The regulator we’ve found so far that offers the best regulation is the Turbosmart FPR2000. It outperforms all the others tested by a factor of between two and four times.
One of the basic functions of an ECU is getting the right amount of fuel into the engine. Since 2009 we’ve been doing this using volumetric efficiency tuning, which requires the ECU to have a decent model of the injector. When we first implemented this model, we expected that people would have such information about their injectors. This was an errant assumption.
Even when people use Injector Dynamics injectors, which have excellent published data, some customers found that they needed to put large numbers into the VE table, especially at light loads, which would indicate that the offset is too low. So we set out to investigate this effect.
Part of this investigation involves building a test system to test the injectors themselves. We started off with a commercial injector test machine, but this one (I suspect this holds true for all of them) was useless for testing modern performance injectors with a high flow rate. Firstly the pump was completely inadequare but secondly the pressure regulation was very poor.
So I went and bought a high flow pump and a pressure regulator that I was told would suit that flow rate.
At first I was surprised because the pressure gauge on the reg didn’t match the 0-150 PSI pressure sensor I was using; I later discovered (by comparing with an accurate pressure gauge) that this gauge was just really inaccurate, and the pressure sensor and the good gauge agreed to within 1 PSI.
My second surprise was how much the fuel pressure changed just when we started to pulse injectors. I thought the whole point of a pressure regulator was to keep the fuel pressure constant, and while nothing is pefect, I wasn’t expecting such a big difference in the fuel pressure, especially on a fuel pressure reg designed for such a high flowing system.
I also discovered that by pulsing the injectors at different rates, you could excite the diaphragm in the pressure reg so that the fuel pressure actually oscillated. During this time, I could see the oscillation on the good pressure gauge, and on the pressure sensor, but the (liquid filled) pressure gauge on the regulator stayed steady. I know that the pressure was actually changing because instead of the injector response being nice and linear once it opened, it had a wavy characteristic that you’d expect of resonances affecting fuel pressure:
ID 725 graph of fuel delivered vs on-time
It occurred to me that people might have this kind of problem on their car, but are tuning around it, thinking “that’s funny, the engine wants 15% more fuel at this condition but I don’t see a bump in the torque curve there… ah well”. Looking at the (liquid filled) gauge that came with the pressure regulator, you wouldn’t know that the pressure was jumping all over the place.
The first problem of the inaccurate pressure gauge I can live with; we can just advise our customers not to use that pressure gauge. However the second problem of the pressure varying with injector flow is a worry, because it means that people’s fuel pressure won’t be stable, and they’ll be adding in higher numbers into the VE map at higher loads, not because the engine needs it, but because they’re compensating for a bad pressure regulator.
I tried to find data on how well the pressure regulators work, but no manufacturer seemed to publish it. The exception was Fuel Lab, who did publish the data in the instructions for the reg.
So I bought several pressure regulators to test, and also borrowed a few.
All of the aftermarket pressure regulators have two pressure ports and one return port. Since the restriction is caused by the fuel flowing through the orifice, I closed up one pressure port, and varied the flow rate of the fuel into the reg, and measured the pressure that the reg developed. The flow rate was varied using a variable power supply on the fuel pump. Test fluid was E85.
I’m not going to say the results of each unit, because I don’t want to make any enemies, but I will show the graphs from various units and which one was the best and the worst.
Pressure vs flow rate of various pressure regulators
The perfect fuel pressure regulator would be a flat line, indicating that the pressure is completely constant against the flow rate. The Turbosmart FPR2000 is the closeset to this flat line. The worst is actually the purple line, which is the factory fuel pressure regulator off the MAzda RX7 FD. See that firstly, the slope is pretty steep in its linear section. In fact it’s four times that of the FPR2000 (dark blue line). Secondly, above about 4 L/minute (240 L/hr), the pressure rises very sharply. This means that if you’re going to put a fuel pump in your RX7 above 240 L/hr, you better upgrade the pressure regulator also.
The following shows the linear parts of the graph with the gradients at the right hand side (click the image to see the full story):
Pressure regulator tests with gradients over their linear sections.
You can see from this that the Turbosmart FPR2000 is approximately twice as good as the next contender (0.0025 compared to 0.0048), and four times as good as the factory FD regulator.
The units under test were:
This wasn’t intended to be a comparison of different brands, manufacturers or models; it was merely supposed to be research into how these units behave and what our customers are dealing with on their installations. However from the testing there emerged a clear winner!
Liquid filled gauges won’t tell you about any pressure fluctuations; and also some of them seem to be really inaccurate, so use a proper fuel pressure sensor and log it through the ECU; looking at a gauge and telling us the fuel pressure is stable doesn’t have much value now with what we’ve just learnt and observed.
The main point is that you should measure the behaviour of the system; although you may just be putting together parts that appear to be commodity items (like fuel pressure regulators), the overall system still needs to be verified. The ECU expects the fuel pressure to be stable to deliver a consistent amount of fuel to the engine.
Now having fixed the pressure pulsation problem on the injector test system, we now have nice linear injector graphs and can properly start measuring flow rates and dead times:
ID725 performance at 3 different voltages. Vertical axis is fuel delivered per pulse (µL), horizontal axis is pulse duration (ms), and the 3 lines correspond different battery voltages (8.0V, 8.8V, 9.6V)
Big thanks to Paul Yaw from Injector Dynamics who helped me understand what the system is doing and how to measure injectors accurately.