View Single Post
Old 07-17-2019, 07:12 AM   #289
John V
No more BMWs
 
John V's Avatar
 
Join Date: Apr 2005
Carmudgeonly Ride: Ram, MS3, CX-5, RX-8
Location: Glenwood, MD
Posts: 14,753
Now that the crank sensor issue is fixed, on to the next problem I've been having. The fuel requirements for this thing exceed the ability of the stock series 2 RX-8 fuel bowl to stay full. The volume of fuel inside the fuel basket is pretty small because the pump and filter assembly take up probably 80% of the volume. It's refilled by gravity. The other problem is E85 requires about 40% more fuel by volume than gasoline to achieve the same lambda. So a good pull at WOT drains the fuel bowl faster than it can refill. I've got a Deatschwerks 340LPH pump in the stock housing feeding a return-style system where the return dumps directly into the fuel basket. Below is a pic of the stock fuel pump basket.



The fuel bowl is filled by two sources - one is the siphon from the passenger's side of the tank, which is not very efficient. But the primary method it's re-filled is by gravity, through a small orifice at the bottom of the basket. Sorry for the blurry pics, they're not mine.



In the middle of the pic you can see a small piece clipped onto a plastic protrusion on the very bottom of the bowl.



That little flapper is a one-way valve that allows fuel to flow through the aforementioned pipe when the fuel level in the fuel tank is higher than the level inside the fuel pump basket. The idea being that the pressure is equalized when the levels in the tank and inside the basket are the same. The problem is that orifice isn't large enough to keep the fuel pump basket fed when the engine is at WOT. When the tank is completely full, overflow through the top of the basket is sufficient to keep it fed (plus the pressure through that orifice is greater), but as the level in the tank drops, that no longer works. It's easy enough to test - at the fuel rail if I disconnect the return hose and run the pump, after a few seconds the fuel flow slows down significantly as the fuel pump basket starts to run dry. Turn off the pump for a few seconds, let the basket refill, and pressure returns.

The flapper is much larger in diameter than the opening in that pipe. I didn't measure it, but it looks to be about 12 or 13 mm. I decided to try to open up that pipe a bit to improve the flow. The plastic is extremely brittle, so I drilled it out really slowly. An 8.5mm bit followed by a 9mm followed by 3/8", which is about 9.5mm. I didn't try to go larger because the plastic was starting to flex pretty badly. Going from 8mm (50.2 square mm of area) to 3/8" (71.2 square mm) is a 42% increase in area. I also modified the flapper so it can swing open a little farther. I'll test this later, but I'm hoping that's enough to keep the fuel bowl fed even when the tank is at 1/2 or less.

I got to test this at the event this past Sunday and it was a success. No more fuel starvation and even better, no misfires with the new crank sensor! The car was great. I had the car down around 1/4 tank of fuel and will try running it even lower next weekend.

I had messed with the boost control algorithm to try to make the boost level more consistent (it tended to hunt around the target a bit which led to more boost than I wanted) but it made things worse. I lost the super fast spool that I had before. I'm going to do some more testing this weekend to try to get the turbo to spool up faster. Overall though, really happy with how the car worked. It was a really hot day and the thing was just a trooper, no issues at all.

Last edited by John V; 07-18-2019 at 05:23 AM.
John V is offline   Reply With Quote