Have had recurring starting problems and just replaced the fuel pump. Started perfectly for a few days and just this morning would not start. I obviously did not fix the root cause of the problem.

Tonight when I get home I'll try to start it with starting fluid, but if that doesn't work, what's the next most likely (electrical) cause?

 

First, what makes you think it is a fuel issue?
If it is really a fuel issue then have you done a "fuel relay" by-pass test?

Hmm, maybe you should just throw in a new fuel relay???

Don't know what it's cause ... no test were done.
But, trying to start it with starting fluid is a beginning.
Just don't get that on the MAF sensor or you will only double your troubles

 

Car starts, then stalls, with instant-start into the intake. Checked fuel pump 15A fuse: Tested "Good". Removed and tested fuel pump relay: Tested "Good". Fuel pump relay clicks in socket.

New fuel pump makes no noise.

Next step: Friday night I'll check to see if fuel pump plug under the rear seat has 12V with the key "on".

The problem is no longer intermittent; the car now won't start. (That's a good thing.)

Anything else to check?

 

Relays clicking does not mean the relay is good.
Only that it clicks. Experience with bad relays tells me that.
However that might not be your problem?
Did you by pass the relay with a jumper to see if the fuel pump turns on?
If you don't know how to do it then look it up on youtube to see videos on it.

When things are quiet, turn key to the "ON" position and listen for momentary priming of fuel pump.
It will only turn on for a couple of seconds.
The ECU (engine control unit) will only power the fuel pump for a moment with the key in the on position if the engine is not running.
So checking for 12 volts at the fuel pump might be difficult.

 

Thx avisitor,

I bench tested the FP relay with a voltmeter and 12V power supply. The high current plugs are conductive when 12V is applied to the low current side of the relay.

I'll bypass the relay after I check voltage. Do the high current sockets in the box have 12V continually even if the ECU timer runs out?

I took great care to keep things quiet. The fuel pump made no noise that I could discern, even for a moment. I'll try again with the seat out and cover removed.

Thx for the tip on the momentary ECU power situation. I'll have my friend try to start the car while I check voltage at the plug. I assume that the pump is running continually when the ignition is in the 'start' position.

What other ECU things are going on???

 

You sound like you know how to test a relay.
On my Elantra (looking at the schematics), the high current side of the relay does have one terminal hot at all times.
If a jumper is put on the high current side then the fuel pump should operate.
If it doesn't then the circuit/ wiring has to be checked.
Yes, the pump should be running continually when the key is in the start position
Make sure there is power and ground to the pump.

Since you have already changed the fuel pump, I should mention that sometimes an O-ring is not transferred to the new pump and this keeps the fuel pressure from building up and the engine from starting.
It has caused many headaches for a few people.

The ECU takes signals from the crank and cam sensor to set the timing of the spark plugs and the fuel injectors.
But, the ECU shouldn't be your problem since you can start the engine from a can of starting fluid.
Well, at least until we know the fuel pump is not getting power to run.
Then, more tests will be needed to determine the source of the problem.

Sounds like you are well on your way to getting your car fixed.
Let us know how it goes.
Thanks

 

Here's a nice detailed pump troubleshooting video including a bonus describing power side switched circuit bias voltage:


Schematic: Wilson - TSB Database

Unfortunately, the Hyundai relay does NOT show a schematic printed on the side so I'll need to figure out the grounds and constant power load side pins. As avisitor said, the control side of the circuit (low amperage) will only power when cranking or when the car is running and the computer is getting a good signal from the crank sensor. The high amperage power side is always 'hot.'

Another nice, simple video:


Looks like I'll be working on this vehicle Saturday.

 

More testing:

Interesting that the high current side of the fuel pump relay is always +12V even with the key out. Control side of the fuel pump works as expected staying "on" for about ten seconds after the key is turned to the "on" position without starting the engine.

Power line measures 11.5 V at the plug to the fuel pump when key is turned on.

Wire resistance from power side of the relay (not constant 12V side) to the plug at the fuel pump measures 136 kOhms when plug is disconnected from the pump... seems way too high a value.

Ground side of the wiring harness plug going into the fuel pump shows no resistance to ground (good thing).

Fuel pump internal resistance measures 3 Ohms (bottom two pins on socket).

Fuel pump will not run shorting the high current side of the relay socket with a jumper.

Fuel pump will run when applying 12V directly from the battery to the pump terminals bypassing the factory wiring, fuse and relay. (Thanks goodness for the Power Probe III: https://www.amazon.com/Power-Probe-3.../dp/B000KPU8MK)

Small 15 Ohm fuel pump fuse looks good and measures 1 Ohm.

Do I have a bad wire? Is there anything else in series going back to the fuel pump?

 

136 kohms is very high resistance for a wire that goes to the fuel pump.
Little or no resistance on ground side is good and the power side should be similar.
It seems like a bad wire. Can't explain why the resistance would be so high.
But, it definitely is not right.

Well, seems like you found your problem.

 

This is a friend's car, rather than mine so he's going to take it to the local dealership for the fix. Maybe there's a recall/campaign due to a bad wire that the dealerships know that I don't.

The wiring harness runs through the interior so it will be a big job to splice in a new wire if necessary.

I feel bad; this is the first problem of this type that I've not been able to fix by myself.