Return Fuel Stumble

Got the car out for the first time this year after switching to a return fuel system over the winter. It cruises around great, but when you step on it, as soon as the turbos start to spool it flat out falls on its face. I'm not talking breaks up or goes lean, it completely misses until you let out. Seeing as it ran beautifully with the returnless setup, I can only assume the issue is fuel system related. Has anyone ran into anything like this when they switched? I haven't got a chance to datalog it but I'd be curious for things to keep an eye out for.

Previous returnless fuel system steup: Cobra tank, dual GT pumps, dual FPDM, wiring upgrade, stock fuel lines.

New return setup: Sumped GT tank, Fuelab 41402 pump, -10 supply, -8 return, Fuelab regulator @40psid verified by mechanical gauge and readout, Fore rails, deadhead return system.

My first gut feelings are that I'm not getting enough flow from the pump due to mechanical issues or insufficient power. The pump is being run at 50% duty cycle, and according to the calibration sheet it came with, I should have more than enough at that. At 65 psi (40psid base pressure, +15psi boost, +10psi restriction) it should flow 90 GPH/9.45 lb/min, which at 10:1 AFR should be enough for nearly 900hp. It also only draws 11A at 65psi, which should be even less than the old GT pumps. Everything about this should be absurdly overkill.

You were retuned for the return style system right?

Yes I was. I tune it myself. I did it to retain the FRPS, and the FPDM is still physically connected. The fuel pump relay is triggered off the FPDM power wire. The fuel pump adaptive PID parameters were set to 0, fuel pump voltage table set to 15. Unfortunately didn't have the fault code switches to turn off so I've got a CEL, which I imagine is for a fuel pump circuit open fault.

Since you kept the FRPS, what do your datalogs show for fuel pressure?

Haven't got a chance to datalog it yet, as mentioned in the OP. Hoping to get a chance today or tomorrow. Fuel pressure and voltage are certainly on the list.

Got out for a datalog session and found my issue. It's fuel pressure. I've got 40psid key on engine off, dropped to 38psid cold idle, by the time I got cruising down to 36psid, and all the way down to 32psid on acceleration.

Fault Tree Analysis (FTA) time. For the most part it boils down (no pun) to either insufficient pump performance or insufficient regulation.

Pump-
Insufficient current/voltage - unlikely. 10ga wiring and 13.5V steady.
Insufficient output - calibration sheet from the factory shows it should flow far more than I need, even at 50% DC, as mentioned above. Mechanical issue is still possible. Could try running at 100%DC, though that's more likely to overheat the fuel, but sure as hell is going to give me enough volume/pressure.
Insufficient fuel feed - Fed from a sumped tank, a <12" long -10AN line, at sump level, half a tank full, Earl's 100 micron filter. Biggest issue could be overheating of the fuel due to the high volume being returned, and would agree with pressure decaying over time. Pump wasn't whining like it was aerated/cavitating. I wouldn't think that it would put enough heat in the fuel in a less than 10 min drive with a deadhead regulator, 50% DC, half a tank, and returning below fuel level in the far corner of the tank, but who knows. I might see if I can put a thermocouple in the sump to monitor the temperature.

Regulation-
Poor vacuum reference - Shows fine at atmospheric pressure, engine off, and rail gauge pressure, not delta pressure, drops when engine is on generating vacuum. If anything you'd think that would mean too much pressure at idle and still being spot on when going WOT as IMP approaches atmospheric. I will note I removed the stock vacuum tree and plumbed my own using 3/16" line. The FRPS, boost gauge, wastegates, and FPR are all teed off the lower manifold.
Poor configuration/hardware issue - Deadhead setup, -10AN supply, -8AN return. It's then -10AN all the way from the regulator to a -10AN Y block and still -10AN to the rails. I haven't seen anyone else see issues with a deadhead setup like that.

Anything I'm missing?

I'm going to guess its something with the regulator. The pressure should drop about 8psi at idle. So if you set it to 40psi with no vacuum connected then connect the vacuum line the pressure will drop to 32-33psi. At least that's what mine does.

Then when you hit boost the pressure should go up roughly 1psi per pound of boost. So make sure the source you're using for the regulator sees boost and vacuum. I teed into the boost line for dash by the firewall.

01yellercobra said:

I'm going to guess its something with the regulator. The pressure should drop about 8psi at idle. So if you set it to 40psi with no vacuum connected then connect the vacuum line the pressure will drop to 32-33psi. At least that's what mine does.

Then when you hit boost the pressure should go up roughly 1psi per pound of boost. So make sure the source you're using for the regulator sees boost and vacuum. I teed into the boost line for dash by the firewall.

Click to expand...

'

Just as a matter of clarity-

When I refer to base pressure or delta pressure, that's the same as "pressure drop across the injectors"- it's the pressure of the rail minus the pressure of the manifold. That is the value that the ECM reports through the datalog from the FRPS. It should stay as a static value all the time, regardless of what manifold pressure is doing. Mine appears to be dropping as time and load go up, which is no Bueno.

As far as rail gauge pressure, that is pressure in the rail relative to atmospheric pressure, regardless of manifold pressure. That's what's 40psig with key on, engine off (reference pressure = atmospheric, 0psig), and with the engine on, idling at cold start is around 25psig. If maintaining 40psi base pressure that would mean the engine is pulling -15psi (30inHg vacuum), which is a pretty reasonable value.

Now for vacuum source, the FRPS and FPR are both tapped into the same source, which is the lower manifold (post blower) That's the actual value they should be looking for, and also the same port as the boost gauge, and what he FRPS is normally tapped into.

Latest update- went through the Fuelab checklist. Theirs seems to point toward cavitation of some sort. It doesn't seem to have any of those causes though.

Excessive restriction/low fuel inlet pressure. The inlet line is -10, very short, and has an 85 micron filter, all of which seems to match their recommendations. If it persists, I might try to put a gauge somewhere to see what the actual value is at the pump inlet. Fuel tank vacuum due to a faulty vent valve is a possibility as well. I didn't get any hiss opening the cap, but I'll try running it with an open cap just to make sure.

Aeration. The return line goes to the corner of the tank below fuel level. That's as far away from the sump as possible, and shouldn't churn the fuel when it enters. The pump is at the same height as the sump, which is to say well below the fuel level, so it's very unlikely to trap or suck air.

Temperature. Took it out for a drive. Weather was less than ideal with a very cold ambient and high wind, but I drove it until it started to lose rail pressure. It took about 15min of city driving. Fuel was drained from the pump and measured with a thermocouple and IR thermometer. Both indicated no more than a 15 deg F increase in fuel temperature, both well below 80 deg F. Lines, filters, and pump were lukewarm at best to the touch, again well below 80 deg F. Granted that methodology is less than ideal, but it would seem to indicate that it's not getting very hot. Cavitation shouldn't be an issue until 120 deg F, so I would think I should have margin. Further testing is warranted with real time temperatures measured in the sump. Probably won't get a chance until later in the week when ambient temperature comes up.

Beyond that the next question comes to regulator function. If none of those efforts bare any fruit, then when pressure loss occurs, I'll try cracking open the return line to see if there's any flow below the set point. If there is, that would mean a faulty regulator.

Just more food for thought. Stay tuned for more testing.

I would try and wire the pumps so they come on full time with the ignition. Also make sure you didn't wire one of the pumps backwards.

Sent from my Pixel 2 XL using Tapatalk

It's a single big pump, no worries of being backwards. It's wired to FPDM power, which is basically with ignition.

Changing it to 100% DC rather than 50% DC right now is an option, but if there is a heat problem that would only make it more pronounced.

Got done calibrating the thermocouple for the sump and installing it and took the opportunity to crack open the pre filter. Found out it's not a true element, or even a large diameter screen. It's a large body filter, but the effective flow area is still about 5/8" with a single layer screen. Here's where I think I found my issue, and it's completely in SVTP fashion. I find some gunk on the screen. I thought I was remarkably thorough in keeping things clean before assembly. But there's all this fiber, and then some really thick fibers...wait, no, those are legs...and...oh god...that gunk was a spider and web. Not a brown recluse, but still apparently got the Mazda treatment. Hoping that resolves my issue.

Lmao

So the spider was in your tank?

I can only assume so.

So did your car survive the flamethrower? Assuming you killed said spider appropriately...

MarkOPKS said:

So did your car survive the flamethrower? Assuming you killed said spider appropriately...

Click to expand...

Well it's already called the dumpsterfire so...

Might have to burn it down- clearing the filter doesn't seem to have fixed the problem. I've ruled out fuel temperature as a cause. Temperatures didn't even exceed 70 deg F.

I'm also fairly confident the issue here is around the regulator rather than the pump. Even after fuel pressure decayed to 34psi, I could adjust the regulator and I was able to bump it up to 50+ psi on a whim. If the regulator was truly chasing the set point and was unable to deliver it (i.e. delivering 34psi on a set point of 39psi), then the regulator would have been fully shut returning no fuel at that point and adjusting it would have made no difference. But I can bump it up all I want and the pump will deliver it. For whatever reason it's not maintaining the correct base pressure. I can't even chalk it up to a vacuum leak, because if anything that should cause the base pressure to be too high.

On the finer points of testing it, my first attempt it dropped 5psi within 10 minutes of slow driving. Even worse than before. So I tried bumping up the regulator set point back to where it should be while idling stationary. That was easy. I let it sit and idle for 10 minutes and it varied less than 0.5psi. But as soon as I took it out on the road it was dropping again. Something is just...odd.

Regulators should be pretty simple devices. You have a valve/seat, vacuum chamber, spring and diaphragm. Something is not behaving the way it should. I can probably rule out diaphragm since it's not leaking fuel everywhere. Which means either the spring is varying, it's getting too much vacuum (if that's even possible?) or something is wedging in the seat. Maybe another spider. Who knows.

Upon disassembly I found a couple of gouges in the regulator seat, one is pretty deep. I can't imagine they would be enough to cause the issue (or even that this small 1/4" hole is enough for a regulator).

I suppose I could re machine the seat at the cost of compromising the annodized layer. After all it appears to have been compromised now anyway. I'm still surprised at the design- it's a ball on the diaphragm and a very shallow taper on the seat.

I suppose I'm open to other recommendations of regulators as well.

I'm running a Fore regulator on my car. A couple friends ran Aeromotive with no issues on their set ups. One was an 800hp beast.