When I first picked up wrenches and started trying to get a handle on how cars worked, it was fairly simple to figure out why an engine wouldn't start. We'd first check for spark and the proper firing order, then we'd yank the air cleaner, prop the choke open (if it was working right it'd be closed on a cold engine) and we'd operate the throttle to watch the accelerator pump jets. From there, we'd jerk the spark plugs out and see if they were wet and sooty or used up.
Hard starting diagnoses typically led down a slightly different path, and because we're talking about newer vehicles, we'll fast-forward a bit and mostly try to leave the carburetors and oil-filled coils in the past.
Spark FirstFor years, contemporary wisdom in textbooks and some older shop manuals said to let the spark jump ¼ inch, and if you had that much spark, it was enough. Working with my dad as a youngster, I discovered the biggest problem with that ¼-inch rule was that a spark can jump that far in sea-level air pressure and still not have enough of a pop in the chamber to light off the mix under compression.
The spark needs to be downright scary these days, with enough blue-white (not orange) energy to punch through nearly an inch of air and pop loud enough to be heard three service bays away. I've seen some really good technicians stumble over that ¼-inch spark test and then call for the "calf rope" because they had been side-tracked by weak coils that wouldn't produce sufficient spark.
So how did they get off track? Well, the spark would traverse that ¼ inch and they didn't stretch it to see when it reached its limit, so they erroneously figured they had eliminated spark and in error they moved away from the ignition system with their troubleshooting.
I remember one hard starting 1988 Taurus that would actually fire up after spinning for 30 seconds or so, and it had me stymied until I connected the ignition O-scope and found that the coil was incapable of producing the voltage needed and it would jump only about 1/2 inch.
That was acceptable on older oil-filled coil systems, but on the Taurus I was troubleshooting, it wasn't enough. A new coil took care of that one.So when testing for spark, make sure you stretch that blue lightning far enough to determine that it's not weak. That's rule No. 1 on a single-coil vehicle with a hard start concern.
As far as coil packs go on Distributorless Ignition Systems, single towers or whole coils on multi-coil packs can stop delivering spark and Coil-On-Plug (COP) coils can drop offline for one reason or another. But weak spark generally isn't the cause for a hard-start on coil pack or COP equipped platforms.
Then there was the Bronco that would start only with the SPOUT connector disconnected. I found that problem with the Service Bay Diagnostic System. The PCM ground traces would go crazy while the engine was spinning, and that turned out to be a voltage drop problem at the battery cable-to-engine block connection.
There were the burned-through rotors and faulty ignition modules, but they're more likely to cause sputtering and stumbling than hard starting.
Understanding Fuel Delivery
With spark either verified or eliminated as a possible cause, the next thing we need to know about our no-start is how much battery voltage is still available to the PCM with the starter doing its thing. Most systems won't operate the injectors if the available voltage drops below a certain level.
Fords typically won't start if the system voltage goes below about 7.5 volts, and the engine will seem to be spinning at a pretty good speed, but it won't even try to fire up (this is also true on Power Stroke Diesels).
If the system voltage is sufficient, it's important to know whether or not the injectors are clicking and what their pulse width is in milliseconds (ms). You can do that with a scope or an enhanced scan tool datastream. Remember, however, that the datastream isn't giving you real-world information like a scope — it's simply giving you a window into what the PCM is trying to do and usually updates pretty slowly. You're really getting the information filtered through two computers on the way to your eyeballs.
Let's think of fuel delivery at startup in a simplistic way: Tecumseh lawn mower engines have that funky rubber thumb pump to wet their tiny manifolds. On carbureted engines, before the engine will start there has to be a surplus of fuel in the intake manifold, thus the accelerator pump and the follow-up of the closed choke.Fuel injected engines are no different when it comes to required fuel. During start, particularly cold start, the injectors either fire simultaneously or bank fire so as to provide the fuel needed to light the engine off in lower temps. And moving the throttle while you're spinning the engine can put too much fuel in there (it adds additional pulses to simulate the carburetor's accelerator pump delivery), so watch for in-range TP sensor failures that might cause hard starting or rich running problems.
Furthermore, it is quite important for the Idle Air Control (IAC) to be wide open during initial start for that no touch starting feature we're all used to. If it doesn't provide the needed air (because of sticking or whatever), a bit of throttle opening typically gets the mill fired up. But as an old-timer, I have to remind myself that there is a whole generation of customers warming car seats nowadays who have never had to touch the throttle to start a vehicle, so you can't expect them to do it instinctively if the IAC valve sticks.
Spark plugs get really wet when there isn't any air to go with the rich fuel supply provided during the cold spin, and wet spark plugs don't fire because the voltage follows the path of least resistance. That's called flooding, and it can wash down cylinder walls in a flash.
A Throttle Position Sensor (TP) that is drifting into the Wide Open Throttle range with no throttle movement can cause intermittent hard starting concerns because most fuel injection engines shut the injectors down at Wide Open Throttle.
Dripping injectors can cause hard starting as well. As a matter of fact, one dripping injector can flood an engine all by itself (I've seen it several times), even if it isn't dripping all that much.
That gasoline steam that fills the plenum tends to condense on spark plug tips when air starts to flow.
Typically, a hard start due to an injector drip will show up hot, and the engine will spin for a bit (sometimes 30 seconds or more) before it starts and it might run rough until the wet burns off the spark plugs. Spin it until the moment it fires up and then shut it off immediately. Pull all the spark plugs out. The sooty or wet ones will be the ones with the dripping injectors. If the injector tips on those same cylinders are washed clean (they should be dirty) then you have found your leaking injectors. I've used that trick for years.
The Stubborn AltimaLet's take a peek at a special case I ran into recently. It helps to understand the principles outlined above, and our experience with this Altima hard start problem will illustrate that need.
She asked me to call the shop where the car was languishing and speak to the technician who had been working on it, and so I did. The tech's remarks were somewhat interesting and provided another clue.
"It's really hard to start cold," he told me. "But you can spray some carburetor cleaner in the manifold and it'll start right up and run just fine for the rest of the day. There aren't any stored codes, but I believe there's something going on with the injectors. I temporarily plugged in a used PCM but it didn't change anything."
Did you pick up on that important piece of data? He could add fuel and it would start normally. What did that mean? As mentioned previously, TP sensors can fail in range (drifting voltages with no throttle movement) and cause that kind of anomaly, but even with the TP sensor near the WOT range and the PCM in Clear Flood mode, the engine will start if the rpm goes above 400.
In the service bay with diagnostic weapons hot, I fired my first salvo.
Spark: Hot and bright, even during the hard start.
Spark timing: Dead on the money, even during the hard start.
OBDII Scan tool: No codes, no glaring discrepancies in the datastream.
Fuel Pressure: Normal (40 psi off and 30 psi running) with no leak-down at shutdown.
Battery Voltage With Starter Engaged: 10.9 volts and holding steady.
After the Altima sat fallow for a few hours, the Intake Air Temperature (IAT) and Engine Coolant Temperature (ECT) readings were about 10 degrees apart. That got my attention until I realized that the 10-degree spread might have been due to the fact that the heaters work so well in our shop. The IAT was the warmer of the two sensors, and it was between the inlet and the air cleaner where ambient air could reach it.
I keyed on the temp sensors because I had worked on a Dodge pickup that flooded the spark plugs and simply wouldn't start at all on a cool day. It turned out that the IAT sensor was completely open and the PCM had substituted 110 degrees, which wasn't such a big deal if the coolant was nearer the reported (substituted) air temp.
What was even stranger was that the IAT sensor had just been replaced. Incidentally, the OBDII datastream won't show the substituted value; it shows -40 degrees. Rule of thumb: All the temp sensors should read the same after a thorough cold soak.
Back to the Altima. With the scope connected to an injector line, I measured about 10 ms pulse while the engine was spinning and not starting (you could run the battery down spinning this one and it wouldn't start unless you released the key and tried it again), and about 3 ms pulse after the engine finally fired up. Without another Altima for comparison, I didn't see much wrong with those numbers.
I thought about it. I slept on it. I reflected on the Altima's wiring schematic and the fact that there is a dedicated start signal wire from the ignition switch to the PCM, and that there is a fuse in that circuit. If the PCM doesn't know the key is in the Start position, it won't know it's supposed to double pulse the injectors. If that was the case, it was firing the nozzles sequentially and not wetting the manifold sufficiently to fire the engine up cold.
Another piece of evidence that supported the improper fuel delivery was rooted in the fact that I saw very little needle pulse on the fuel pressure gauge with the engine spinning (usually the needle bounces pretty good during start).
Based on my hunches, we measured start signal voltage at PCM pin 20 with the key in the START position. There was no voltage. We searched for the location of the suspect fuse in ALLDATA, then discovered that it was fuse 23 and it was plainly labeled on the instrument panel fuse block cover as Start Signal. But there was no fuse in that slot. We installed one, and the hard start problem was gone. It was as simple as that.
Interestingly, the literature we had on hand didn't mention that as a possibility. It simply took a basic understanding of gasoline fuel systems to make the connection. The owner paid our basic charge and bought a fuse.
Richard McCuistian is an ASE-certified Master Auto Technician and was a professional mechanic for more than 25 years. Richard is now an auto mechanics instructor at LBW Community College/MacArthur Campus in Opp, Ala. E-mail Richard at [email protected]