Gradual hard starting on my 87 led to some diagnostics that others might find valuable.

Here is what happened: Normally, when cold, the car would start within three seconds of cranking the starter. Over time, it took little more time -- maybe four seconds of cranking. Then even more time. No ECM codes were set. Once started the fuel injection system worked as one would expect during a cold start. The idle would advance and, based on driving away immediately, the mixture for cold running before warm up worked just fine. As the engine warmed up the idle would reduce and the engine would run normally. During warm starts the car started normally, within the usual three or so seconds.

Very puzzling so I looked at the manual.

When first started the ECM turns the fuel pump on for about two seconds to build pressure regardless of all other conditions. Cold starting problems are often thought to be related to this prime function on the theory low fuel pressure will cause less fuel for the required rich mixture. To test the two-second prime you simply insert the key, turn to "on" and listen for the pump to hum for the two seconds. Also, you can hear the fuel pump relay on the driver's side of the firewall click in then out. This seemed to operate correctly but I was a bit suspicious so every time I drove the car I would listen for the fuel pump to prime. On a couple occasions I **thought** it did not do so -- I did not hear the pump or the relay click. Cranking the car a bit more than usual however got it started. Cheap relays often have problems - during cranking there is low voltage and, over time, contacts get pitted, corroded and make poor contact. Maybe the relay was working most of the time but not providing enough juice to the the pump.

Off to the Chevy dealer for a new fuel pump relay, about $25.00 and a ten minute change out. Inspecting the terminals on the relay it appeared that they were drenched in black goo -- probably adhesive or potting compound dripping down into the connector. This was a good hint that it might be bad, or at least intermittent. As relay contats go bad their resistance goes up and they can get hot, maybe causing the goo problem. The connector terminals on the wire side were cleaned with alcohol and burnished a bit with a relay contact cleaning tool and reinstalled. The new relay click audible from the driver's seat was noticeably louder. As a double check on the available fuel pressure, I waited for five minutes after the pump primed, put a rag around the fuel line port containing the shrader valve and gave the valve a single press. Lots of pressure so it was evident that all was well on the fuel supply side.

Still the problem persisted.

On to the cold start injector. For 85-88 the fuel injection system includes a cold start injector, for a total of nine injectors.

For some reason known only to the General, the regular fuel injectors were unable to reliably get the mixture right during cold start cranking so this injector was added. Looks like a major kludge to me since after 88 it appears they fixed the problem and the cold start injector disappeared.

The cold start injector is completely independent of the ECM so no codes are set if there is a problem.

Here is the way it works: power is provided by the starter circuit through a fuse. That is, power is only applied to the cold start injector when the starter is turning over. Once the starter stops, the cold start relay is deactivated. On the ground side of the injector there is a thermo-switch, also called the cold start relay. The thermo-switch controls the period of time the injector squirts gas while the engine is cranking. The time is determined by both the manifold water temperature and how long the engine has been cranking. Manifold water temperature above 90 deg F. or so inhibits the injector so no added cranking fuel is delivered. And, if you crank for more than 8 seconds the thermo-switch turns off from internal heading to avoid flooding. This allows flood recovery when you use full throttle during cranking to tell the ECM to use the lean flood recovery mixture.

First on the check list was to look at the fuse. Naturally the fuse is located behind the front dash panel behind the idiot light panel. I am wary of digging into this due to how fragile the plastic parts are so decided to check at the injector itself. Naturally the injector connector looked like the tab to release the connector was inaccessible. Not wanting to break the connector I looked for the thermo-switch. Naturally it is buried under the throttle body behind smog thingys. In short, no easy access.

So, back to the injector itself I got a stout needle with a plastic ball at the end to stick through the wire insulation to the bare wire to meter the circuit. Using an analog meter so I could see the meter needle momentarily deflect (digital meters aren't so good at this) the 12V side of the injector was getting juice while cranking. Then I checked the ground side for voltage. It **should** stay at ground (no volts) as long as the thermo-switch provided a ground path to start then go to 12 volts when the thermo-switch cuts off fuel. It appeared to cut off the fuel almost immediately after the starter was engaged.

So the next cold morning I did a second test. Using the same needle on the ground side of the injector I jumpered it go ground -- bypassing the thermo-switch entirely. This would give full injector squirt for as long as the engine cranked. The engine started instantly.

Conclusion: the thermo-switch was not providing a ground path long enough for the cold start injector to pump enough gas into the manifold for a cold start. It was sorta-kinda-maybe working but not quite long enough.

Off to the Chevy dealer and a shock: $81.00 for the little pup.

To get to the thermo-switch I removed the air hose between the throttle body and the mass air flow unit, then disconnected the mass air flow sensor connector and removed the whole mass flow unit, then pulled the wiring connectors out of the way with a twisty. The hot water hose between the manifold and the throttle body was pulled (used a pan underneath the engine to catch the coolant) and all the remaining wires were disconnected and pulled aside. Then the smog control thingamabob, via the bolt facing straight down, was removed keeping track of where all the hoses went.

Using a deep socket the thermo-switch simply unscrews out of the manifold. I waited for most of the water to drain out of the hole, applied a bit of thread sealer to the new thermo-switch and screwed it back in. It uses pipe threads so it just tightens up as it is screwed in. All the connectors got a squirt of Deoxit cleaner and were then reconnected.

After putting everything back together again it started right up, just like new.

Couple notes: the Chevy manual has two wiring diagrams in two different places that show the thermo-switch in the normally open and normally closed position. The correct diagram is the normally closed position when cold. Also the wiring diagram says that the fuel injector relay is activated on start up only. At least on my 87, it is on while the engine is running.

If you have a cold start problem look into the relay/switch area -- these don't seem to last very long and cause goofy problems that are hard to diagnose.

 

Good diagnosis. I believe that my cold start switch is wonky on my car. One of these days I might get around to fix it.

 

r pogo, nice write up, thanks for taking the time to let us know the procedures and diagnosis. Another one gets printed of for my book.

 

The goo is normal. It is a form of sealent to keep the water/etc. out of the plug since it is not a weatherpack type connector.

 

Saved to hard drive. Thanks!

 

Went straight to print and in my nugget file. Thanks for the time and great write up.

 

Dittos! Thanks for the time and a great write up.

 

My 87 starts funny when its cold also.

 

Wow, I wish all posts were as thorough as this one! I like that you explained what led you to each diagnosis.