Expert: Ernest (Ernie) Kenward - 11/17/2007

Question
QUESTION: Ernie,
I am a not so proud owner of a 94 Firebird, well lately anyway.  Options of importance are 170,000 miles, manual tranny, 3.4 V6, California car.  I am a certified aircraft mechanic, and just cant figure out a misfire my car has above 3000 rpm.  Load or no load, not temperature related, and it is a solid failure, otherwords not intermittent.  I have checked the spark plugs, wires, ignition coils, fuel pump, fuel pressure regulator, replaced fuel filter, checked MAP sensor (MAF sensor not installed), checked O2 sensors, compression check, air filter replaced, Checked EGR (flow and electrically), checked for vacuum leaks, and finally I checked the EVAP and AIR systems.  There is no service engine soon light.  The engine will accelerate the car up to 5 grand, but at 3 it breaks up and starts to fall on its face.  A coworker suggested valve float, but at only 3000 rpm on up it sounds unlikely to me.  Another suggestion was partially clogged injectors, but the car idles fine.  Any other ideas?  Thanks for looking!
JJ   


ANSWER: The best I can do for you is pass on the "eureka" moment that I was lucky enough to have after some weeks of pursuing a problem of sudden, dramatic and complete cessation of power in a car that otherwise performed beautifully.  Sometimes we spend a great deal of time checking out and rechecking out the presumed obvious candidates for problems, and miss something that turns out to be very low-tech ...

The Ford FE "big block" in my 1959 Edsel is an engine I take pride in tuning to the nines in terms of drivability, economy, and like-new emissions performance.  

The engine pulled 19+ inches of vacuum at idle, ran about 2% CO at idle and less than 0.5% loaded, had a low HC reading which corroborated the good vacuum reading (no leaks) and had a highway range that indicated up to 18 miles per Imperial gallon on highway cruises.

Due to its rear-end gearing (2.65:1), it had a respectable but not stellar pull away from a stop, but had a good passing-gear kickdown to second that could be achieved up to 75 mph; this is documented in the glove-box manual for this car.

All in all, a very nice ride.  However, there was this one problem that surfaced.  One day, during high acceleration and once over 60 mph, the engine abruptly lost power and shut off!  "What the ... ?" was all I could think as I coasted to a stop.  I attempted to restart, and ... no problem.  I took off gently, and no problem.

It seemed to be a solitary event, an anomaly.  That the problem was electrical seemed clear, but checks with a timing light showed steady timing and advance, and voltage and megger checks of the nearly-new coil indicated no problems or breakdowns there.  The cold insulation resistance was high enough that, even compensating for a running coil temperature of 75 degrees C, the insulation resistance was well above the minimum required at that temperature.  (I did not have my Fluke 1520 then, and this makes this sort of thing now very easy to assess!)  The ignition harness was a premium copper wire set, no arcing or grounding issues.  More to the point, given the complete-failure nature of the problem, the main ignition lead from the coil to the cap gave no clue to a problem.

A day passed, then two, then a week, and driving returned to business as usual.  And then, one day, it happened again.

And now (finally) to cut to the chase ...

This time, while I had the distributor cap off, I began idly playing with the points and then the breaker plate.  I rotated the plate, more to see that it moved smoothly than for any other reason.  It did.  The timing light had already shown that the centrifugal advance worked to spec (the distributor ported advance being disconnected) and the addition of the carburetor signal similarly worked as hoped.

And then, I saw it!  As I turned the plate all the way to its maximum travel, the wire lead that passed through the case sidewall to connect to the points was seen to lift ???!

Could this mean ... ?  I lifted the wire myself, and twisted it so as to see its bottom surface ...

The insulation was neatly and completely shaved away, apparently by many instances of occasional abrasion, abrasion that occurred only when the breaker plate turned to the maximum of its travel.

"Eureka!"  The grounded breaker plate was shunting the points, solidly shorting the coil primary to ground, and doing so ONLY when the engine was at maximum loading!

The fix?  I pulled in some excess lead, cut off the shaved section, and reterminated the connection to the points.  

Future prevention?  I modified a grommet and used that to replace the original.  My modified grommet held the wire a little higher than it had been, keeping it clear at all times from the breaker plate.

James, all you can do is what you most likely do in your work anyway: Keep your eyes open and be step-by-step methodical.  Touch everything.  Think laterally.  And on it goes.

Thanks for the complement in asking me my opinion on this.  You will find your problem in short order, and I would be eager to hear about it.

Best regards ... Ernie Kenward


---------- FOLLOW-UP ----------

QUESTION: Ernie,
Thanks for the quick reply.  I have been in this type of situation in the past, but man this one is kickin my butt.  Unfortunately as our schooling doesn't teach us much about electronic fuel injection.  I have found some more info to share with you.
This morning I read through my Haynes manual on o2 sensors, and learned that this system goes into closed loop at 59 degreess of coolant temp.  It started out a 55 degree morning, so I decided to start the car and drive off (no warm up period).  MAN this car really goes when its cold.  In 1st gear at wot is shot right up to 5 grand without a hickup.  After about 20 seconds of driving time, it went right back to the usual.  HHHMMMMM....
On my way home from work, I tried some extreme ideas.  I unplugged both o2 sensors, egr valve, air intake temp sensor and coolant temp sensor.  To my amaze it changed nothing..  
Also need to add, I have replaced the coolant temp sensor.
I'm sure you understand that I'm doing everything to keep this away from a dealer.  Seriously thinking about switching over to a carbureted V8!
Thanks again for your time.  This is a great program.
JJ
   


ANSWER: Don't you just hate that situation where you find a plausible cause-and-effect, only to have the model suggested by that clue break apart when you test it further, taking you right back to square one!  I can certainly relate to that great feeling of "gotcha" you must have had on that initial test run on that cool morning.



I really am going to have to get into electronic fuel injection and digital engine management myself at some point; my area IS in large part electronics, but is so in terms of power supply, inverter and convertor design and in industrial controls.  I do a lot of specialty circuit boards, some of them analog, some of them digital, some even using good old relay logic where voltages and the need for signal isolation is an issue.  The higher-end logic I usually do with programmable controllers.  When and if I finally get an ECU-equipped car (or perhaps modify my E250 to use a Holley ProJection package or something similar), I will be motivated to dig into these in much more detail.  For the time being, I enjoy the legacy automotive systems because much of how they work can be described by basic physical principles, and it kind of satisfies the side of me that almost chose mechanical engineering over electrical early on.  (Physics - and math - bridges both!)

Anyway, it is all too often that problems of the type that you are having is much less due to some flaw in the higher logic of the design, and much more due to some basic issue with some component that has gone "flaky" due to heat, vibration, or even just a manufacturing flaw.  The same could be said for the circuit board that component is mounted on, or for the connector tying that board to the harness.  

There could be an overstressed driver transistor, perhaps not adequately mounted to its heat sink.

There could be a poor bond to a circuit board - cold or cracked solder joint - which, on heating due to its own high resistance opens and which then perhaps just as quickly recloses as the trace cools - that would provide a rather rough mode of regulation!

It is even possible for an oxidised connection, in a connector say, to behave like a diode junction, in this case raising havoc with signals due to its non-linearity, producing harmonics or mix products from them.

The above could also apply to some strands of wire in a crimped-connector, oxidised by the minute bit of water that gets drawn in by capillary action.  (This is why I always solder, as well as crimp, such connectors used for automotive use, and then finish the job by enclosing them with the epoxy-lined type of heat-shrinkable tubing.)

Problems of the type above develop so easily in any system that moves and vibrates and is subject to wide ranges of temperature.  And, of course, these problems can be absolute devils to find!  Freeze mist can help, and so can applying heat selectively.  Of course, doing this in the driveway is one thing; doing it on the road with a real load on the engine is something else!  Do you have access to a dyno?

This has all the hallmarks of being a problem that will cost something in time, but probably very little in money.  Provided the time is available, it is good exercise.
Wish I had an afternoon and was there.

Yes, this is a good program, and fun!  I fell into this myself only this past May and, time permitting, do what I can.  Dealing with people who are striving to understand, rather than just browsing for a quick answer, is what engages.  There is always room for more to get involved, by the way.

Keep me posted as to how this resolves.

Best regards ... Ernie Kenward


Hello again!  I really don't have anything further to offer at this point - I have had to set my own vehicle projects aside for the season, as term end approaches for my electrical equipment and systems course to no less than ninety mechanical engineering technology students.  Nonetheless, I am wondering if you ultimately resolved this problem and, if so, how?

Too, we should see YOU on this forum!

All the best ... Ernie Kenward

---------- FOLLOW-UP ----------

QUESTION: Hi Ernie!  What a suprise to hear from you, I give you a 10 for customer service!  
I also have put my projects to the side for a while, just took custody of my 3 yr old nephew, on top of my 3 kids.  Yes, we are busy.  Good timing, the Firebird is a summer car anyway.  
I have been brainstorming and searching for a fix with this car, I'm sure it's just something simple that I have missed.  I've seen many people troubleshoot by throwing parts at a system, which can be an expensive way to go.  Pretty soon that little light bulb will light in my head and it will run like new.  A dyno run may be the answer, but I'm trying everything else possible first.  The next task I have planned is to remove the fuel injectors and check their flow.  Wish me luck.  
Recently, my 96 Olds LSS check engine light turned on, storing a code of PO171 (system bank 1 lean).  After hours of troubleshooting, the fix was a clean mass air flow sensor.  These systems are incredibly sensitive, sometimes we just have to find the sweet spot.  
Thanks again for the reply.  You will recieve a message when I find the problem!
JJ  

Answer
Incredible sensitivity is one attribute that gives these contemporary combustion control systems their finesse - as I said before, I am going to have to get one of these late model (ie: from the last twenty years!) vehicles - or else upgrade my E250 motorhome - so that I will have the incentive to dig more into the details of these systems.  I do a lot in my work life with electrical and electronic systems anyway however, and I guess my older vehicles provide the switch-in-focus to satisfy my desire to work on mechanical stuff!

Checking the injectors for flow probably is a good idea, especially if more than one type or size of injector could be fitted to the engine; it might be conceivable that the engine could have received units that were not quite right for the application at some point.  (I am just hypothesising here.)

Thinking about restricted flow gives me another thought:  is there a pressure or flow regulator anywhere in the system?  A flow regulator set too low would pass fuel with little restriction up until its cutoff point, and then prevent further increase in flow beyond that and, if the limit is reached prior to the engine reaching its maximum intended power, the result will be diminishing returns and a bog-down.

I use these with my carbureted vehicles as one more hedge against float needle valve problems, and they have served me well in avoiding those.  Whether there is a need or a good reason to use them with pressurised injection systems, I am not sure but expect not.  Still, it is just a "what-if?" thought.

Keep in touch, and all the best ... Ernie Kenward

Thanks for the kudos, by the way!