Window clearance, but not an index issue (I believe)

[MPRCTCL]

My passenger window does not drop the same distance each time. More often then not, it drops too little to clear the roof bezel. I have re-indexed it several times, and I am convinced that is not the problem. See the photos below:



The above photo shows tape placed on the window in its fully-up position. Spoiler alert, the three lines show the three different levels to which it drops with the purple one being the only one that clears the roof.



8 out of 10 times, it drops this ^ far and hangs up on the roof trim. Occasionally, like 10% of the time, it drops to the other, not purple line which is really does not clear.



Once in about ten tries, it will move to this ^ position. This is the only position that clears the roof. When it doesn't clear, and you force it past the seal, you can't close the door. Let me re-phrase that... my unassuming passenger is very displeased and unimpressed when the door will not close.

Again, I have re-indexed it according to the service manual procedure. That makes slight tweaks, but does not change the inconsistency of where it stops.

Here are some more bits, but don't let them over-influence your opinion because some may be red-hearings and others I may not have noticed yet.

• My roof trim is not perfectly aligned. See below. That said, I'm inclined the think the inconsistency in how far the window indexes is the problem. If anything, the back section is higher than it should be and the one the window stick behind is "normal."

• I removed the interior door panel. That does not change anything. I was thinking that the inside window welt was dragging, (which it may be) but it is not attached to the panel, but to the door itself.
• I had also considered the known problem with door module relays, but in my own mind, I don't see a way that a marginal relay would alter how far the window indexes each time.
• If you jiggle the window as it is going down, it may go farther. By that time, the door solenoid has re-engaged, so you have to trip the door handle again.

Okay... What are you thinking?

GO!

 

[MPRCTCL]

Here's an update, but I hope you won't assume it is solved and not give your input. I did get it working today. I focused on the uneven bezel on the roof, above the window glass. I got thinking, "I wonder if how the bad side compares with the driver's side." There is a 2mm difference between the driver seal width and the passenger side seal width. Coincidentally, there is about a 2mm difference between successfully and unsuccessfully opening the door. I partially opened the top to see if there is any adjustment, and there does not seem to be.

With some trepidation, I grabbed my GM alignment jig. I hesitated because it the tool calibration certification was slightly expired, so I proceeded slowly. With 22 applications of the tool, the seal receded 3mm, and the door operated properly.

All this said, I have not yet reinstalled the door panel and I am operating in a temperature-controlled environment, so this may be a temporary masking of the original problem. If you have any experience with this issue or have solid ideas, please do give me a reply.

Here is the required tool:
1 lb Non-Marring Mallet (Blue-Point(R))