A lot of the tutorials on using SPI to talk to SD cards seem to show code that sends a 8 or so "reads" to the device before writing data.
Almost always anyway. I have gotten the beginnings of an explanation here and there, but nothing really descriptive.
The best I can gather is that you need to do some fake reads just to get the device a look at the clock so it can be ready when the writes come in.
First, is that about right?
Second, should I precede EVERY write operation with some of these pretend reads?
Third, 8 reads seems common. Is that a good rule of thumb?
Finally, if I'm completely wrong about the whole thing, care to set me straight?