Strange Starting Issue

[AZJeff]

As my sig line shows, my MJ has an engine/control system from a 2000XJ, so it's OBD-II.  It also has a strange starting issue, and has had it for a long time.

 

When I crank the engine to start it, it will crank just fine, but not catch (ignite), even after about 7 seconds of cranking.  If I release the key, and then return it back to the start position, it then starts right away.  It does this whether hot or cold.  

 

The only time it will start right away (on the first crank) is when the engine was shut off and left to sit for perhaps no more than 5 minutes, in which case, as soon as I turn the key to "start", the engine ignites immediately.   The rest of the time, it's two cycles of the key to the "start" position before the engine will ignite.   Once it DOES start, it runs smooth at appropriate idle speeds immediately, and has normal behavior otherwise.

 

I replaced the CPS a few years ago, since it died, with a new Mopar unit.    The vehicle never throws any OBD codes, so I am pretty sure all else is normal.  (It has a new coil, new plugs, new MAP sensor, and new idle control solenoid, BTW.)

 

Any thoughts?

[jdog]

Sounds like the keep alive memory is getting reset, possibly the constant power wire is either not there or keyed power

[AZJeff]

On 8/6/2021 at 12:35 PM, jdog said:

Sounds like the keep alive memory is getting reset, possibly the constant power wire is either not there or keyed power

I would say your suggestion sounds plausible, except for one thing:   if I shut the Jeep off, go away for some short interval (5-10 mins), and then go out to restart, it catches right away.   If I wait some longer interval, like 30 mins for example, then we are back to two crank cycles to get it to start.

 

Keep alive memory with the issue you described would become lost every time I shut the machine off, I would think...

[silvertwinkiehobo]

Of course, the cranking/starting strategy is for the engine to spin a few revs for the computer to determine TDC, then do fuel and spark. I do not think the CKP sensor is misadjusted, because otherwise the engine won't run. So what popped into my mind is you're losing the fuel prime, or losing power to the injectors and/or the coils when the ignition key cylinder is in the START position. Give each of those a check, see what you come up with.

[AZJeff]

I am certain my fuel system is pressurizing, since I can turn the key to “on” and hear the pump run.  The idea that the coils or injectors are not getting a signal at the correct time is an interesting concept.   Now how to figure out how to measure that in real time…..

[jdog]

Possibly cmp

[AZJeff]

7 hours ago, jdog said:

Possibly cmp

Hmmmm.....good point.  I replaced the "distributor" shaft/housing a while back due to worn bushings, and it came with a new CMP.  The one that came with the unit is of unknown brand, and we all know how Jeep control systems don't like aftermarket parts very much.

 

It looks like the OEM part is discontinued, from my on-line search.  Among the known brands that I see available, there is

1.  Dorman (ugh, probably relabeled junk)

2. Crown (they are hit or miss)

3. NGK (usually good, if actually made by NGK themselves

 

Any recommendations?

[AZJeff]

Bueller?    

[89 MJ]

Maybe @eaglescout526 could help you locate a new Mopar one. 

[eaglescout526]

New Mopar what?

[jdog]

Cam position sensor

I haven't got a clue if any of them are good

[eaglescout526]

So if it’s the sensor I’m thinking that’s in the distributor itself, there’s 6 left. 

[AZJeff]

Ugh.   I have a new "distributor" I got through Rock Auto.  It came with a cam sensor of unknown source.  I would hate to buy a whole Mopar "distributor" just to get the cam sensor that attaches to the top.  (The cam sensor used to be available as a separate part.)

[eaglescout526]

Oh that’s right. You have the HO stuff. I looked at Renix. Let me look at HO but it still should be separate just like Renix. 

[eaglescout526]

Will this PN work? 53009077

[AZJeff]

Back to resurrecting this thread.  So I put in a new NGK cam position sensor, and confirmed it's "timing" with the toothpick method per the FSM.

 

No change.  In fact, over the past month or so, it's actually gotten slightly worse.   Now, if I leave the vehicle for more than 10-15 minutes, I have to crank the engine twice before it's starts.  Once to get the control system to get a sync, and then the second time, it starts.  It sure is acting like a loss of volatile memory is taking place.

 

So, I  am just spitballing ideas here:

 

 

---I checked the 12v "keep alive" power feed to the ECU, and it's working normally (at least up through the connector).  However, I don't know if there might not be an internal issue with the ECU.  Perhaps the feed is not getting into the actual circuit board inside the ECU.  

 

---Or, maybe the capacitors in the ECU are getting old and bleeding down rather quickly with old age, and thus the ECU looses some memory settings due to capacitor drain.   I seem to remember reading about someone who went in and replaced the caps in an ECU on some MJ or XJ, but I cannot remember which forum I was on where he reported doing this.

 

I need some suggestions from our electronics wizards on this.

[AZJeff]

Bump so that someone with ECU/PCM knowledge at a deeper level can comment here.

 

I know this is not very MJ specific, so if someone can point me to another forum that deals more with OBD-2 PCM’s on XJ’s, that would be appreciated.

[zomeizter]

How about a quick PCM swap, I reckon one can be had for cheap at a boneyard? My C4 did this for a long time but I lived with it since it didn't get used that much and I blamed the issue on this low use condition, but I have a feeling I have a fuel priming prob...good luck pinpointing the cause.