zachu812
New member
Title says it all, I've had this idea for years but haven't had the time and resources to start on anything until recently. Go figure it's after I ppumped the truck, now it's more about seeing what these pumps are capable of and learning something along the way. I'm an electrical engineer by trade so I've got a good start on the electronics portion of things but my software skills are pretty rusty, gonna have a lot to figure out there. I'm open to ideas when I get to that point.
This all started when I took a VP44 apart a few months back and came to the conclusion that the pump itself is capable of moving a decent amount of fuel, the biggest problem it seems is the timing piston galling and locking up. Once this happens the PSG gets all confused and starts doing weird stuff then people blame it on the electronics. I've got some ideas to fix the timing piston problem, or at least to help it live, but that can wait until I get a pump running on a truck. I want to set the code up so the pump still runs fine with static timing, none of that dead pedal crap.
My plan to start is making an Arduino Due based standalone ECU that only references the internal pump position sensor and fires the fuel solenoid. Timing, map based fueling, and all that can wait until I've got something running. I started by pinning out the pump sensor, it has a ground, 5V supply from the PSG, and two outputs. I was hoping for a nice 0-5V logic level signal when I spin the pump but wasn't that lucky. The output almost looks like something from a variable reluctance sensor with some DC offset but the voltage doesn't seem to change with speed. At first I thought it was a differential signal but I think there's actually 2 sensors offset slightly. You can see each sensor output on the blue and yellow lines in the scope plot below, the red line is what happens when the wires are connected together. The higher points on the sensor outputs are actually dips on the pumps internal tone wheel, there are 6 of these evenly spaced so the plan is to use these as reference points to time the injection events.
Next step is to turn whatever the pump sensor is putting out into a usable 0-5V signal. I put together a circuit simulation with a couple op-amps and it seems to do exactly what I'm looking for. The red trace is a simulation of whats coming out of the sensor, green trace is after I remove the offset, and the blue trace is the 0-5V signal I'll be able to feed into the Arduino. Getting these results on an actual circuit might be another story but it's doable, I'll get started on that when I'm back from this business trip next week.
This all started when I took a VP44 apart a few months back and came to the conclusion that the pump itself is capable of moving a decent amount of fuel, the biggest problem it seems is the timing piston galling and locking up. Once this happens the PSG gets all confused and starts doing weird stuff then people blame it on the electronics. I've got some ideas to fix the timing piston problem, or at least to help it live, but that can wait until I get a pump running on a truck. I want to set the code up so the pump still runs fine with static timing, none of that dead pedal crap.
My plan to start is making an Arduino Due based standalone ECU that only references the internal pump position sensor and fires the fuel solenoid. Timing, map based fueling, and all that can wait until I've got something running. I started by pinning out the pump sensor, it has a ground, 5V supply from the PSG, and two outputs. I was hoping for a nice 0-5V logic level signal when I spin the pump but wasn't that lucky. The output almost looks like something from a variable reluctance sensor with some DC offset but the voltage doesn't seem to change with speed. At first I thought it was a differential signal but I think there's actually 2 sensors offset slightly. You can see each sensor output on the blue and yellow lines in the scope plot below, the red line is what happens when the wires are connected together. The higher points on the sensor outputs are actually dips on the pumps internal tone wheel, there are 6 of these evenly spaced so the plan is to use these as reference points to time the injection events.
Next step is to turn whatever the pump sensor is putting out into a usable 0-5V signal. I put together a circuit simulation with a couple op-amps and it seems to do exactly what I'm looking for. The red trace is a simulation of whats coming out of the sensor, green trace is after I remove the offset, and the blue trace is the 0-5V signal I'll be able to feed into the Arduino. Getting these results on an actual circuit might be another story but it's doable, I'll get started on that when I'm back from this business trip next week.