So I had lots of very hot weather to test the VP44 project. On a 103 degree day the controller hit 175 sitting in traffic with AC on and 120s while moving along. No problems it never missed a beat I'm happy about that I was a bit worried.
I got one of those piezoelectric sensors that clamps around the injection line and had to build something called a charge amplifier and insulate it from ground and use a shielded cable to get rid of all the noise so I could see anything reasonable.
Timing wise things didn't add up. I finally got frustrated and pulled #1 injector and used a dial pointer to mark TDC and made a pointer mounted to an oil pan bolt. Next I connected a meter to the cam sensor and crank sensor and very slowly turned the crank and confirmed my timing waveform (I drew it on a piece of tape on the balancer😏) with what I had with it running and found a 10 degree error.
That explains why my timing just wasn't adding up. It was actually close by coincidence. What I found was that engine has too rotate about 10 degrees from the moment the inj valve seats because the cam donut has a very gentle ramp at first, the plungers have to move about 10 degrees (crank) to reach the pop off pressure of the injector. I can see the pop off with the piezoelectric sensor.
Then there is the line delay from pump to injector it's how long the pulse takes to get there it's based on the speed of sound through diesel which is 1357 meters/second. The lines are 32 inches long. That's 0.81 meters.
So that's 1/1357 = .737 mS to go 1 meter X 0.81 equals .597 mS delay.
At 1000 rpm the crank rotates 6000 degrees in 1 second so 6 degrees every millisecond. So 6 X .597 = 3.6 degrees at 1000 rpm that same delay equals 10.8 deg at 3000 rpm.
I have adjusted the way the controller calculates timing now. I suspect that the line delay, which is determined by line length is sent to the vp from the ECM in one of the CAN bus messages. But my ECM is toast.
BTW, I figured out another pair of bytes in one of the CAN messages from the pump. It's the delay in microseconds from the time the injection solenoid current is shut off until the time fuel actually stops injecting. It's runs around 760 uS. On mine.
So I had lots of very hot weather to test the VP44 project. On a 103 degree day the controller hit 175 sitting in traffic with AC on and 120s while moving along. No problems it never missed a beat I'm happy about that I was a bit worried.
I got one of those piezoelectric sensors that clamps around the injection line and had to build something called a charge amplifier and insulate it from ground and use a shielded cable to get rid of all the noise so I could see anything reasonable.
Timing wise things didn't add up. I finally got frustrated and pulled #1 injector and used a dial pointer to mark TDC and made a pointer mounted to an oil pan bolt. Next I connected a meter to the cam sensor and crank sensor and very slowly turned the crank and confirmed my timing waveform (I drew it on a piece of tape on the balancer😏) with what I had with it running and found a 10 degree error.
That explains why my timing just wasn't adding up. It was actually close by coincidence. What I found was that engine has too rotate about 10 degrees from the moment the inj valve seats because the cam donut has a very gentle ramp at first, the plungers have to move about 10 degrees (crank) to reach the pop off pressure of the injector. I can see the pop off with the piezoelectric sensor.
Then there is the line delay from pump to injector it's how long the pulse takes to get there it's based on the speed of sound through diesel which is 1357 meters/second. The lines are 32 inches long. That's 0.81 meters.
So that's 1/1357 = .737 mS to go 1 meter X 0.81 equals .597 mS delay.
At 1000 rpm the crank rotates 6000 degrees in 1 second so 6 degrees every millisecond. So 6 X .597 = 3.6 degrees at 1000 rpm that same delay equals 10.8 deg at 3000 rpm.
I have adjusted the way the controller calculates timing now. I suspect that the line delay, which is determined by line length is sent to the vp from the ECM in one of the CAN bus messages. But my ECM is toast.
BTW, I figured out another pair of bytes in one of the CAN messages from the pump. It's the delay in microseconds from the time the injection solenoid current is shut off until the time fuel actually stops injecting. It's runs around 760 uS. On mine.
Edited by Great work!
More info