles within a few moments and that system is identified in the Engine Management System. Does that clicking occur or not? . It'd be a clue whether the relay pulls in on run position II. I also am not sure when the fuel pump gets primed by the ECM, but this circuit is not identified in the EMS nor that relay.. All we know is there fuel pressure arriving on the scene and probably at the priming as normal and maybe in start position III. Anyway, I'd expect that the Security system has been satisfied at position II and permits the starter engagement, but I have to wonder if the ECM really checks on the EMS in position II. I believe it goes down the list of these sensors and actuators, but you know what they say about the word *** U ME.

Two minor points, I doubt seriously there's anything wrong with the EMS relay since it seems to work perfectly well in position III. Also, the color coding I mentioned may not be valid if this is the 2000 with the 1999 wiring harness.

You say you'll check out the ignition switch contacts when the battery is back in, but the ignition switch only connects circuits to ground, so you can check many different circuit continuities straight to ground through the ignition switch without much disassembly, but I'd leave the battery disconnected. We want to keep Lucas smoke within the wires, as old timers would say.

 

See the Workshop Manual 3rd edition, section 419-01B. Gus has it on his site of everything here:
http://www.jagrepair.com/images/Auto...p%20Manual.pdf

 

I thought I’d seen 2000 XK8 for this thread, which would involve the much simpler immobilizer system The best description I found for this is in the Service Coursse 881. One whole paragraph. They devoted one sentence to the KTM. The Megamos 13 stores an alphanumeric string as I recall. Y’know, key chips made car thieves turn from hotwired thefts to car jackings with assault & battery and murder. That’s really helped reduce crime.

I pulled my hair out trying to diagnose the immobilizer business on my ‘06. They took a complicated security feature and made it convoluted. The Megamos 48 chip requires AI, Artificial Intelligence to interpret the changing ID. The 2003 Updated Training Manual and Workshop Manual XK8 2003-2006 shed more light on this system plus offer trouble shooting with DTCs, but this particular problem falls into the cracks. There are no tattletle DTC codes showing up and Troubleshooting section doesn’t recognize such a problem. It does provide a very quick test that can be performed...

The best method to confirm the correct operation of PATS is to check the LED (located in the center of the instrument panel). The LEDshould illuminate solid for 3 seconds when the key is turned to the run position and then extinguish. This validates the PATS functions (thekey transponder matches the key code stored, the challenge/response sequence between the instrument cluster (IC) and the engine controlmodule (ECM) was successful resulting in the ECM being enabled). This ought to definitively tell us if the ECM and Engine Management System are normal in run position before hitting crank position. I’m still curious if the throttle body goes through its motions clicking away when the key turns to run, also.

I have more questions than the manuals have answers, like does the ECM prime the fuel rail before it goes through it immobilizer sequence? At the end of that paragraph in the Service Coursse 881, it peeved me to no end to read “ The ECM initiates engine start EMS values for the duration of the cranking signal.” OK, THEN WHAT? I know it should stay running with the Engine Management System and ECM in key position II, but I wish they spelled it out.

 

Correction; I was working under the premise that this was the 2000 model year XK8 with the 1999 wiring housing. This subject is a 2004, not 2000.

The earlier starter is activated by the BPM and the ECM "engine start" is signaled by that starter relay wire.

The latter system gives the ECM total control of the starter, so it decides whether the Security guards, sensors and actuators are happy and hits the starter after the BPM transmits the "engine start request.".

It'd still be good to know if the Instrument Cluster LED lights up for 3 seconds.and throttle body clicks off its Morse Code in the RUN key position II.

 

I got the KTM, Key Transponder Module Out of my 2004 XK8, not that bad a job, took about 2 hours. Next time I could do it in under an hour.
The shop in Toronto was going after that because it shows on the SSD module report with a question mark so that is what lead them there and they got a replacement used one, it did not change the problem with the start and quick shutdown.
I called them to find out if I they had put the original (OEM) back in or left the used replacement. They told me they had left the used one in and did not return the cars original one to me.
On looking up on the web for this module I see some reference to new ones that are unprogrammed.
My question, does this module have to be programmed to match the keys if you put a new one in.
The behavior of the car did not change with the module change, but I am wondering if it has added another something that could be wrong.

 

In my experience, it is very common for modules to display in SDD with a question mark (?) even when there are no serious issues with the module. Unless there are DTCs stored for the module, I've learned to ignore the question marks.

My understanding is that if the KTM was not functioning and authenticating your key transponder, the starter would not be engaged at all. So I doubt that the KTM is your problem.

The shop should retain the original KTM and return it to you.

Have you done any tests to see if either the ignition or injectors are being disabled after 2 seconds of runtime?

Cheers,

Don

 

SDD on an x103 never worked properly, use IDS. KTM is just a couple pids in datalogger only and it's also programmable as a new module.

If you want to see if the ECM has a starter interrupt just unplug the KTM and see what happens

 

YES!!!
There is what is called an exciter ring in the column. It radio signals the key (or something) and the key (chip) responds or provides an identifier as the right key. That information is sent to the module and from there the module sends the "OK TO START" to the BPM. From there, the BPM signals everything else in the prestart checks. I am almost completely certain that IF the key, the module or BOTH are replaced the units need to be programmed...

I believe any hi tech key replacement company should be able to do this programming with a "mobile" service.

Here's the thing. In my experience you wouldn't GET a start up and a shut down. You wouldn't get anything. So I am confused (which is not unusual). So this is strange unless my entire premise is incorrect...?

Q: Have you checked anything to do with the inertia switch? Tested it at all?