Mr.T

Disabling sounds like a decent idea, but suggest testing it immediately afterward in case the problem somehow goes away.

The voltage staying higher on the cigar lighter-plug meter (compared to a meter at the battery) during cranking is almost certainly from the meter in the lighter-plug being slow to respond. The best test is with a meter that can trap and hold/display the lowest voltage (during cranking in this case). That feature was expensive many years ago, not so now -- Your existing meter might even have it if it's mid-range quality or better. In any case, I'm doubtful that low voltage during cranking is the root cause because cold-starts (which normally drop voltage the most) are fine, plus all the other work/testing you've done.

Best of luck!

KClight01

"The DTC's cover a wide range of faults, every input sensor gets checked, often a variety of ways including wiring shorts to ground and power. Each module (computer) has internal self-checking, and communications are checked too. I did a quick search for low voltage codes and found them for the ABS and Dynamics sensor. There's a PID for voltage that could be graphed too. EDIT: There's also P0627 for a fault with the Fuel Pump Control Circuit, but it looks like this one lights up the check engine lamp."

I just reread this again, how did you do a quick search for low voltage codes and checks for shorts to ground and power? I did not see that immediately. Maybe I need to spend more time with the app. I cycled through most of the options, and did the advanced scan and then jumped right to the wcm and tip start options.

Mr.T

By running the advanced scan (in JScan), it will automatically check for every DTC (Diagnostic Trouble Code) in every module (PCM, TIPM, WCM, etc.). This includes DTC's that could be caused by low voltage, shorts to ground, open wiring, etc.

Virtually every device/sensor has self checks that will set a DTC upon failure. If any wire for a sensor gets grounded, shorted to +12, or if the sensor fails to produce a proper signal, chances are very good that a DTC will be set.

On a side note, I used to do programming for process control, and always did virtually the same thing -- Check for failures, then generate an alarm and reject the relevant control from automatic to some backup or manual mode if necessary. The problem you are having is unusual and difficult; Apparently it either wasn't anticipated, or was complicated enough that detection wasn't cost effective -- probably the latter.

Mr.T

A couple thoughts since you have figured out the time and temperature conditions to force the problem. While cranking in the temperature-time window for failure:

• Check for spark with the tester (previously mentioned).
• Use JScan to check injector pulse width (any injector) and throttle position. I normally use HPtuners for this type of stuff because of the detailed trend charts, but it should be doable with JScan as well. This will verify that the PCM is actually trying to start the engine with fuel and air.
• Check that the fuel pump is energized.
• Check that the ASD relay is energized.

Obviously, the latter two are not so easy. A few ideas: I haven't checked JScan, but there may be a way to look at the ASD and fuel pump outputs -- However they might show energized even if they are actually not, depending on the failure. There is a very thin insulated wire called wire-wrap wire, and it may be able to slip in with the blade of a fuse or relay to check voltage to ground. Another trick is to pierce a wire's insulation with a straight pin and attach a volt meter.

Hope this helps!

KClight01

Yes, all suggestions help. I have the spark tester and I checked that during the normal and expected long crank conditions. I saw good spark each time. I was only able to test one lead at a time, each seemed to give good spark. I really haven't wanted to mess with the plugs or wires because it runs so smoothly, no misfires or anything.

I used the advanced scan with JScan, still no codes. I'll have to dig deeper but I didn't see option to check injector pulse, that would be interesting. Throttle position is normal.

I hear the fuel pump (its new) and I already checked the ASD and starter relays. I verified they were good but went ahead and swapped the ASD with a new one I bought. I also did the TPIM fuel pump bypass, but that made no difference other than make the fuel pump run constantly when you turn the key to on.

I did check all the grounds including the ASD ground, and the TPIM positive connection. Also rechecked the battery and connections, I have a battery tester that tests the battery, alternator charging, crank, etc. All checks were good.

I looked into ways to run a wire at the ASD so I see what it was doing during the long crank, because it seems to make sense that the ASD is preventing it from getting power or signal.

I disabled the TIP start which was interesting for a trial. It was strange to start the engine without the TIP function, but it did start.

Mr.T

To view values like the injector pulse width and throttle blade position in JScan:

• In this case, goto the PCM module (each module has it's own list).
• Select live data.
• Select the PID's to display or graph (select ... for a list). In this case it's Injector 1 pulse width and throttle position. Make sure it's the throttle blade and not the pedal.

What I was thinking is that if the injector pulse width goes to zero or the throttle closes, then maybe a sensor input necessary for that function is failing. Might be the cam position, the throttle, etc. We know if there's spark and fuel it will start, so the idea is to find out if the PCM cut the fuel and go from there.

Here's the generic documentation for JScan: https://jscan-docs.readthedocs.io/en...d.html#modules

KClight01

OK this app is powerful, lots of info. I have a few questions...
what is cranking pulse width? reading was 8352. Does that relate to crank time? Do you know if that a normal value?
Cylinder 1 injector pulse width reading was ~3100, I tried to watch during cranking, and it did seem to go to 0 for a split second but I'm not sure if that was just my cycling the key or not.
Cylinder 2 thru 6 injector pulse width readings were ~3100 to 3200 with Key OFF cranking pulse width goes to zero, all Cylinder injectors go to 7736 reading
I looked for misfires and under MF CAT ACCUM CYL1 the reading was 24919, does this mean cylinder 1 is misfiring? Could that be the counts during the long cranks? Surely that can't be. Still no misfire codes, stumbles or anything after redoing the advanced scan.
It was starting good this week I heard, but today (rainy) when I drove it, it did a really long crank a couple of times, so I went in and disabled the tip start again. It did not let me reduce the tip start time, saying "failed" each time I tried to reduce the 10 secs duration, so I am just going to keep it disabled for a while to help save the starter.
I think my plan now is to look closer at the cylinder 1 sparkplug and wires to the ignition coil, and then maybe the main plug and wires once I figure out which wire goes to each.

Mr.T

I think cranking pulse width is the injector pulse the pcm wants during cranking. It likely varies with the MAP (manifold absolute pressure), ECT (engine coolant temp), IAT (intake air temp), etc. Everything's a darn acronym... I'd have to guess on the scaling of the number, but bigger is almost certainly more pulse width, more fuel.

I'd be guessing more on the rest, but it sounds like you're on the right track. Check if that missfire count increments, maybe it detects misfires during tip start cranking??? A little common sense and intuition goes a long way! I'll try checking those parameters on mine tomorrow, but note that mine has a manual trans without tip start.

KClight01

Raining so I can't check much today. I did get values at cold engine and again after warmed. Jscan gives the pulse width values in microseconds for some reason. I deactivated tipstart and even the WCM, so I might be closer to your setup now even though you have manual. I watched for misfires under livedata and it showed cyl2 misfire - 1 count, but then zero after trying again a couple of times (shutting off engine and restarting), and no misfires for any of the others. the MF CAT accum stayed at 24919, I guess i'd have to reactivate tipstart and get the long crank to see if that changes with the long crank.
crank pulse width - ~29000 (cold)
cyl inj pulse width - ~4500
crank pulse width - ~8300 (warmed)
cyl inj pulse width - ~3000
MF CAT ACCUM CYL1 24919

Mr.T

Here's some quick data for comparison:
crank pulse width: ~40984 (cold ~30F)
cyl inj pulse width: ~4000 (decreases rapidly after start, which occurred after about 1.5 seconds of cranking)
MF CAT ACCUM CYL1-6: all Zero

The first thought is those miss-fire counts on yours are not good. Mine were all zero and stayed that way. Early on in this thread Eddie (EHarris) mentioned having the same symptoms and changing the coil fixed it, and typically ignition problems cause miss-fires.

There's a ton of parameters (pids) that might help, and graphing them, which is call "Diagrams" is sometimes the better method if just looking for a change. At the top of live data, it says LIST, BOXES, DIAGRAMS to choose the display method. Pids that have the word "crank" in them (use the search feature) may be interesting, for example "crank pulses detected". I had that one in the list this morning and it changes state immediately after turning the key to the crank position. However, the obvious stuff like miss-fires detected sounds like a promising lead, and needs to get fixed eventually anyway.