OK, game over. Thanks for playing
We discussed this in Group Therapy last night too.
A lot of good ideas and many of you are on the right track. Loren (Maki) nailed it with the most complete and accurate response. He said it better than I could, but I'll blab anyway...
1. The asynchronous pulsing of the injector is a normal injection strategy. It can do this on accel or trying to catch a stall. It's just a distraction and can be dismissed for now. There are much more important elements in the capture that tell the story.
2. The lifting off ground of the reference (chan B) is a circuit voltage drop caused by the coil current. It's normal. At time 4420ms you can see clearly a coil ramp and fire in the reference voltage drop. Zoom in 2x vertical on channel B to see better. You see the induction ramp, current limit and cut off.
3. The voltage is present on channel A injector drive (neg side) when the injector is not active so the circuit is not open. However, during the event, when the injector is grounded, no current flows. So, we know there is excessive resistance somewhere. We have a bad connection.
4. The distance between the lowest point in the injector voltage waveform on channel A and ground is the circuit voltage drop from the test point at the neg side of the connector through the PCM driver, the PCM grounds and all the wiring to bat neg. Just like using your DVOM to do voltage drop testing on a starter circuit, the circuit must be active to see the voltage drop. If the bad connection was anywhere between the channel A test point and ground (PCM driver, PCM grounds, etc) then we would see that in the voltage drop. We don't. The injector is flat on ground the whole time. Therefore, we know that the bad connection is not anywhere after the channel A test point toward the PCM.
5. Channel C is on the power side B+ of the same connector. It shows no drop in voltage during the events. If the bad connection was toward the power supply from the channel C test point, we would see the voltage fall. We do see this voltage dip when current flows in the circuit. It's minimal and to be expected. We can even see an upside down injector current signature there if you look closely. But not when there is no current flowing.
6. Conclusion: The bad connection is between the channel A and C test points. Since all that is under the upper intake plenum, it's time to get out the wrenches. We know what we are looking for and where to find it.
Waveform databases are great, but we can't do lab scope diagnosis by simply comparing waveforms. We have to be able to analyze, using the system knowledge and the behavior of circuits. I always say, slow down you'll be faster. Don't rush off to do another hook up until you fully understand what you have captured already. You may not need any other captures. The answers to all your questions may already be right in front of you.