I don't know yet if this applies to the SE, but on other detectors I've used, a phantom signal can often be brought back by lowering the discrimination. When the signal was first found, the 'halo' effect allows the signal to make its way past the disc. But when the ground is opened up and the matrix is disturbed, the halo disappears, and no more signal. But, if it's due to disc kicking in, then a null in the threshold should be present instead.
Also, from what I've read the SE seems to not like air as much as soil. Is it possible that some of these phantom signals are caused by the absence of soil in the hole between the target and the coil? One possibility might be that there's a bit of a no-man's land between how deep the pinpointer can reach, and the coil also not seeing it due to the air gap now in the hole. So if the target is at 10 inches, and the hole is only 7 so far, neither the pinpointer or the coil see it. You fill back in the hole, and poof, there's the signal again! The coil sees it now since the soil is back. Does anything like that ever happen?
HH,
DirtFlipper