SDHC card of 4GB causes problem.

[daid]

While using an SDHC card of 4GB with SDSUPPORT then you run into strange problems. My print suddenly restarted on 12% of printing. The Sd2Card.cpp seems to have proper support for SDHC. But the bug might be in the fat implementation.

[bkubicek]

yes, SDHC is much more tricky, because you need fat32. You can try to reformat to 2GB partition size with fat16.

Bernhard

On Fri, Mar 16, 2012 at 4:25 PM, daid < reply@reply.github.com

wrote:

While using an SDHC card of 4GB with SDSUPPORT then you run into strange problems. My print suddenly restarted on 12% of printing. The Sd2Card.cpp seems to have proper support for SDHC. But the bug might be in the fat implementation.

---

Reply to this email directly or view it on GitHub: https://github.com/ErikZalm/Marlin/issues/116

[bkubicek]

sorry, my reply was false in content. I have actually no clue whats going inside the fat libraries, however I did initially changed them so the work/compile in the Marlin firmware (basically just trying to get rid of unused junk). What I can tell however:

I have a 4gb sdhc card. And a canon powershot camera (sd support) and a dell monitior with sd card reader. Ok, so factory ways, the card was formated with fat32. Both devices did not work with it. Neither after I tried to use a 1.5 gb fat32 parition. And neither if I tried to format the 1.5 gb partition with fat16. I hence have not tried at all with the ultipanel, thinking that if profesional devices fail, most likely the arduino libraries also fail.

Bernhard

On Fri, Mar 16, 2012 at 5:43 PM, Bernhard Kubicek < bernhard.kubicek@gmail.com> wrote:

yes, SDHC is much more tricky, because you need fat32. You can try to reformat to 2GB partition size with fat16.

Bernhard

On Fri, Mar 16, 2012 at 4:25 PM, daid < reply@reply.github.com

wrote:

While using an SDHC card of 4GB with SDSUPPORT then you run into strange problems. My print suddenly restarted on 12% of printing. The Sd2Card.cpp seems to have proper support for SDHC. But the bug might be in the fat implementation.

---

Reply to this email directly or view it on GitHub: https://github.com/ErikZalm/Marlin/issues/116

[daid]

Well, I can tell you, SDHC always has been a bit of a problem for these kinds of libraries. I've seen it happen with the GP2x and with Wii homebrew.

I have an SD 1GB card, which I formatted fat32, and I haven't seen problems yet with that... maybe it's a problem with an int/int32 mismatch in the fat library? Maybe it just fails to address the right blocks?

[daid]

The plot thickens.

I actually made an image from the 4GB SDHC card that was causing the above problem. And yesterday I picked all the sd card code from Marlin, and added some wrappers around it so I could run it on x86. With some hacking I added a simulated SPI interface which read from my image. However, when I do this the bug does not happen. I also replaced all the occurrences of "int" with "int16_t", so the ints would be 16 bit. But this still keeps the code working.

So I'm not yet sure what is causing the problems with SDHC...

[nophead]

Changing the types to int16 on a 32 bit compiler will not emulate a 16 bit compiler. Expressions get promoted to int in C and C++, so it will still do 32 bit calculations but store 16 bit results, which is not the same as doing 16 bit calculations.

[daid]

True true, but this means there is no variable that is too small. There could still be an over/underflow in a calculation somewhere.

[github-actions[bot]]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.