i continued and asked me why
the test in atray was not representative because the geographic position was constant and finally the position if the sd card has no importance
after real tests in a river i think that if the speed is too hight and/or the way is not enough straight, the helix consider that the the si result is not valid and stop it's record
but
a long time ago, i recorder successfully a long way with many forward and backward at 2 m/s
but the record of an other straight way at also 2 m/s failed
so the problem is not solved
deos somebody has this problem ?