Team Rose Motorsports' Porsche 914 has an turbocharged rotary engine from a third generation Mazda RX-7 (Series 5 13B), making for an extremely fast and engaging car to drive. The engine conversion was performed entirely by TRM students like me who learned through the experience of building and failing then rebuilding.
ECU Tuning
The Porsche runs a Microtech LT-9c chosen for low cost. There is very little manufacturer documentation for this ECU, so I become familiar with the RX-7 internet forums to learn details of its operation. It uses serial communication, so even connecting it to my laptop was a challenge. I first configured a list of items on the ECU: lambda input, fan control, scavenge pump control, and others. The engine ran poorly because a number of sensors were not calibrated, so I fixed those too. I tuned fueling at low load and idle so the team could break in the rebuilt engine.
ECU Tuning
The Porsche runs a Microtech LT-9c chosen for low cost. There is very little manufacturer documentation for this ECU, so I become familiar with the RX-7 internet forums to learn details of its operation. It uses serial communication, so even connecting it to my laptop was a challenge. I first configured a list of items on the ECU: lambda input, fan control, scavenge pump control, and others. The engine ran poorly because a number of sensors were not calibrated, so I fixed those too. I tuned fueling at low load and idle so the team could break in the rebuilt engine.
Once the club finally completed breaking in the engine, it was time to turn up the boost. While driving, the engine would stumble and break up right at the transition from vacuum to boost pressure. The engine would not spin above 3500 rpm under any load and it caused a violent sensation similar to an aggressive ignition cut. Unfortunately no engine cuts or warnings (easy answers) were active when I reviewed the data logs. Logging is another issue with this ECU. Data sampling is locked at 2 Hz which is not fast enough to diagnose transient response issues. Data logs are mere text files, making me pour over them line by line. At this time I had not learned enough to write software to graph these parameters, I would do that now that I know.
I chased the problem through engine cuts, spark and fuel acceleration enrichment strategies, and many more dead ends. I eventually spotted the cause: the injector output postfix changes from sT to mS at positive pressure. After days of digging I learned these characters signal the ECU firing the secondary fuel injectors. The Porsche's secondaries flowed 3x more fuel than the primaries which immediately flooded the chamber and caused the stumble felt when driving.
This realization prompted an even deeper dive through documentation on staged injection and Microtech's fueling strategies. This ECU has no gradual changover; secondary injectors are either on or off meaning the violent transition must be solved in the main fuel table. I reverse engineered the ECU's secondary injector open time calculation to figure out how to smooth out fuel delivery between the mismatched injectors. Below is my work graphing fuel flow through each injector set. Armed with my new knowledge I was able to change a few operating parameters to eliminate the massive jump in fuel delivery (blue line on right chart).
This realization prompted an even deeper dive through documentation on staged injection and Microtech's fueling strategies. This ECU has no gradual changover; secondary injectors are either on or off meaning the violent transition must be solved in the main fuel table. I reverse engineered the ECU's secondary injector open time calculation to figure out how to smooth out fuel delivery between the mismatched injectors. Below is my work graphing fuel flow through each injector set. Armed with my new knowledge I was able to change a few operating parameters to eliminate the massive jump in fuel delivery (blue line on right chart).
I planned on implement my idea and fix the issue in spring of 2020. Unfortunately I never had the chance to take the Porsche out to the testing grounds and test my hypothesis. I passed my ideas on to younger members of the team so they can fix the issue in the future.
Wiring / Electrical System
Team members who originally put together the car did enough to get the engine running. In the years between their work and me joining the team everyone originally involved graduated and took their knowledge with them, and others later appended or removed circuits as equipment changed. Since it was a rotary it ran unreliably, but diagnosing issues in the Porsche's existing fuse box and wiring harness was difficult at best.
Team members who originally put together the car did enough to get the engine running. In the years between their work and me joining the team everyone originally involved graduated and took their knowledge with them, and others later appended or removed circuits as equipment changed. Since it was a rotary it ran unreliably, but diagnosing issues in the Porsche's existing fuse box and wiring harness was difficult at best.
I started by tracing each and every wire to understand each function the wiring supported. With reverse engineering complete I split each device into circuits based on fuses and relays if needed. My goals for the new wiring were to fuse every circuit, eliminate high current flowing through the switch panel on the dash, and complete documentation. During this process I learned extensively about automotive wiring and how the ECU controls devices through the minimal manufacturer's documentation and internet research.
The car is actually very simple and only needed 7 circuits switched either by the driver or ECU logic, or a combination of the two. The fuel pump has a manual shut-off and the fans have a manual override on the switch panel, but are normally controlled by the ECU. Over spring break my sophomore year I built this fuse box from scratch to incorporate everything I learned, power every feature of the Porsche and leave room for future expansion. I added a wideband O2 sensor to help tune the engine, moved the radiator fans to the ECU's output (were manual before), and included a scavenge pump for the relatively low mounted turbo oil drain. Not pictured is the main power relay activated by the shutoff switch in the dash. Every wire is labeled with it's function, and I left detailed diagrams and information with the team.
The car is actually very simple and only needed 7 circuits switched either by the driver or ECU logic, or a combination of the two. The fuel pump has a manual shut-off and the fans have a manual override on the switch panel, but are normally controlled by the ECU. Over spring break my sophomore year I built this fuse box from scratch to incorporate everything I learned, power every feature of the Porsche and leave room for future expansion. I added a wideband O2 sensor to help tune the engine, moved the radiator fans to the ECU's output (were manual before), and included a scavenge pump for the relatively low mounted turbo oil drain. Not pictured is the main power relay activated by the shutoff switch in the dash. Every wire is labeled with it's function, and I left detailed diagrams and information with the team.