As David said... need more info. I've scanned a number of my own books into my Sony Reader. Here's the process I use.
Flat bed scanner (I use a
OpticBook 3600) but most flat beds can handle this, it's really a matter of how long it takes.
I set it to gray-scale 600dpi. I've found that black and white cause too many strange artifacts and since it's all just black or white the OCR program can't ignore them easily. Also, I've just had better luck with 600 dpi when it comes to the OCR.
OCR is good not just because it makes your text more usable (converting to other formats, etc) but also because it drastically reduces file size (usually.)
The OpticBook actually comes with very good OCR software, but there are other options available (including Acrobat Pro - though it's not too good in my experience.) Anyway, use OCR on it somehow - it may help to combine the scanned images into one PDF first.
Most good OCR (aside from Acrobat) has an option to export as a word file. I used to do this all with basic text, but found this was a faster process and was a more reliable option.
Opening in word, you begin the process of "fixing" which can take anywhere between a half-hour and hours.
Once this is done, I save the file and open it with Book Designer. I then set the chapters up and then save as an LRF.
I've had 200 pg books end up only about 300 kbs this way, and work like a dream.
Now, if you're dealing with travel books with lots of pictures I'd handle things a bit differntly. I'd still do all of what I suggested, but when you get to the book designer page find the images you want in the scanned images and drop them in the appropriate places before turning it into an LRF.
So, that's really all I can do to help - hope it does. As to why any given PDF doesn't work - no clue. I try to avoid using PDF files at all.