I've got a 1994 Elantra with the 1.6L DOHC and manual transmission. It doesn't matter if the engine is warm or cold, this is always happening. While driving (it happens the most often between 2000 and 3000 rpm while accelerating) the engine will shudder really violently like someone just kicked the crap out of it. It seems like it just shuts off for a moment then comes back on. You can hear a clicking noise from behind the driver's side dash. It seems to be the ecm (mpi) relay, but I ran the a complete test on it using direct voltage from the battery and a multimeter, and it seems to be working perfectly. I also checked all the corresponding connections from the ecm relay to and from the computer, as well as the voltage coming in from the battery. Its all working perfectly. I really don't have any other ideas on what it could be. I figure if the ecm was screwed, the car wouldn't run at all. Also, if the fuel pump was cutting out, the car shouldn't surge like that because there should be enough fuel and pressure left in the line to let the car run for a little more even if the fuel pump completely gave out. I thought it might be a fuel pressure related problem, but I don't have a fuel pressure guage unfortunately. However, I was thinking if it was fuel pressure related, I'd have a hard time starting the car, like if there was a clogged fuel filter or something. But the car starts beautifully everytime, hot or cold. And this doesn't happen when idling, only when driving. I replaced the clutch last year, but I doubt that has anything to do with it because the car ran great for like 6 months with the new clutch. But if the clutch was sticking for whatever reason, I don't see why it would affect the ecm relay and make turn on and off like that.

The first time it happened was a few months ago, and it would just get progressively worse and worse, until finally it blew out the 20A MPI fuse under the hood. I replaced the fuse a couple of times, but then the alternator went. Now I don't know if the two are related, or if it was just coincidence, but when I replaced the alternator with a new one, the problem seemed to disappear for about a month. Now its happening all over again. I pulled the alternator out and ran a complete diagnostic on it, and its working perfectly.

So I'm out of ideas. Any insight will be greatly appreciated. Thanks.

 

To me these are classic symptoms of a dying fuel pump on a multiport fuel injected system. Not all the injectors run dry at the same time, so there's momentary mis-firing on some of the cylinders and not others. When fuel is restored, not all injectors get equally fed until the fuel rail is fully pressurized. The engine runs smoothly at idle because the pump only runs intermitently since there's little demand for fuel. At speed it's running constantly or near constantly. It could be argued that some ignition component is involved, but the wiring schematic for this car shows the fuel pump motor directly wired to the MPI relay. When the pump motor overheats, the MPI relay kicks out (which you hear). Since the pump appears to resume operation so quickly after the relay resets itself, I think this is only the beginning of the end. As you've noticed, frequency will increase and eventually the pump may overheat to the point it takes several minutes for it to cool down and resume operation instead of a few seconds. Also, this situation should get worse on hot days and with less gas in the tank. A fuel pump with a partially corroded ground will sometimes act in similar fashion, but the relay will not respond like you've indicated and you've established that voltage is good.

A 1994 car with a bad fuel pump, I would surmize that there's a fair number of miles on the vehicle. I think the bad alternator was coincidental.

 

My mom had a 1984 Camry long time ago that would all but stall as soon as it reached 40mph. It turned out the fuel filter had so much goo it looked like mud. New filter and the car gave her many more years until the tranny went out... at 250K miles.

Fuel pump going bad (could be caused by the above mentioned filter) sometimes can be detected by an increased current draw; it is needing more energy to move.