Walbro 255 Problem

Confirm if your Cobra is a return (fuel pressure regulator) or return-less (Fuel Pump Driver Module). I suspect it is return style but apparently not all 98 Cobra's are.

IMO, it is plausible there is a bad high-speed fuel pump relay. For example, the PCM calls for high-speed and then due to a bad relay, power to the FP completely goes away.

My knowledge is not as good on the 1996-1998 DOHC. Does anybody know what causes the PCM to switch the fuel pump from low-speed to high-speed mode? Is it fuel pressure? RPM? Load? Or perhaps the PCM switches to high speed mode when the intake runners (IMRC) are opened.

FWIIW, I found a couple of posts that indicate the stock setting for when the IMRC's open is 3250 RPMs. Is that when the motor's power goes away?

Recommend that you data log the fuel pressure, RPM, TPS, MAF, Mode(open/closed loop) for starters. It would be handy to know when the PCM is switching the fuel pump from low-speed mode (6v) to high-speed mode (12 volts). What the car is doing when the switch occurs could provide clues to the problem. There is an ODB2 PID for that.

It is possible that it is switching into high-speed mode too soon? Since the Fuel pump is so much larger, this is causing excessive fuel heating. IMO, this is a long shot but as good of a guess as anything.

IMO, a fuel pump input restriction is also a plausible explanation. Please review the link below with information regarding drilling holes in the basket. Per the link, it is not a cut and dried solution.

http://www.fuel-pumps.net/fuelpumpsfaq7.html

Again, this is just guessing without some information to go on. Need test results.

Disconnecting the battery to make the O2 sensors come back on line is mearly going to delay the PCM re-learning. Each time it is done, the PCM has to start back over from square one. This is not doing yourself any favors.

SCT is good. As are many others. In my case, I use an older single line Auto X-ray unit. Can't do data logging which is a problem sometimes. If I had to do over, I would get a multi-line unit with data log capabilities.

However, it is surprising what can be done with even a single line unit when it is used correctly.

Bottom line is that what ever ODB2 scanner you have, there is a learning curve to get the most value from it. Some of these scanners can display a huge amount of data that can be very daunting.

IMO, this problem can be solved without an ODB2 scanner. However, it will be much harder.

For example. Hook a volt meter to the IFS in the trunk. Run a long set of test leads into the passenger compartment. Monitor voltage at the IFS. Drive the car. If you see the voltage drop (or remain the same) when the motor cuts off, this confirms it is the fuel pump speed relay.

An external fuel pressure guage can be used to monitor FP. Or you may be able to attach a VOM to the fuel pressure sensor and monitor inside the car. Again, the VOM will display voltage (not FP). It would be necessary to convert the voltage display to pressure. Not straight forward.

However, it gets to be a problem when driving around. Get the picture?

I also suspect that you may not fully understand what happens if the FP is too large for the requirements and the effects of heating the fuel. Remember that in a return system, the fuel is pumped in a continious loop. Each time through, the fuel is heated just a little. If the fuel gets too hot, it will vaporize in the lines. The gasious fuel is not dense enough to keep the AF where it needs to be. Note, this is not very likely. I was just trying to make a point about what it COULD be.

Did you put a pressure gauge on the feed line where the schreder valve is? You need to get a reading of the PSI and see where it is at idle vs. WOT
Also, make sure that the pump in not bottoming out in the basket. On my 98 there are spacers to avoid this exact issue.