Carpokes

Verify that any suction strainer in the tank is clean and free-flowing, as almost any restriction in the main pump suction risks the little bubbles that cause the screaming.

On the vapor pressure question, the answer is, as always, "it depends...". Refiners are required to manage the component blends of both fuels to manage the Reid Vapor Pressure of the fuels particularly in summer months. A typical max allowable pressure allowed is 7.8 PSI, but in some areas that number can be as high as 9 PSI, to limit more of the evaporation that's contributing to the GHG problem. We typically cycle between winter and summer fuel blends, since the colder winter temps reduce that evaporation on their own. No need for a great 90ºF pressure if it doesn't get that hot in winter I guess. Bottom line for you might be whether your tank still has a winter-blend fuel in it, and that's allowing a lower vapor pressure in hotter months. It may be that your fuel supplier or local station doesn't move a lot of E85 and there's residual winter blend remaining in their tanks.

The seasonal formulation changeover times vary depending on local climate, so check local/regional APCD rules. In SoCal, October and April seemed to be the shoulder months when I lived there. The difference between summer and winter temps was minor at best; winter tended to happen in the week between Christmas and New Years holidays, notable as the only period when I wouldn't need AC in my black car. Summer temp ranges in "Silicon Valley" are, um, all over the map. Sausalito could easily be more than 20º cooler than South San Jose on the same summer afternoon.

The higher flow the pump capacity is, the more fuel circulates through the lines to/from engine bay. It causes the fuel to heat up.

While cruising, fuel consumption is low. However, more fuel is returned to the tank. Perhaps, controlling the fuel pump based on engine load is an option to keep fuel cooler.

Of course, determining the cause of the pump noise is yet to be determined. Then based on the results the proper action can be taken.

Took the car out today (mid 80's ambient) and drove it about 10 miles or so. Fuel pump was quiet the whole time. When it got loud before, it was after a much longer drive (90 miles +/- round trip). I assume that means the fuel didn't get hot enough today to create the issue? One day this week, I'll head out for a longer trip and see what happens....

Hi Tom,

I am somewhat late to this thread, but both @Fast951 and @dr bob raise salient points about the relationship between fuel temperature and pressure.

A couple of quick tests you can try: when the temps are hot and the pump is screaming, turn off the engine and remove the fuel filler cap. Do you get a hiss of air as the cap is removed? If so, you might have a fuel tank venting issue, which would effectively create a vacuum in the fuel tank (from fuel pump suction), which would slightly reduce the boiling point of the fuel as the fuel level goes down and hot fuel is reintroduced to the tank via the return line. However, this condition would likely also cause fuel delivery issues.

Another quick test would be to recreate the symptom (presumably with a low fuel level), fill up the tank and go flog it again in hot weather and see if the condition improves.

If you have a spare $1K burning a hole in your pocket, you could try out a 928 pre-pump, looks like it screws right into the stock 944 fuel pickup hole, though it might interfere with the factory fuel level sender

https://www.fcpeuro.com/products/porsch ... 2343002001

The last Porsches with fuel return lines (996TT/997TT and GT2/GT3 with the "Mezger" engines) all had a dedicated fuel cooling circuit which used the A/C system, the cylindrical heat exchanger can be clearly seen in the engine bay.

---Chris A.

The symptom I noticed in a CIS car with the fuel heating issues was a actually a pressurized tank. First time it happened, the car stopped along the side of the road just south of Bishop Ca. in the summer. Fuel level was low, and we'd planned on filling in Bishop anyway. Casually opening the cap "just to see..." launched it 20+ ft. onto the sand. With the cap off and the pressure suddenly reduced, I could hear the fuel boiling with a rumble inside the tank. I found that cap and put it back on. It was about a hundred summer degrees that day, but the quick venting actually cooled the fuel in the tank some. I iced the CIS metering unit and the exposed fuel lines in the engine bay, jumpered the pump relay for a bit for circulation, and after ten minutes or so was able to get the engine going again. Made it into Bishop with AC off, filled the tank with the local higher-altitude blend, and on we went.

The Bosch "fix" for the problem was to offer a higher-pressure/flow pump, which tuned out to be the wrong way to go. Or maybe it just wasn't the right way, at least for my Saab turbo car. Friends with similar-vintage BMW cars were reporting the same issues, finding themselves stalled in L.A. traffic on hot days. The little in-tank pre-pump in the 928 is a small part of the big solution. The fuel cooler is the big part really. It's a coaxial exchanger with the evaporator suction/return line passing through the center, fuel passes in the outer shell and gets cooled by the "spent" refrigerant passing through the middle. It only works when the AC is working of course. Porsche fitted the little in-tank pump to the car on and off throughout the production run, but had the cooler very soon after production started.

After working on a few upper-performance Mercedes cars, I found that there are some very interesting options for complete in-tank pump solutions, ones that drop in to replace the original level sender too. Downside is that they also include the FPR and bypass internally rather than looping the fuel circuit through a FPR in the engine bay, so there's no fuel pressure adjustment with changes in manifold pressure. Knowing what I know, I'd be tempted to fit one of those packages inside the tank to pressurize the suction side of an external pump. A lower-duty external pump would do just fine, way better than an 044.. Remember that pumps are rated with total discharge head and flow, so adding the in-tank pre-pump would boost the flow and outlet pressure of that external pump significantly. You'd definitely still want the external fuel cooler.

If 944s and 928s without the inlet fuel pump use the same in-tank fuel filter then I suppose the inlet pump used on earlier 928s would fit in the 944 tank?

Lots to unpack above, thanks all.

When the Bosch 044 pump started making this noise, I was due for a smog check. The car passed no problem on a hot summer day with the pump screaming. That was on 91 octane pump gas by the way, not E85. I mention the smog check because part of the test is to check the fuel cap's seal. Removing the fuel cap that day did nothing to make the pump quieter -- it screamed all the way through the test, with cap on or off.... So I'm thinking the tank pressure (or vacuum) is an unlikely culprit.... Next time it screams, I'll confirm that though.

When the true Bosch 044 went bad, after a while, if memory serves, it basically screamed all the time. Once it starts making this noise, is it damaging the vanes inside or something, such that I will need a new pump now regardless?

The 928 in-tank pump sure looks like a fit. I might look on eBay just to see. Wonder how they do on E85? I'll have to mull fuel coolers. Surely doesn't help that Porsche routed the fuel lines directly over the headers.... I'm also going to take a closer look at the plumbing on the external pump. To get it to fit, I used a fairly tight aluminum elbow fitting, which might be contributing? This car at one point made 500rwhp on a dynojet (on 100 octane petrol fuel), so I hesitate trying a smaller pump. Even though I don't run that much boost/power regularly anymore, it's now on E85, so I'd hate to scrimp on the flow. Maybe the 928 in-tank pump with a less-than-044 external pump could get me back to 044 flow levels?

Tom --

The more serious issue will be cooling the fuel. Adding another "pre-pump" temporarily solves a symptom so might work, but longer-term solution will still be with cooling the fuel. Your E85 flow for the same power level is a bunch larger than the gasoline flow. I have no pump data on the 928 in-tank pump assembly for flow and pressure, so don't know if it will present another restriction in main pump suction rather than actually enhancing flow at full power levels on E85. The 928 discussions, unverified, are that it presents a few PSI at the main pump inlet at spec main pump flow.

928 WSM Pump delivery test is 1350cc in 30 secs @ 2.5 +/- 0.2 bar, test is for the two pumps running together. I don't have the performance curve for the matching 928 main pump and it's really irrelevant to this discussion with the flow test numbers available.

The fuel cooler in the 928 is as I described above. It's mounted in the firewall on the engine side, and has a threaded and o-ringed connection for the hard line from the expansion valve at the driver's side end. At the other end, it's threaded again with an o-ring fitting to accept the hard fitting on the end of the hard line that runs around towards the passenger fenderwall and forward to where the compressor suction hose connects. Both hard lines are aluminum.

The fuel connections on the cooler are small female threaded sleeves, brazed to the outer cooler shell. IMO they are too small for the volume of fuel flow you'll pass with a 500HP+ E85 demand, and would need to be upsized some to match your return line plumbing at minimum.

If you have a continuous hard line from expansion valve around to the fenderwall, above the existing fuel return connections somehow, my temptation would be to add the outer shell for the fuel around that existing AC suction pipe by welding it, and avoid cutting/sectioning that exiting AC pipe at all. It's a bit of tig work to join a couple sections of larger tubing as the outer shell, but that way you can make it longer so more heat transfer area. And get your favorite flavor of fuel fittings too. The 928 cooler has a wrap of closed-cell neoprene insulation around it, hangs in a couple P clamps from the firewall bulkhead. I can snap some pics tomorrow if my description wasn't clear enough.

A picture or three would be great. I looked on Summit, and see that there are any number of radiator style fuel coolers, mostly for diesel, but any port in a storm...? I may also just try re-routing the lines under the hood so they are not being baked by the headers (Porsche's idea, not mine, and super fun when they leak).

p.s., I've jumped on flimsier excuses to practice my TIG welding...

Here's an image from the WSM pages, handy because the air cleaner blocks an actual camera view on the car. I'll get some better pics from mine when it's in the workbay. This one should give at least a concept to consider.

Rather than trying to add AC fittings and build leaky connections for same, I'd be more tempted to find some larger aluminum tubing, section it lengthwise, roll the ends so they fit closer to the original aluminum line OD, and then TIG the pieces in place around the existing line from evaporator/expansion valve to compressor suction. In the picture there are fuel return line connections shown together at one end of the outer shell. I have no indication of what might be used for baffles in that outer shell, so might look harder at having fittings closer to the outer ends of the shell. With everything in aluminum, the outer shell will transfer heat from the fuel and conduct it to the AC tubing from the ends, in addition to the heat that transfers directly from fuel to the AC tubing inside the shell.

For sure you'll want to insulate the fuel lines and the cooler shell, and probably do any other exposed sections of compressor suction piping from the expansion valve/evaporator anyway.

On the 928 anyway, the fuel injector rails over the heads were originally fitted with a foam layer and a cover to help diminish injector noises transferred to the rails. At my last injector cleaning session, I fitted some more insulation so it wraps around to protect the bottom of those rails from radiated heat from the cylinder head. It's adhesive-faced closed-cell foam in my case, with holes punched for the injectors and retainer clips to poke through. Tiny cable ties secure it. It's hidden pretty well, and might be noticed only if you look hard from a low angle. All with the idea o reducing transferred heat into the rails. The rest of the under-hood metal fuel plumbing is still factory-exposed, so I'm not sure there's a lot of benefit. But I felt I was doing something even if I can't prove a benefit.

Again, I'm not familiar enough with the 951 fuel plumbing to know if this is a practical option for you.