A few easy mods for the SD4

Re. the well nut.  I've always wondered how people end up with backlash problems on the Z axis.  Seems like gravity should pretty much prevent it.  I am not seeing any banding to speak of in the Z (see below).  I suppose, for $2, why not.  But OTOH, if it ain't broke...

layer height is 0.25mm

I stand corrected.  You are correct on both counts.  I did a quick experiment (two actually) to test the assertions of the well nut and integer step improvements.  I created a 40x41.6 cylinder with 1.2mm walls where the lower 20mm was printed at .25mm(~35.433 steps) and the upper 21.6mm was printed at .254mm (exactly 36 steps).  I then printed twice.  Once with the well nut (on the left) and once without (on the right).

Two observations:
1) The .25mm does show some deformation manifesting in the form of diagonal bands.  Harder to see on the print w/o the well nut in the image due to the lighting angle, but it is there on both about equally.
2) There is a improvement in the horizontal defects with the wellnut in place. I was surprised at this (as evidenced in my previous comment) but there it is.

All of that being said, your prints are looking good so keep it up. Now you need an e3d to make those start points less horrendous.

layer height won't help you with that seam, that's just what the stock hotend does. it likes to blob all over your prints because it's a jerk.

It didn't eliminate it, it just covered it up.  If you get things properly calibrated you can worry a whole lot less about "I try to make my geometries integer multiples of the bead dimensions".  Slice software has come a long ways in the past year so let it do its job and you might be surprised at the results. 
Moire is not a real issue as long as you don't mind the appearance.  Some like it and some hate it.
With the stock hot end you can come very close to eliminating it with slight adjustment of the VREF.  Don't let turning a POT intimidate you.  Once you adjust it you'll ask yourself why didn't I do that sooner?
As far as the blobs at the start point go...
You can select random start points in your profile
Or you may want to increase the retract distance and or speed.  I have found speed to have the most impact with the stock hot ends.

Now if only I could score a ceramic DMM probe to prevent *it* from shorting things out... ;-) 

I did check the Vref on the extruder and it was off.  I measured .438.  Solidoodle support says it should be 0.250!  I dialed it down to 0.250 and the Moire pattern is almost eliminated even with the 0.25 Z height.  Yay! - I can go back to easy math.  My understanding from the A4988 part sheet is this sets the max current for the DAC.  So am I correct to think that the distortion from too high Imax is caused by non-linearity in the stepper response to high current?  Seems to make sense - would result in a modulation in the flow.