'78 928 -- Fuel pump not priming

[MrGreenJeans]

Thanks for all the help so far. Just completed a fun road trip to the mountains and we made it there and back no problem.

My remaining issue is that ever since my fuel pump relay went bad my priming function has been absent. It did get hot and discolor the relay socket, but the wiring appears to be fine. I have 3 FP relays that seem to be functioning otherwise (car runs), but cold starts sometimes require some cranking to get fuel running through the system.

I have a new ignition switch installed.

Side note…my cold start injector was causing hot start flooding that a new thermo time switch did not fix, so for the time being I’ve left it disconnected

[dr bob]

The actual priming function for the pump is managed exclusively by the fuel pump relay. The internal protection that differentiates the relay from a plain old SPST relay has a circuit that monitors ignition pulses in parallel with the signal to the tach, looking for the ignition box to pull that tach line low for each cylinder firing. If the relay doesn't see a pulse within a second or so after the key is turned to run or start, the relay opens and the pump stops. On cold start, the pump relay energizes for that first second or so, running the pump to pressurize the system. If you don't actually crank the starter, no pulses so the relay opens. The timing circuit depends on a small capacitor to help it remember that it already primed the system, so you'll find that multiple consecutive off-to-runs/start transitions won't close the relay for system priming again.

The cold-start injector really needs to be functioning for reliable cold starting. The CIS system uses a mechanical airflow paddle that operates the injector 'plunger' pistons in the metering unit. No airflow means no fuel flow, and the linkage between the paddle and the metering unit is set with a tiny amount of low-flow deadband so that it doesn't allow fuel until a minimum plunger movement is detected. The cold-start injector offers just enough fuel to bring the engine above cranking speed. The thermo-time switch determines the length of injector pulse needed, but even a warm engine can use a short spray to get the engine to fire and turn fast enough to draw air and lift that paddle. If you have a cold-start injector that sticks open, closes OK when disconnected, and a new thermo-time switch doesn't solve the symptom, it's time to look at he wiring itself between the two. Specifically, verify that the injector sees battery voltage on one side with key in start position only, but the other terminal goes to ground only momentarily and only while in start position/cranking. If it stays low (towards ground for more than that short pulse time, unplug the thermo-time switch and test again. If that wire stays close to ground even with the switch unplugged, you have a wiring fault to ground between the injector and the W terminal on the switch, and the injector will stay open whenever the starter is engaged.

With the switch plugged in, verify that you have battery voltage on the G terminal with key in start position only. This the same wire from pin 14 in the 14-pin connector by the jump post that engages the starter solenoid (circuit 50, yellow wiring)

----

Over the years, some issues have been reported with wiring at the starter solenoid, particularly with replacement starters. There are two smaller connections at the solenoid. One is the yellow wire at the 50 terminal, the other is an ignition booster at the terminal marked 15A on the solenoid. This circuit returns battery voltage while the solenoid is drawn back, and bypasses one of the ballast resistors in the power to the ignition coil. Intent is reduce the effect of ignition voltage draw-down while cranking. It's pretty important that the solenoid and the connecting wiring are connected correctly. That 15A current passes back through pin 10 in that same 14-pin engine harness connector. It's a little confusing when testing there, as key-on supplies some voltage there via 15 and one ballast resistor with key in run, so testing needs to be done with key off (no voltage), then with key in Run (close to battery voltage), then with the connector pin extracted from the top half of the connector shell with key in start, looking for close to battery voltage that disappears as soon as the starter is no longer cranking.

Hope this helps. It's a bit complicated but can be essential to troubleshooting both the cold-start injector circuits and the power to ignition while cranking. Both affect starting.

My last CIS car would fire up almost instantly on a bump of the starter motor. If it took longer than a second of cranking I was worried, a by a few seconds and I was planning an engine overhaul. Yours should be working that well too.

[MrGreenJeans]

I’ve learned a few things. One is that the early relays don’t do the the prime in position two. The one that I had in that went bad was a different relay, so perhaps the lack of priming in position 2 is not a fault after all.

I plugged the cold start injector in this morning and it cranked right up, so I think my next move is to determine why it’s activating when the engine is hot.

I’ve started by checking the time switch according to the WSM, which had me disconnect the switch and check resistance at both pins. Below 30C both pins are supposed to measure 0 ohms, which I get at the W pin but but I’m getting 33 then measuring both together it’s supposed to be between 25 and 40 which it was. I haven’t checked it hot yet.

Does a ground resistance reading when there shouldn’t be one tell me the switch is bad?

[dr bob]

I don't know. I'd have to consult the WSM which I'm away from right now. I know there's a test spec, one I have tested/verified in the same handy workshop toaster oven I use to warm lunch sometimes. IIRC, that switch has only two connector pins and relies on the body path to ground via the coolant manifold. I readily admit it's been a long time since I dug deep into a K-jet 928 like yours. In my advancing years I struggle remembering even half the stuff I've forgotten.

[MrGreenJeans]

Dr Bob I hope you see this. After many trials and tribulations I have determined my starting woes have been related to two instances of aftermarket wiring mods. The cold start valve was wired to be always on start up, which caused hot start flooding. The fuel pump relay circuit had a Push/pull switch wired in between the ignition and the O connector at the fuse panel I would assume as a theft deterrent. I have removed both and set it back to stock wiring and all is right with the 928 world.

Throwing this out there for anyone chasing the same problems. Wiring mods are guilty until proven innocent.

[dr bob]

I will remind other readers of some actual crimes against future owners. One high on the list is undocumented wiring changes, things done in the heat of battle to try and solve a symptom rather than drill down and solve an actual problem.

Folks who have endured my "logical problem solving" seminars in the power industry may be only slightly more familiar than my Carpokes friends, and my focus on using "symptoms" to narrow down actual "problems". "My car won't start!" is a symptom. "The battery is dead" is another symptom, one that might contribute to the first symptom, but not the real problem, the root cause of all these symptoms. As car problem detectives, it's our job to look at the various "symptoms" and build a probability matrix of possible "problems" that might cause those "symptoms".

Then there are "wildcard" conditions, like prior wiring hacks, that can easily mask the symptoms and partially get around the problems that cause them. Like jumpering around a safety system like the full CIS fuel pump relay protective functions. I guess, so long at the paddle piston in the metering unit shuts of fuel flow completely when the paddle at rest, some of the relay's protection function is achieved. But it doesn't always do that, especially if a less-than-expert owner or technician decides to adjust the mixture screw.

The "fuel pump relay doesn't close to pressurize the system on cold start" describes a "symptom". I can just put a normal 53b relay in the socket, and waste no more thought on needing to know if the engine is actually spinning beyond cranking speed. What could possibly go wrong with that?? Start-up flooding, and that annoying fire after an accident. The engine stops, but the fuel pump continues merrily feeding the fire even after the engine stops. Bypassing the thermo-time switch to get full cold-start valve opening is another serious risk.

------

Glad you were able to identify the wiring mods/hacks. I have some small sympathy for adding a manual anti-theft switch to the fuel pump relay, but I struggle with how to make it effective without some serious 'religion' on actually using it every single time the car is parked, without fail.

On a prior car, only coincidentally a CIS car, I added a relay and a bypass around a small 5A self-resetting thermal circuit breaker. On start, the breaker is a thermal 'time delay' that causes the fuel pump to stop for ten seconds or so, run for a few again, then stop again. Makes the car hard to drive very far or very fast. If the driver closes the relay contact around the breaker, car drives normally. Importantly, at all happens at the fuse socket, with no actual wiring changes. "Setting" the bypass relay is done via the high-beam switch and the high-beam fuse, again no hacks. Key on, engine start, pull the high beam lever within 5 secs of fuel pump start, drive away normally. Forget, and the car stumbles and stops. Yeah, a thief or carjacker could figure it out after a while after staring at the CE panel and a wiring diagram...

Anyway, inspecting a car and finding permanent and particularly undocumented wiring hacks very quickly diminishes its value to me. My 928 has only one wiring change, a now very well publicly documented change to the AC clutch coil circuitry, with a higher-capacity relay attached to the dash controller. To avoid the wrath of the originality police, I did buy a replacement stock/original controller should it be "needed" by a future owner. Along with the logbook entries the exactly describe the modification. It's easily reversible with a few easy tools, reversing the directions shared in the logbook.

Again, thanks for sharing back the details, and glad you were able to solve the odd symptoms.