Change TABLE settings for mod files to cover wider voltage range

[pgleeson]

Currently most mod files have expressions like:

DEPEND Aalpha_n, Kalpha_n, V0alpha_n, 
       Abeta_n, Kbeta_n, V0beta_n, V0_ninf, B_ninf, celsius FROM -100 TO 30 WITH 13000 

i.e. precalculate a table of 13000 values for Aalpha_n etc. from -100mV to 30mV

This led to some difficulties comparing the mod file behaviour to the NeuroML2 channel behaviour, e.g. the value for tau is fixed at 30mV to 100mV.

@theiera is there any reason (apart from efficiency) that I shouldn't increase the bounds of this, -100mV->100mV (and increase table size)? Note that the peak of the spikes are slightly above 30mV. Also, erev of Na is ~87mV and so getting much higher than 30mV is not unrealistic.