Injector measurement
I was interested in injector opening/closing time, flow regarding to flyback setups.
NEWS 2007.02.05
Previous measurements were obviously inaccurate, so I started over with better equipment:
It's now the "what we've expected" chart, linear, with some degradation in flow at low pw values.
I started with the smallest possible pulse (which actually made the injector open). Holley is low imp, so it's faster to open/close. It needed almost 50% pwm duty at 11.5V battery, otherwise it couldn't stay open.
Still seeking the answer to the question: Why we couldn't get stable idle with lower than 3ms with the Accel inj above? (And why it's using 90% duty to feed ~300 horses on a 6cyl/6inj engine?)
2007.01.26
I did some measurements with setting cranking RPM range high. Basically it worked fine, but two problems arised:
1. Pulsewidth simply disappears (goes to zero) after some amount of time. Time's between ~40-120 seconds of "cranking". Sometimes it doesn't need to have RPM signal at all just change settings in megatune, and pw on LCD/megatune/in_reality changes to 0000. Switching off/on resolves the problem for the next ~100 seconds.
2. Above 20ms of cranking pw, scale changes to 0.5 msec (so 20.0 + 55*0.5 = 47.5 msec can be specified at max: rarely needed, but it is by design). Above 25.5 msec, megatune displays pw - 25.6 msec.
About the results:
I'm not sure whether this is normal or not. I expected a linear growth of flow with pulsewidth (at least in the higher range).
- Results are certainly somewhat wrong. Did you measure actual pulse width. Measuring opening time with oscilloscope or graphical multimeter etc. would clarify measurements. //BengtR
All of the above pulsewidth values includes all timing (I have set all open times to 0)
2007.01.22
I've put together a flowbench-setup
First tests showed that injector opening time increased slightly according to the fuel pressure introduced (3 bars). No flow measurement data yet.
Question: How can I force GenBoard to a specific pulsewidth?
-Tried "mxofemxofemxp05mxp05" and similar, but no effect at all.
I would use a soundcard ElectronicDesign/TriggerSignalGenerator with 60-2 or similar pattern, and use "cranking RPM" (both low-temp and high-temp values set to same pulsewidth). you can bump up cranking_thresh, but best to apply 290 RPM or similar. Take care (and have multiple fire extinguishers at hand). -It good-and-working idea, thanks.
- You might want to measure (if measuring 'acutal' flow rate) at around 6000 1/min (ie. 100 Hz) or similar (depending your engine maximum revolutions etc.) and around 80-85 % opening time (ie. 8,0-8,5 ms at 100 Hz). //BengtR
- Good idea, I'll check it later, first I'm interested in the idle-area.
All of the below measurements made on a 12 Ohm Hi-Z injector dry -without fuel pressure-, and PowerFlyback installed inside the ECU. No PWM-ing was enabled, pulse width is ~14ms.
Voltage was 12.1-12.2V, voltage/current measurements were made after a 0.3Ohms resistance, so test setup is like:
+12.2V____[0.3Ohms]____*Measurement*____[injector]____INJ-OUT
The results:
Red channel means measurement voltage (or current).
Blue channel means mechanical movement/noise.
- 1. graph is one open-hold-close pulse with no external flyback. Closing time is 1.7ms
- 2. graph is similar open-hold-close pulse with 18V Zener flyback. Closing time is ~0.95ms
- 3. graph is similar open-hold-close pulse with 30V transient diode flyback. Closing time is 0.9ms
It's visible on the red channel that after opening (about 1ms later) current has a pulsation which shows the movement of injector iron core. At the same time mechanical noise started (blue channel).
At closing phase there's not visible current-pulse in red channel but a spike goes well over starting (+12V) voltage which is the inductive "kick-back" or flyback current. Blue channel's mechanical noise shows here the exact position of mechanical injector closing.
As my movement/noise detection (knock sensor) was nearly installed on the injector it's coil slightly picked up injector current, so on the blue channel it's also visible when injector current switched on/off (a single spike before noise).