The most significant change done lately was to reverse the polarity of the tip bias, so that now, once again, the tip is +ve relative to the sample. The circuit being used was not designed to do this and so a little re-wiring was required. Both the circuit diagrams and the software have been updated accordingly. There was also an error in a circuit diagram indicating the point of connection for injecting a sinewave onto Vb.
The tip was made -ve for a while on the suspicion that, while scanning in air, there will be less of a tendency for stray particles to deposit themselves on the tip. Unfortunately the circuit did not behave as previously. The appearant asymmetry of the analogue circuitry about midpoint-ground might possibly be due to some diode action at the tunnel gap itself. The change had brought about a need for higher gain and also resulted in an increased drift in slewing on open loop. This drift in turn interfered with the coarse-approach mechanism and sometimes stopped it altogether. It was therfore decided that using the present circuit, the potential advantage of the -ve tip was not justifiable.
Another recent change is the removal of the interleave "padding" within the scan data. It was originally thought that this padding would keep the 3-D renderings smooth while keeping scanning times short. Experience proved however that this merely created scans that were fuzzy (and less accurate). Moreover, when the smoothing interleave filter was ommitted, the default renderings had unreal ridges emphasising the surface features. The rendering routine has now been modified to use only the rows containing actual scan data.
And finally, the tip-etching section now describes the benfits of using HF for oxide removal.