I would not think the lattice constant is a problem if you are using a ribbon or similar. When averaged over the XY plane, the plot of the voltage drop vs. Z shows a more smooth shape that tends to show quite well where the resistance is located in Z. But of course you can easily plot Vdrop(x0,y0,z) as function of z by calling evaluate(x0,y0,z) on the voltage drop object (obtained by subtracting the electrostatic potential at bias minus the zero-bias potential). Like so:
# Read configurations from NC files
device0 = nlread("zerobias.nc", DeviceConfiguration)[-1]
device = nlread("finite_bias.nc", DeviceConfiguration)[-1]
# Evaluate the voltage drop as a grid object
v0 = ElectrostaticDifferencePotential(device0)
v = ElectrostaticDifferencePotential(device)
voltage_drop = v-v0
# Point in XY plane
x0 = 5*Angstrom
y0 = 5*Angstrom
# 100 points along the z axis; cell_z is the length of the central region
cell_z = device._cell()[2][2].inUnitsOf(Angstrom)
for z in numpy.linspace(0,cell_z,100):
Vz = voltage_drop.evaluate(x0,y0,z*Angstrom)
print z,Vz.inUnitsOf(eV)