September 2023
Successful Full Power Fuel Pump Tests
Over the past ~4 weeks I've been testing and iterating the fuel pump for the bandit engine. It was a challenge to reach full power due to some nuances with the power system and the motor efficiency that I cover in the video. I ran 22 tests over the course of a couple weekends, and generally took the rest of each week to analyze data, and implement design changes between classes and such. These tests from 9/30/23 (tests 6.1 and 6.2) actually reached about 125% of the minimum pressure requirement, leaving some margin which is good. Next steps for the fuel pump are to finalize and order the sheet metal brackets to replace the plastic parts and also to fix some pinhole leaks on the welded fittings etc.
I had struggled to reach full power for a couple of test days after my first tests at the beginning of the month. The first cause for this struggle was due to the internal resistance of my batteries. I was able to Identify this problem using data from a new programmable VESC I bought to support testing at higher wattage. The data showed large drops in battery terminal voltage when under high current load which is something I overlooked when sizing the voltage, quantity, and c rating of my batteries. The solution was a combination of adding additional cells in series to increase voltage to compensate for the drop, and also to use batteries with a higher c rating (lower internal resistance). Both these solutions can be seen when comparing test 3.2 to the full power test 6.1.
The second problem was basically an incompatibility between the pump and motor operating points. Since the pump is a positive displacement gear pump, the speed of the pump is set by the desired flow rate. For example, to achieve the flow rate for bandit, I calculated that the pump shaft must spin at 2411 RPM. In the first iteration, the motor shaft was directly connected to the pump shaft meaning that it would have to also spin at 2411 RPM. However, the vendor performance charts show that maximum mechanical output of the motor occurs between 3200 and 4000 RPM while at 2400RPM, the motor efficiency is a relatively low 30% and the output at 36V is listed at only about 1000W which is insufficient after considering all the other losses in the system. Additionally, when running the motor at a lower speed, more torque is needed to achieve the same power and current is proportional to torque so very high current is pulled from the batteries which exacerbates the problem with the internal resistance and also heats up the motor more quickly which is bad. My solution to this problem was to implement a gear reduction between the shafts so that the motor could run at a higher speed where it could output more power at a lower torque while the pump could stay at 2400RPM. The ratio I sized was a range around 1.5-1.6. I found a pair of cheap, steel helical gears on amazon and designed a new bracket so they could mesh properly.
In the end, both solutions contributed to greatly increase the power imparted to the kerosene and even exceeded the minimum pressure and flow requirements for the engine by an impressive margin. I was very excited to see the discharge pressure exceed 300psia with an even lower current draw than previous tests that only produced 200psia at the outlet. One drawback of the helical gears though, is that I noticed they produced a thrust force on the motor shaft, and in the videos you can see the motor shift downwards, but no permanent damage was done so I don’t know if it matters.
