A log scale but not the log of your data, that is exactly what semilogy is doing.
In fact I think you are not stating your problem very clearly either. What do you really want?
If you have a log axis, but then your data is not logarithmic, do you mean that you don't want to transform the data through a calculation process of your own/the Matlab or, you want to plot the 'original' data under a log scale? That would be quite absurd since your data would be jam-packed into a bundle of dots 'collapsing' onto each other.
If you really want to do this, you can achieve it by transforming the axis. Do something like,
ylimit = [0,35];
ytic = [0 5 10 15 20 25 30 35];
ylimit2 = log(ylimit);
ytic2 = log(ytic);
ytic2str = str2num(ytic2');
ax1 = gca;
ax2 = axes('Position',get(ax1,'Position'),'Color','none');