10 ms·

Remember to always read at least twice, then compare the md5s before flashing Many SOIC8 clips can give an intermittent/bad connection. You will only realize

Remember to always read at least twice, then compare the md5s before flashing Many SOIC8 clips can give an intermittent/bad connection. You will only realize it when you try to restore and ... oops!

userbinator 7y ago
Also worth opening the file itself in an editor to at least see if something reasonable was read --- a consistently bad connection can give you a consistently bad file!
- eloy 7y ago
  or use `strings`
  - TiltMeSenpai 7y ago
    Specific to BIOS images, `binwalk` is probably a better choice than `strings.` There's often data in LZMA streams or otherwise compressed, and chances are you're gonna binwalk your image at some point if you want to modify it. Also, specific to Intel powered motherboards newer than a certain generation, the image should show as an "Intel PCH File Descriptor" or something if you run `file` against it. Don't use this as an integrity check though, `file` is just looking at an Intel-specific header at the beginning of the ROM file.
- devereaux 7y ago
  Yes, among the basic check with hexedit or strings as suggested below. When you are flashing a bios as the other poster mentioned, you want to be sure you have the right file and you will be able to restore it. About SOIC8 clips, if you can't get a reliable read, consider soldering with stripping wire straight to the IO ports. Stripping wire is thin enough that you can't break a leg of the chip.
neilv 7y ago
I actually had better luck with cheapest SOIC clips from eBay, than the Pomona clips. While trying to seat/reseat on the package, the lower pins in the Pomona clips would tend to get pushed in, bent, and pop out of their channels. The cheapest clips, with the pins that go straight through, however, didn't seem to do this. To make space to attach individual header hookup wires to the pins at the top of the cheap clip, I bent the pin tops left and right, in an alternating pattern.
mindslight 7y ago
And keep the leads short. This got me, and I should know better! Setting a lower SPI speed didn't help. It felt like there must be a flaw in the Linux SPI implementation (/BeagleBoneBlack driver) that even though I lowered the clock speed, the delay between data and clock was still tight. Either that or the signal times were so fast, they were coupling enough to override an active driver?? Rather than setting up the scope to investigate, I just tidied up the leads.

Remember to always read at least twice, then compare the md5s before flashing Many SOIC8 clips can give an intermittent/bad connection. You will only realize it when you try to restore and ... oops!

userbinator 7y ago

Also worth opening the file itself in an editor to at least see if something reasonable was read --- a consistently bad connection can give you a consistently bad file!

eloy 7y ago
or use `strings`
- TiltMeSenpai 7y ago
  Specific to BIOS images, `binwalk` is probably a better choice than `strings.` There's often data in LZMA streams or otherwise compressed, and chances are you're gonna binwalk your image at some point if you want to modify it. Also, specific to Intel powered motherboards newer than a certain generation, the image should show as an "Intel PCH File Descriptor" or something if you run `file` against it. Don't use this as an integrity check though, `file` is just looking at an Intel-specific header at the beginning of the ROM file.
devereaux 7y ago
Yes, among the basic check with hexedit or strings as suggested below. When you are flashing a bios as the other poster mentioned, you want to be sure you have the right file and you will be able to restore it. About SOIC8 clips, if you can't get a reliable read, consider soldering with stripping wire straight to the IO ports. Stripping wire is thin enough that you can't break a leg of the chip.

neilv 7y ago

I actually had better luck with cheapest SOIC clips from eBay, than the Pomona clips. While trying to seat/reseat on the package, the lower pins in the Pomona clips would tend to get pushed in, bent, and pop out of their channels. The cheapest clips, with the pins that go straight through, however, didn't seem to do this. To make space to attach individual header hookup wires to the pins at the top of the cheap clip, I bent the pin tops left and right, in an alternating pattern.

mindslight 7y ago

And keep the leads short. This got me, and I should know better! Setting a lower SPI speed didn't help. It felt like there must be a flaw in the Linux SPI implementation (/BeagleBoneBlack driver) that even though I lowered the clock speed, the delay between data and clock was still tight. Either that or the signal times were so fast, they were coupling enough to override an active driver?? Rather than setting up the scope to investigate, I just tidied up the leads.