Thread: (Nuvi 660) Is it possible to convert OEM FW to regular one?

@ddabcd277

I'm out of town now -- will upload the image later this week when I get home.

The way I put it together was to just take the header out of the logo.bin you extracted from the "PouchX dump" up to the size of the image (the 4 bytes turboccc mentioned in the other thread), and then appended a 480x272 RGB565 image after it; then added a few FF bytes to the end to round the size up. Per turboccc the last byte should be the checksum; being lazy I did not calculate it, rather I just threw in a random byte. Then I loaded it into GIR_editor, which accepted it; then replaced the image with another 480x272 (actually the same one), and save it -- GIR_editor seems to calculate the correct checksum in the course.

This followed the recipe turboccc posted on the other forum. This may not be exactly what the 660 wants (looks like not) but this is how GIR_Editor accepts.

BTW I compared your extraction with the larger bin file, and you did not seem to break anything; what I didn't understand is where the extraction ended -- if we take turboccc's spec as the rule, then there should be exactly 480x272x2 bytes for the image after the header, followed by some numbers of FF filler bytes (and a byte for checksum). In your logo.bin, the "image" part is larger than 480x272x2 (and that is why GIR_Editor would crash). How did you decide the size?

Sean

p.s. Now that I think about it, it could be that I/we did not fully comprehend the meaning of the header. As I said the garbled images all look like images with wrong widths (e.g. if we display a 6x6 image as 9x4, we would have similar effect). So maybe there are some bytes somewhere that specify how the second display should work. I'll try more when I get back.

p.p.s. I noticed on the other forum that someone also had issue with the second display (being completely dark in his case). And someone noticed that for that device logo.bin contains 2 images (or 4). That is why I tried a 2-image version, to no avail. Note however that I was merely following a GIR_Editor compatible format; that may not be the correct one.