Stuff

The low pressure side of the commonrail fuel system often gets overlooked, but plays a crucial role. I was pleased to see BMW specs nominal CP3 input pressure at ~5 bar. They probably used a fairly good quality lift pump in the tank which will likely never need replaced, assuming you ascribe to the fuel lubricity additive concept. Fuel supply flow capacity needs to increase proprtionally with power-adding mods. That does not appear to be an issue with this setup. My lift pump holds a rock-steady 60psi with the go pedal matted and JBD at 100%. But there are other things to consider. Cold weather for example, induces fuel viscosity increase, severely reducing pumpability once temps drop near zero and below. That’s why we have a fuel heater, but it can only do so much. Anything constraining fuel supply flow can quickly lead to injection pump damage. Unfortunately there’s no shortage of different ways to get things FUBAR’d on the upstream side of the CP3: cold weather, bad fuel (water especially), filter degradation, various lift pump malfunctions, etc.

The failure mode varies, but sudden catastrophic loss of the CP3 can take out everything in it’s path. Many parts are subject to damage, potentially requiring replacement of injectors, sensors, and at a minimum, a complete fuel system flush from head to tank. Dealers confronted with it often elect to simply replace the entire fuel system. It means lots of parts and labor hours, but ironically, ends up being quicker and easier in the long run if one thing after another starts turning up bad during a piecemeal approach. But this is all easily avoidable with a little monitoring. I understand the car’s sensors and software will set codes for a plethora of fueling paramter anomalies. But I doubt it’s ability to forewarn impending problems or long-term degraded conditions in the fuel supply like a simple, mechanical guage.

There are a couple of fairly serious issues involved with using a mechanical fuel gauge in the cabin. There’s a small chance it will fail in some way as to allow a leak. A worst case scenario might be something like it explodes spraying fuel in your face while driving in heavy traffic smoking a cigarette. My customized combination shutoff valve/snubber mitigates this issue and another problem fairly well. The other problem is CP3-generated pressure wave reflections wreaking havoc with the gauge readout. Installing a ball-bearing poppet in the valve’s gauge port stifles that phenomenon. The rubber valve seal gets soaked in gear oil while doing up the valve assembly to induce swelling to the point of actually making the thumbscrew harder to turn after reassembly. Done right and reassembled using JB Weld for thread locker, the valve will never leak and just cracked open provides a safe, stable analog/mechanical fuel pressure reference to the gauge.

Tapping the fuel line post-filter is not exactly straightforward. There are lots of candidates for places to do that, but BMW’s proprietary faux plastic lines and fittings everywhere take all the fun out of it. A good way I found was to slide the stock filter back a couple inches in the mounting clamp and clock it about 15 degrees. This creates enough room to splice in the adapter quite nicely. Remember you are working on the clean side of the filter, so parts and activity related to opening it downstream of the filter must be meticulously maintained in pristine cleanliness. If not, the car will immediately tell you about your failure to do so upon restart.

I expect fuel filter maintenance intervals to increase significantly with the ability to actively monitor fuel pressure. Filters filter better the longer they are in service, right up to the point where they can no longer meet the flow requirement. I believe changing filters of almost any type on a schedule is just dumb. Of course there are specific exceptions to the rule, but they are limited to special types of equipment and conditions not in the consumer automotive realm.

Dumpsters behind quick lube joints across the country are always full of perfectly good filters. Also, it should be interesting to see pressure fluctuation (or not?) with a hotter tune at some point. I have a feeling fuel supply is not going to be an issue. The 9mm ID tubing should be adequate for up to around 400hp.

The w/m injection nozzle coupling worked it’s way loose somehow. I discovered it after dumping about 3 gallons of juice on the road between the time that started and when I identified the problem. So while I was at it I decided to upgrade it to a 2-stage system. The trucks use progressive controllers to modulate delivery, but I really doubt the efficacy of those devices in normal use. The main purpose for having it on the “D” is keeping the intake clean. It can also add a little power. I theorized a more practical approach would be two injection events, a smaller nozzle coming on earlier more often, and a bigger nozzle coming in only at or near full load.

1st phase of the project was installing a 2nd smaller injector, size M2. The original M3 stays put in the intercooler coupler, after of course modifying the connection to keep it secure. I suppose those push-in 1/4″ pipe couplers are OK, because I never had this problem until now, after years of using them on the trucks and continuing to use them at various locations in this system as well. A hose barb reinforced with JB Weld is not coming apart without a forcefully applied tool. Here’s the new nozzle mounted right at the manifold flange:

Phase 2 involved setting up a solenoid valve to fire the original M3 injector off a secondary pressure switch set for the max boost range starting ~24psi. I began by testing with the spare pump to ensure it would effectively atomize 2 injectors, because it’s a pretty small pump. No worries there. Setting the boost switches accurately was a frustrating moving target at first. The engine will see somewhere around a 3psi delta on the high side parameter depending on a number of factors. 22psi is where it typically tops out, but I’ve seen it pull 25 briefly under heavy load.

I used 1-amp fuses on the pump and solenoid circuits to dial it in pretty close to my target settings. Those switches from McMaster Carr are quite sensitive and accurate. I started with them set high and repeatedly ran the car to the desired boost setting, stopped and checked the fuse until it blew. Took three tries on the high switch and about 6 or 7 on the low switch. I haven’t run it with a boost gauge installed since removing Burger’s JBD tuner last summer, but I imagine it should go higher with that extra fuel.

After a couple weeks of using too much water while randomly fiddling with the switch settings I ended up making a tool for pressurizing the system manually to see exactly where the switches were tripping. It probably would be better to tap into the MAF sensor for this purpose, but I’m still a little skittish about the electronics on this high-tech Teutonic contraption.

The pump supply line tees off to the primary injector before the secondary solenoid valve:

The last phase was simply getting everything wired up, plumbed and tested: Pressure Switches:

The air, or “boost side” of the system literally comes together on the driver’s side shock tower. Both switches and the gauge reference the same charge pipe feed at this junction:

The car has a Boost Gauge mounted in the instrument cluster. It’s more of a testing tool that anything else, necessary to fine tune the pressure switches. I used the high-pressure gauge on my compressor to set the switch before, but it’s not good enough at lower pressure to see the 1-psi increments needed for this application with any accuracy.

Compared to the single-nozzle version which came on at 20psi, the system now provides >60% more water/meth on the top end, while delivering about 30% less at moderate load between 18 and 24psi boost. I expect the real-life consumption rate to go up spraying less more often, based on my driving habits. Before I was using about about 1gal/tank of fuel, really only hitting it with go pedal down hard. I’m hoping this new setup equates to better cleaning in normal driving.