Jump to content

Recommended Posts

T-shooting an intermittent starting problem.  About 80% of start-ups are perfect.  When failing to start, it will turn over and over, but won't fire up.  Sometimes starter fluid will cause it to fire up and then it runs fine, but that's only 50% effective. The rest of the time, only push-starting will work, but push-starting does work every time.  Most obvious assumption is fuel problem, but fuel system checks out fine (replaced almost everything over time anyway). Ignition/electrical is fine.  Also, the problem seems to be completely independent of outside and engine temperature.

I started reading about ECU inputs, and discovered that during startup, an engaged WOT switch causes ECU to cut off fuel because ECU thinks it will flood the engine.  The WOT switch is worn, but shouldn't prevent starting if disconnected.  Would any other ECU inputs prevent fuel flow during start-up if their signals are outside normal limits?  Service manual says during "Ignition On", CTS, MAT and MAP sensors begin providing input to ECU, but doesn't list MAT or MAP sensors during actual start-up. See attached page.  So which sensors matter to ECU for start-up fuel supply calculations or decisions?

Also, during push-starts, the ignition (key switch) is in the Run position, and not the Start position, which changes the ECU inputs, too.

My shop has a DRB II diagnostics reader, but unless they hook it up when it happens to be misbehaving, shouldn't everything check out ok?  Unfortunately, I've never been able to get it to fail at the shop.  

I also tried replacing the fuel relay.  I also noticed that it is no longer cycling the ISA motor when it fails to start. It used to, but I can't figure out why it did that anyway.

What are my next moves?

image.png

Might be trying to oversimplify this, but what kind of intermittent problem would occasionally prevent a normal startup, but NEVER prevent a push-start from succeeding?

Link to comment
Share on other sites

  • Administrators

Msongy...You've been thorough in presenting this issue and your progress...I'll make some comments based upon the troubleshooting you've done so far, we can go from there:

19 hours ago, Msongy said:

T-shooting an intermittent starting problem.  About 80% of start-ups are perfect.  When failing to start, it will turn over and over, but won't fire up.  Sometimes starter fluid will cause it to fire up and then it runs fine, but that's only 50% effective.

This hints that the problem is in some way related to fuel supply...When the engine is cranking and not starting, have you checked the spark and fuel pressure at the TBI unit?

The rest of the time, only push-starting will work, but push-starting does work every time.  Most obvious assumption is fuel problem, but fuel system checks out fine (replaced almost everything over time anyway). Ignition/electrical is fine.  Also, the problem seems to be completely independent of outside and engine temperature.

This suggests that either the ignition switch or the CPS (crankshaft position sensor, a common issue) might be defective.  If the ignition switch is involved, note that you get a normal run condition in the KEY-ON position but not the cranking position.  Use this as a lead to your troubleshooting.  There is either a voltage drop in cranking mode or a trigger for reducing fuel supply or spark.  Voltage drop can be on the hot "+" side or the ground "-" circuits.

Again, I would check spark during the no-start condition as you crank the engine.  Try using a remote starter switch at the starter solenoid.  Turn the key to the on position, and see if you get a consistent start condition.  IF SO, you have an issue in the cranking mode position of the switch.  If the CPS is at fault (wrong resistance reading or intermittent signal due to poor connections) or simply dirty/oily, you might get the fire condition when the engine spins faster during a push-start but not during the slower crank start.  Check out both possibilities here.  Keep in mind that voltage naturally drops during engine cranking, a poor battery connection or defective battery (dead cell, etc.) can increase the voltage drop during cranking.  This could impact the ECU performance...Check the battery voltage and supplied voltage during engine cranking. 

I started reading about ECU inputs, and discovered that during startup, an engaged WOT switch causes ECU to cut off fuel because ECU thinks it will flood the engine.  The WOT switch is worn, but shouldn't prevent starting if disconnected.  

Good research.  IF the engine still does not start with the WOT switch out of the loop, you're right, that's not the problem.  A separate issue, you do need a WOT if defective.

Would any other ECU inputs prevent fuel flow during start-up if their signals are outside normal limits?  Service manual says during "Ignition On", CTS, MAT and MAP sensors begin providing input to ECU, but doesn't list MAT or MAP sensors during actual start-up. See attached page.  So which sensors matter to ECU for start-up fuel supply calculations or decisions?

Crank mode still uses information from the CTS, MAT and MAP sensors.  The engine simply needs to fire, though, at which point the increased rpm takes over, and ISA comes into play.  At this stage of your troubleshooting, I would do three things:  1)  verify spark during cranking mode, 2) verify fuel pressure/supply and 3) confirm available battery voltage to the ECU during cranking.  (The concern is sufficient voltage for the ECU to work properly.  It needs adequate voltage!)   Additionally important:  With TBI, the charge of fuel depends on the constant fuel pressure at the pressure regulator to provide an injector fuel pulse.  A "Tee" at the pressure regulator fitting with a test gauge could indicate what happens to fuel pressure during cranking.  Consider both fuel pressure and actual fuel volume.  Volume can be impeded by a clogged fuel filter or tank sock, or poor fuel pump performance (defective pump, bad pump grounds, poor connections at the pump, etc.).

Note:  A common source of trouble with these TBI era models is a poor engine ground(s).  The ground near the dipstick often comes up as a trouble spot.  Check the ground continuity/resistance!

Also, during push-starts, the ignition (key switch) is in the Run position, and not the Start position, which changes the ECU inputs, too.

My shop has a DRB II diagnostics reader, but unless they hook it up when it happens to be misbehaving, shouldn't everything check out ok?  Unfortunately, I've never been able to get it to fail at the shop.

There should be stored codes if there is a general issue.  The DRB II is a great tool for picking up trouble codes but even more importantly, for cycling and testing individual devices!  Use this tool to its fullest capacity, and test individual sensors through the DRB II tool if possible.  This can be done with OBD-II, see what degree of function is available for OBD-I testing on a system like yours. 

I also tried replacing the fuel relay.  I also noticed that it is no longer cycling the ISA motor when it fails to start. It used to, but I can't figure out why it did that anyway.

Possible clue.  Don't rule out the obvious issue of a defective ISA motor or an ECU failing.  ECUs do fail on these models, and there is a thriving "rebuild" program in the aftermarket.  There are also several firms that specialize in rebuilding and repairing ECU units.  Considering the vehicle's age, this is a prospect.  Check out other possibilities before condemning the ECU.

What are my next moves?

Might be trying to oversimplify this, but what kind of intermittent problem would occasionally prevent a normal startup, but NEVER prevent a push-start from succeeding?

See each of my comments above...Moses

Link to comment
Share on other sites

Thanks!  I don't know whether to be relieved or concerned that you didn't mention any vacuum lines.  Is that because checking for broken, dirty lines is obvious or because the logic doesn't support it?

Also, does the occasional success with starter fluid (key start) indicate anything in particular?  I must also add that the frequency of success with starter fluid has been slowly declining over time, indicating either increased corrosion, pitting at contacts or buildup of material in lines, filters or orifices. Or all of the above! U agree?

Link to comment
Share on other sites

Msongy...Vacuum circuits are always a concern.  However, you're sharing that the engine runs great once started, and it starts every time you push-start the vehicle.  Of course, you do want the vacuum circuits to be routed correctly without vacuum leaks, as this can affect overall performance.  Leaks also include the brake vacuum booster or even the brake booster's check valve.  Worth a glance...

If you have any suspicions about the vacuum circuits, you have the FSM for diagrams.  To save time, there is an earlier post where I added a PDF of the vacuum circuits for a 1987-90 Jeep YJ Wrangler TBI 2.5L:

Follow my highlighted comments above, and you will systematically eliminate possibilities.  Among those possibilities is the success or failure of starter fluid.  There's a fundamental mechanical problem, and you're going to find it!

Moses

Link to comment
Share on other sites

Super! Thank you. I hurried outside to resume troubleshooting, and found that my wife drove it to work!  Aside from the fact that I can't work on it, she also has no hill to park near in case she needs a rolling start!  But I won't ask your advice regarding my wife...:D

Link to comment
Share on other sites

New info. Just failed to start, but for the first time, I was able to start it with very slight pressure on the gas pedal while cranking.  I thought that only worked with 6 cyl., and it's actually 50 to 75% depressed for a hot engine start.  I was only pressing about 1/8"!  

Sounds like the extra air/fuel helped, but throttle might actually not be at best starting position.  ISA is definitely resetting after shutdown, but whether or not it's setting the throttle to the best location is the question. Turns out I was right. Read on.

If I accidentally turn the key to start only for a fraction of a second (just long enough to turn over once), the idle actuator used to "pump" the accelerator pedal several times when the key was returned back to the Run position, as if to prepare for the next startup attempt.  It rarely does that any more.  Is that the "ratcheting" that occurs if the ISA motor runs when the actuator is at full extension? (Yes it is!) I am guessing that the ECU interprets a failed start as a shut-down and attempts to re-position the ISA - or is supposed to.  Is that correct? (Yes it is!)

I'm so close I can feel it!

And more: just made some progress. When startup fails, ECU sends power to ISA again, opening throttle a little more. If startup fails again, ECU does it again, until actuator is at full extension, and sure enough, the ISA motor ratchets, shaking the accelerator pedal.

Now, rewind back to the the start of this latest post.  Stepping on the gas just a fraction simulated the ISA opening the throttle a bit more.  On shutdown, does the ECU aim for a pre-set physical throttle position (and if so, does TPS or CTS provide feedback) or does the ECU simply energize the ISA through the B+ relay for a preset amount of time, regardless of throttle position?

I just read that last question to my wife, and she just stared at me, then asked, "So, is it fixed?"

 

 

Link to comment
Share on other sites

  • Administrators

Msongy...For the benefit of those without a factory service manual, see the PDF below.  My ISA motor focus would begin with identifying the described inputs to and resulting outputs from the ECU.  The third page has a description of how and why the ISA motor functions.  It's clear what should be happening.  The "ECU Input" signals determine the response.  

The steadily extending ISA plunger sounds like a lacking engine start, which does not allow the ISA motor to actuate normally and reset as it should when the engine runs and is then shut off.  Instead, as the ignition key gets cycled, the plunger just keeps resetting further out due to lack of a start:

Jeep 2.5L TBI ECU Inputs and Outputs.pdf

Once the engine fires, the ISA motor's plunger position responds immediately to ECU processing outputs.  Failure to start is why the ISA motor continues to move the plunger out.  This will create a fast idle upon startup and then an immediate response to the engine sensor/ECU inputs.  This allows the idle speed to stabilize.

TBI is an electronically responsive, carburetor-like device with the ability to control air/fuel (A/F) ratios far better than a mechanical carburetor and a human foot on the gas pedal.  This consistent control of A/F is TBI's contribution to emission control and improved performance.  (Check the canister purge system on your Jeep, make sure it is functioning properly.  EVAP troubles or vacuum leaks can affect start-up, too!)  TBI shares many mechanical features and behaviors with carburetors, including a throttle position valve and precisely matching fuel flow without the need for a carburetor's venturi.  Check the fuel pressure at the regulator and see if there is an issue.  You need spark and fuel to start, and once the engine has started, the ISA motor should function correctly.

Moses  

Jeep 2.5L TBI ECU Inputs and Outputs.pdf

Link to comment
Share on other sites

Are you stating that the fuel/air ratio and quantity at startup are pre-programmed and not dependent upon initial throttle position?  Opening the throttle means more air, so if the ECU holds the fuel quantity constant at startup, regardless of throttle position, then a slightly more open throttle creates a leaner mixture, more easily combustible. Sounds like the ISA continues to move the plunger out in order to increase the engine's ability to fire up.  It shouldn't have to work that hard, though. I'm keeping at it. Going to continue cleaning and testing!  Thanks!

Link to comment
Share on other sites

  • Administrators

Msongy...TPS (throttle position sensor) is a major ECU signal at all times.  The ECU weighs the throttle opening position.  TPS voltage needs to be in sync with the actual angle of the throttle valve.  That's the indication of throttle position.  

The TPS voltage reading represents the range of TB throttle valve angles from the closed/idle position to WOT.  This is why TPS readings and settings (on a movable/adjustable TPS) are so important.  Note that TPS adjustment and diagnostics are always expressed as voltage readings, that's the way an ECU interprets the angle of the throttle valve in the TB unit.  The voltage output from the TPS must reflect the actual throttle valve position or angle.  Throttle valve opening is driver controlled by the throttle cable and gas pedal.

Moses

Link to comment
Share on other sites

  • 8 months later...

This looks like a dead thread but I have been reading it as I have a very similar problem, my 88 yj tbi will not start cold unless I press on the throttle pedal while cranking, it usually cranks for 15-20 seconds then sputters to life, eventually clears up and idles and runs fine for the rest of the day. I have been reading thru these pages for a few days and I have done very extensive troubleshooting, I now have an O2 Sensor on order as I am pretty sure that is linked to my problem. I have done ohm readings and pressure tests and even hooked up a snap-on "brick" scanner. Scanner shows no codes, reads "this vehicle is not capable of storing codes". One thing I noticed on the data menu, engine running, is the O2 stayed at lean and the computer never went to closed loop for the duration of the troubleshooting.ON a separate topic, I used to have flickering gauges and through checking all the grounds at firewall and engine block that seems to have gone away, all gauges are working properly now. Any input is appreciated. Thanks

Link to comment
Share on other sites

  • Administrators

matador75...Could be the O2 sensor with the lean reading and open loop.  These sensors generally stay in the system too long, won't hurt to replace it.

Have you checked fuel pressure during cranking mode?  At the TBI test port?  Is the fuel filter fresh?

Moses

Link to comment
Share on other sites

  • Administrators

You might consider a fuel filter replacement...It only takes one bad tank of fuel (higher water content or debris) to clog a paper matrix fuel filter.  These filters are designed to swell in the presence of water to stop fuel flow and protect the precise EFI system components.

Here's an anecdote...I change my Ram/Cummins fuel filter regularly.  I had a relatively new filter in the engine bay and drove from home to Green River, UT en route to Moab EJS last year.  While I was fueling at a truck/passenger car station, a tanker truck was filling the station's tanks.  I didn't think much about this, finished filling and drove on to Moab.  When I pulled into the trailer park, the engine's idle was rough.  I quickly realized that the fuel filter picked up debris and water from the station's tank.  The tanker delivery had stirred up these contaminants.

That filter barely lasted 55 miles.  I carry a spare and changed it.  The new filter is still in the truck, and I drove right past Green River this year, making it to Moab before filling!

Moses

Link to comment
Share on other sites

Yes indeed, it's easy to replaced and also very easily overlooked item as it's under the vehicle. I probably should be changing it yearly especially with today's fuels but it's just not something I think about until something is wrong. I looked underneath today and it is pretty rusty, won't hurt to replace that's for sure. I wish they were still underhood like the good Ole days. Thanks Moses 

Link to comment
Share on other sites

Wouldn't a filter swollen from water contamination also prevent a push-start?  Recall that even when failing to start normally, a push-start always works and the engine runs normally afterward.

Regardless, filters are cheap, so worth a try. Will also fully drain and clean out tank.

Link to comment
Share on other sites

Update, replaced filter and O2Sensor today, I am now hooked up to scanner as I write this and it just went on closed loop. O2 is switching. Idle is showing about 635rpm on scanner as opposed to 575rpm before, sounds better too, doesn't have the occasional putt putt as it did. As I normally do when I change a fuel filter I blew thru the old one in the direction of flow and there wasn't much resistance at all, I don't believe that was my problem, but it was due to be changed anyway. Fuel pressure is the same before and after about 14psi. I checked cranking pressure and it shows 15psi. Gonna take it for a little ride now and hopefully tomorrow on a cold start everything will be good. Thanks for the help

Link to comment
Share on other sites

  • Administrators

matador75...You're probably right about the fuel filter but better to have a new one in place.  Did the old O2 sensor look contaminated?  Did you replace the sensor with an OE brand sensor?

Sorry for the continued anecdotes, but I replaced the O2 sensor on our Jeep XJ Cherokee 4.0L with a Brand-X sensor.  It functioned okay for a short while, then started acting up.  In 1999, Mopar outsourced O2 sensors from NTK.  I took the NTK number from the original O2 sensor and purchased an NTK replacement sensor.  It's still in there and working flawlessly...Didn't have to be Mopar, but I did need the NTK part.  

Aftermarket or generic O2 sensors often cover a broad range of applications.  When lucky, the benchmark happens to be your application.  To avoid this issue, I use an OEM replacement (Mopar) or at least the Mopar outsourced equivalent...Same with other sensors.  I've had a generic, off-shore replacement TPS switch that could not pull the right voltages and a temp sensor that threw off the gauge readings by five to ten degrees (still living with that part)...That's my experience.

You mentioned push-starting without a glitch.  There could be too much resistance on the ignition switch circuit or the starter motor draw might be excessive.  Grounds are a possibility, again, recall that poor grounds cause a voltage drop.  This could include the fuel pump ground if there are possible fuel pump fluctuations.  That's why I was curious about the fuel pressure and volume/flow while cranking.  

There could be too much electrical load when the starter draws.  Be aware that poor battery connections, resistance in the battery leads, a poor engine to battery ground, or a battery with a dying cell can drop voltage too much during cranking mode.  This could reflect as ECU trouble, TBI and ISA misbehavior, or ignition faulting.  Look for items affected by the starter motor draw.  You may have a strong voltage drop during cranking mode.  Remove the battery and have it load tested.  Check for poor battery connections, worn battery leads or a starter motor and solenoid that draw too many amps during cranking.  The starter motor can be bench tested, too.

Moses

 

Link to comment
Share on other sites

I think you got me mixed up with a previous member. I never had to push start mine, I just have extended crank time, like 15-20 seconds, battery and starter are less than 2 years old as well. I don't have a cranking problem I have a cold start problem. The O2 is an ngk unit, yes the old one looked like crap, probably original judging by the hex part being rusted away undersized, but like I said, it wasn't switching and the new one is. I haven't messed with the jeep today, hoping I would get some idea on what to look at next. I did check the grounds last week and that fixed my fluttering gages.

Link to comment
Share on other sites

  • Administrators

I read Msongy's comments posted Saturday at 12:12 P.M. and thought the reference was to your engine.  Difficult to juggle two balls here...You can see where my comments originated...Sorry for any confusion, we should have moved your initial post into its own topic.

So, it looks like the O2 sensor is from the day.  NGK oxygen sensors are now NTK.  Regardless, yours needed help, a good call to replace it.  Looks like a lot of external corrosion from air or road salts.  Glad you were able to get it out.

Did you test the ISA function?  I have provided ISA testing and adjustment procedures in PDF format.  Go to this topic and see my August 29, 2015 reply with two instructional PDFs:

Moses

 

Link to comment
Share on other sites

Drove the jeep to work this morning, started hard but drove fine. Drove home from work spitting and sputtering about every 30 seconds or so. I just did the wot test at terminals d2-6 and d2-7 and it appears the wiring is shorting to ground. I get 0.67V with switch in both positions With the key on and the switch itself tests fine. Even with switch unplugged I get the same reading at the test terminals. I was thinking of just snipping it at both ends and using a new wire but once it gets to the main harness I cannot follow it as the wires are all taped under the wire loom. I see no easy way out of this one. Do you know where the wire ends up from the wot? Thanks for the help

Link to comment
Share on other sites

  • Administrators

matador75...Here's the OE wiring schematic depicting both sides of the WOT switch on your '88 YJ Wrangler 2.5L TBI series 80 chassis:

YJ Wrangler 2.5L TBI Wiring of WOT to ECU.pdf

Note the ground circuit on this device and the junctions for grounding the WOT switch.  Also see the ECU connections.  This may provide a clue.  Most TBI and EFI systems have KEY ON hot leads to devices and injectors.  The ECU/ECM/PCM provides the ground to complete the 12VDC circuit.

Moses

Link to comment
Share on other sites

  • 2 months later...
  • Administrators

I don't believe so.  You cannot get an O2 reading/signal until the engine fires.  Cranking mode is not looking for a cat/O2 reading.  However, as you may be thinking, entering the WOT or limp modes could be a cat related issue.  In any case, the engine needs to come off the warm-up cycle before a meaningful O2 reading occurs.

A sub-standard or defective cat could create a downstream or second O2 related signal issue.  Your 2.5L TBI engine does not have a second O2 sensor downstream past the cat, right?  The single O2 reading is below the exhaust manifold and before the cat.  The type of O2 sensor (heated, not heated, etc.) and its parameters must match the OE computer/ECU calibrations.  Confirm the O2 sensor part number and application.

Have you checked the EGR system and valve? If the EGR valve is stuck open, shut or leaking off, the cat load and O2 readings could send the system into limp mode.  Can you read the tailpipe emissions?

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...