I gave up for the moment although I have enlisted the help of a variety of Microsoft resources including the support channel and others in the product team to answer the central question of how to correctly display a PDF stored in OneDrive in the PDFViewer control.
I developed a workaround by:
- adding a field of type "File" to the CDS entity - (you may note that there is now an Image type field too. I tried that and it didn't work. I just haven't deleted it yet;-).

- adding a step in the Flow that creates the PDF to upload the "file content" of the CreatePDFfromWordDoc flow step to the new CDS field.
- Then I was able to successfully refer to that selected record from the record selected from a gallery pointed to the CDS entity. By using the .Value from the selected record/field as the input to the Document parameter in the PDF Viewer, I got it to correctly display the PDF.

Definitely "going around the block to get next door" but in the end it is probably a better design and will minimize runaway "intermediate" or dated versions of the printed document accumulating in the OneDrive folder.
I don't have time to solve the original question as it stands which remains, "can one display a PDF stored in a OneDrive or SharePoint Document library in a PDFViewer control?" I just don't know but hopefully, someone with deep product knowledge can answer that question. In the meantime, I have an effective workaround, and perhaps a better overall design.
I probably could have also gotten an Azure Storage account and used it to store the PDF temporarily to display it but that would be even clunkier and heavy-handed, plus my client hasn't committed to any Azure usage so far so that would have meant a lot of other lengthy discussions that I wanted to avoid.
My only fear is that I have been using the PDFViewer control now for just about 4 years and it is still classified as "Experimental"
. This approach, syntax and usage is not sanctioned in the documentation
as a supported input to the Document property. That states that the input to the Document property of the PDFViewer control is supposed to be only a direct https:// link to an unsecured URL with no redirects. I worry that 6 months ago from now something will change and suddenly this won't work then like my solution from 6 months ago doesn't work now but I will cross that bridge when I get to it. In the meantime, this looks like a pretty stable workaround but we will see as we test it with some 100 users or so.