Having looked at Ian's a lot I think the problem is that the pickup is on the left side of the tank and all the fuel just sloshes out of the sump area and into the right side of the tank during left handers... Mazda simply didn't design it for sustained 1.4 G turns followed by 0.5 G of acceleration all the way down the straight. Which brings us to what I think the right solution is:

It's called baffling. There are a lot of designs out there. The simplest ones are just some walls around the pickup to prevent the worst of it:

https://cimg6.ibsrv.net/gimg/www.mia...e91f5c975b.jpg

Labyrinthine designs work better in sustained corners:

https://cimg6.ibsrv.net/gimg/www.mia...5081241e3e.jpg

The very best ones have flappers to let fuel in from the high side, but not let it out on the low side:

https://cimg5.ibsrv.net/gimg/www.mia...60c6574c97.jpg

The downside to the above is they require welding inside the tank. It's a great idea for a new setup, but really a pain to add into the OEM tank.

Another solution is to add a swirl pot / surge tank:

https://cimg6.ibsrv.net/gimg/www.mia...8c1b407318.jpg

This is supposed to work flawlessly, but it requires two fuel pumps and a lot of pressurized fuel fittings outside of the main tank.

The most practical solution may be the Holley HydraMat. If they live up to the hype then $200-300 you get a bolt-on good-enough solution. https://www.holley.com/products/fuel_systems/hydramat/

All of the above is based on my research into ways to solve this problem. I don't actually have any experience with any of these so I'm still eager to hear from anyone who's tried solving this on a Miata.

My solution,should hacking that flange off and attaching a hose prove ineffective, will be to simply maintain more than 1/3 tank on track days.

That's what I've been trying to do in the short term. With the 300+ at the wheels and RR grip, I'm getting stumbling at 5 gallons left, which means it's probably going slightly lean without me noticing before that as well. 7 mpg on track times 6 gallons means 40-45 miles. That's 13-15 hot laps of Thunderhill, and at a little over 2 minutes per lap that gives me 30-35 minutes of hot laps.

moto IQ did a very cool surge tank/swirl pt on their FD project car, but, um, yeah. It's a lot of work.

Project [Rotary] FD RX-7: Part 3 - Fuel System (Chassis Side) > MotoIQ - Automotive Tech, Project Cars, Performance & Motorsports

--Ian

Took the FD to a "Japanese Super Car Meet/Cruise" today. I just did the pre-meet, didn't go on the big group drive up the peninsula to Treasure Island (too far, plus I don't do big group drives -- too much attention).

150 Supras, NSXs, FDs, and a few other types of cars there. It was like 1999 all over again, except without the Civics. :)

http://www.codrus.com/cars/fd/fd-jscc1.jpg
http://www.codrus.com/cars/fd/fd-jscc2.jpg
http://www.codrus.com/cars/fd/fd-jscc3.jpg

--Ian

Thats why I called it a splash guard and not baffling the tank. I'm thinking something square or cylindrical mounted to the fuel pump assembly.

But looking around, maybe aeromotives idea with "foam and bladder" is a better one.
https://www.aeromotiveinc.com/product/phantom-flex/

Or, as said before, the hydramat.

That's pretty amazing. Why haven't you guys jumped onto this yet?

Because for the most part the 3 people experiencing fuel starvation in miatas in the world, are in this thread. Its not exactly a common problem. And other cars have run similar setups without this issue.

I mentioned that several pages back and got no response.

I installed that mat in a fuel cell, it's a small PITA, but does seem to do that job, no more fuel starvation. The kit I used had cool glue down clips that secured the corners of the mat, but what they don't tell you in that video is that the NPT/AN hose for the pickup has to be positioned correctly so it doesn't move the mat around.

But Aidan's right, on our enduro car we use a completely stock fueling system that fuel starves on right handers, makes it back to the pits, and we fit an entire 10 gallons in a tank that's supposedly 11.9 dry.

This is not the correct part. It meets their print on Amazon, but does not fit NB fuel rail or return.

On this part, CtC mounting holes is 32mm. On Miata parts, CtC is approximately 39mm.

Does anyone know the correct P/N? I know that people do use these to turn to AFPR.

EDIT: This one, , appears to be more correct. And I found it much cheaper at Amazon.

Missed it! Here, have a cat. :)

So at the last few track days the wideband has been pegging at 22:1 about halfway through the second session. It fixes itself once it cools all the way down at home, and seems to work properly after that, but clearly this isn't good. I've got a replacement sensor, but rather than put it in and do the same thing to it I wanted to try to fix the problem first.

So, one piece of aluminum sheet metal and some bending in the vise:

http://www.codrus.com/miata/fm2r/wb-sink1.jpg

http://www.codrus.com/miata/fm2r/wb-sink2.jpg

Of course, then I took the car out for a drive and it started spewing sync errors and shutting off once it went into boost. I had the wiring apart last week trying to reverse-engineer the serial protocol going to the racelogic TC unit, and must have screwed something up. Doh.

--Ian

I suggest putting some kind of insulating washer between your heatsink and the exhaust pipe. Maybe a thick stainless steel washer?

You mean, other than the big wrinkly piece of metal in the picture?

I think Jason is suggesting trying to insulate the heat sink from the exhaust itself, so that it works more effectively at cooling the sensor.

I dunno what good a stainless washer would do, though, since the whole exhaust is stainless.

I have also heard that these kinds of sheet metal "heat sinks" are actually more effective as radiant heat shields. I dunno, we'll see how it does at the track next week. :)

--Ian

.

this actually sounds like a really good idea.
fm dp mounts it at the very top iirc, the hottest spot (but least likely to get damaged)

Yes, the FM downpipe points it straight up. It could probably lean over 20 degrees or so, but I don't think there's room to mount it horizontally. The LC-2 instructions say it should be between 9 and 3 o'clock.

Apparently Innovate sells this guy too:

https://images-na.ssl-images-amazon....L._SL1500_.jpg


$65 -- I may try that if the heat sink/shield doesn't help.

--Ian

Would that not cut down on the flow of gas to the sensor itself? Possibly slowing down the readings?

FWIW I've got an 1/16" thick piece of brass about the size of compact disc with a hole in the middle the perfect size for an O2 sensor. Gesso had it in his random box of scrap so I tossed it on when we did my exhaust and its been there ever since.

Well, Innovate makes it, so I assume it works with their sensors. It's got holes in specific locations and needs to be indexed a particular way.

--Ian