A few comments from someone doing microscopy (but not waffers) for a living and build his own experimental microscopes in the lab:
1) In the beginning dave said it's an stereoscopic microscope, which is off course not true. The Mantis would be stereoscopic, e.g. having two slightly different angeled beam paths for both eyes to give 3D perception. This microscope however doesen't have that, both eyes see exactly under the same angle (and i guess dave nows this). I assume he meant that it has a binocular tubus (which is better than a monocular tubus, you have to close one eye all the time which can be tiring). Unfortunately a mistake often made.
2) The fact that your Objectives are not parfocal (you have to refocus after switching the lens) is caused by wrong adjustment of the camera distance and eyepice distance. You should follow the following procedure:
a) set the eyepices to the right distance for your eyes
b) read the eyepice distance (in mm) at the scale above the eyepieces. The barrels around the eyepices have the same scale, set it to exactly these values
c) change to the highest magnification, focus on your sample. If you need glasses, do this step with the glasses or cantacts.
d) check if switching the lenses still requires focus adjustment. If so, slightly adjust the barrels of the eyepices till you get rid of this effect
e) when this works, focus on sample with high magnification and adjust camera to be focussed.
This procedure is quite complicated with this type of binocular tube, other manufactureres and later olympus scopes had other constructions that adjust the length of the optical path automatically when you change the width of the eyepices.
3) as mentioned before, it seems the table is slightly tilted. This could have happend in transitt, but can be solved by a skilled repairman (by shimming the table or slightly losening screws holding the table mount to the z machanism and tapping it back). But the effect seems to be very slightly, so i wouldn't attemp to correct this. Chances are you make it worse.
4) Adapting a LED is a good way to upgrade old lightning. There are a few important notes:
a) you don't need much power (as you experienced). 1W is more than enough for brightfield and polarisation microscopy if adapted correctly (see below). Darkfield might use a bit more power, but dave doesn't have any of the necessary components anyway.
b) You should use a LED with a single emitter, you want to have a homogeneous light source
c) the size of the emitter should be the same as the size of filament in the old bulb
d) the LED surface should be at same position the bulb filament was in (and don't remove any lenses from the lamp housing). Idealy you should have an image of your light source in the back focal plane of the lenses (which you can check with special tools if necessary).
5) The optical quality of the middle class microscopes from india/china is quite good nowadays and can compete with the entry/mid level lenses from the big four (Nikon, Olympus, Zeiss, Leitz/Leica). If you need special lenses, you will end up with the big four (ans paying a few thousand bucks for one lens). The main disadvantage of the cheaper microscopes (like AM Scope) is the not so good mechanics and aviablility of aditional parts. The Olympus scope dave gought can be modifies for transmitted light, can do polarisation with a rotary table, Differential interference contrast (DIC), Phase Contrast, Fluorescence, darkfield, .... These microscopes are like a lego set, you have one frame and can add lots of different parts to create basically every technique you like.