Hi, thanks for the answer.
I did some "debug", and found out, the problem was caused by my BIDIR ports of the SPI slave (my design enables QSPI). once i defined the port as input only, the voltage was O.k.
after some more debug, i realized, that for some strange reason, the FSM read flag was raised => the inputs were turned over to output (thus explaining why the voltage fell after the first byte...).
this is very wierd since i sent a"write" command, and according to my test block, the command was accepted correctly by the slave.
do you have any idea how could a signal go high even though on simulation it didn't?