injector pulse width to verify MAF/AFM operation

[picowight]

Working on an older car and I took a reading of the (petrol) injectors -peak and hold type, looking for a misfire, not going to get a chance to get back on it for a while and in the meantime I was thinking about using the maths channel to plot the pulse width of the injectors - ignoring the frequency as this rises with RPM but the pulse width should change in relation to O2 sensor and air mass.. I think
At a steady rpm unloaded this should reflect the oscillations of the O2 sensor, but I was thinking if there was a glitch in the MAF/AFM this would show up in the curve. What would be the best way to do this, can I do it on 7 or should I go back to 6 ?
Thanks!

[Steve Smith]

Hello and thank you for the post

This is an interesting topic and one we have discussed in the past for port injectors

Graphing the negative pulse width with math's is challenging due to the inductive spike of these injectors

Applying a heavy low pass filter will remove the inductive voltage spike but also have a negative influence on the pulse width. The same applies when applying resolution enhancement to 16 bit

What we need is a threshold within math's whereby calculations are only based on values above or below the threshold (using unfiltered data)

We have this feature with Deep Measure within PS6 topic17711.html whereby we specify the threshold of interest

Below I have graphed the negative pulse width of a post injector at idle speed using Excel

Image 1

Here is the psdata file showing the negative pulse threshold

Image 2

I will pursue using math's as an alternative for you in the interim

I hope this helps, take care.....Steve

[picowight]

Thanks Steve.
I did miss an important part, I measured the current of the injectors - the actual injectors were behind a 30 year old plastic cover on an iconic (read very expensive) car, so I had to content with hooking round the wire in the loom as I was not risking removal.
I am guessing that the threshold is the same issue though. I can see here visually for example that the middle pulse is much shorter - each pulse is 2/4 injectors as it group fires.
I guess if there was an upper/lower value on the existing duty cycle calc then that would do it.

[Steve Smith]

Hello and thank you for the feedback

No worries re the positive pulse (injector current) the same should still apply when using Deep Measure where we graph the high pulse width instead of the low.

Thanks to Martyn at Pico Support, we have the following math channel in which to graph the negative pulse width of Port Injectors:

For PicoScope 7, we have negduty(sign(A-10)) for 12 V port injector switching voltage captured on channel A (Change the letter A) for your chosen channel

For PicoScope 6 we have duty(sign(A-10)(-100)) for 12 V port injector switching voltage captured on channel A (Change the letter A) for your chosen channel

To avoid confusion (and I will here) we can use duty(sign(A-10)(-100)) for both PS6 and PS7
Below we have the above math channel graphing injector pulse width in response to our narrow band O2 sensor (System is closed loop)

Image 1

Please note, increasing the scale and offset of the math channel is required to reveal the sine formation above as our negative duty is very low at approx. 2% (idle speed)

Referring to the above we can see closed loop fueling in operation as we cross the threshold voltage of 450 mV (Stochiometric 14.7:1)

Note how during the rich “cycle” (above 450 mV) injector pulse width duty is reduced

Likewise, during the lean “cycle” (below 450 mV) injector pulse width duty is increased

Below we have the same scenario at increased engine speed where our negative duty has increased to approx. 8.9% based on the demand to keep the engine running at a higher rpm.

Image 2

Note closed loop fueling still applies above as we can see in the increase and decrease in negative duty, just at a higher percentage

I hope this helps, take care…..Steve