It would seem that the OP answered his own question.
" From the SD bus transaction point of view, this is a standard
single block read transaction and the time out value of this command is 100=
 ms,"

I assume that the controller times out so that any writes in
progress can be completed before a card gets yanked if the external
controller mucked up the command sequence or hangs.

Did you read all the bytes the card expected you to read, or did you run in=
to
a boundary condition error in your loop and only read 511 of 512 bytes
expected for a block?

Thanks for asking the question. I learned something important about SD
card I/O on
slow processors (e.g. bit banging).



On Sat, Mar 7, 2015 at 2:23 PM, IVP <joecolquitt@clear.net.nz> wrote:
>> Here is another spec sheet I was able to pull up.
>> http://dlnmh9ip6v2uc.cloudfront.net/datasheets/Components/General/SDSpec=
..pdf
>
> Thanks. I have some like it, they're all a little bit different !!!
>
> There may be a glimmer of a clue in Section 4.7.2
>
> "The period after which a time-out condition for read/write/
> erase operations occurs is (card independent) either 100
> times longer than the typical access times for the operations
> given in Table 4-6 or 100 ms (whichever is lower). The times
> after which a time-out condition for Write/Erase operations
> occur are (card independent) either 100 times longer than the
> typical program times for these operations given below or
> 250ms (whichever is lower)."
>
> Perhaps though this might refer to actively waiting for the FE
> token before the real data. Some data sheets say operations
> should be abandoned or the card rejected if there's no response
> after a second.
>
> Really ? A second ? This is supposed to be modern high-speed
> smart memory. You wouldn't put up with that in a RAM IC
>
> Harumph
>
> Basically my set-up is
>
> (lots of CMD17) ..... 0.7s to process ..... (lots of CMD17)
>
> It's during the '0.7s to process' that the card decides not to accept
> the next CMD17
>
> One thing I might try is a little watchdog function that does a
> regular dummy read to keep the card interested
>
> Thanks
>
> Joe
>
>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2015.0.5751 / Virus Database: 4299/9249 - Release Date: 03/07/15
>
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.