I worked on a professional laser that used ~60 optics first on tiny mounts. Mounts were constrained by slots on the baseplate. These slots ran along the optical axis of the laser.There were fold mirrors at the end of each slot, so the laser was compact.
The idea was each optic was professionally machine aligned at the factory and can be moved at will. Revision two, probably inspired by customer complaints, caused each mount to no longer be magnetic and now had a pin on the base that dropped into a grid of holes in the slots. The lid was stiffened and held the mounts from escaping, but did not rest on the mounts. Again the idea was the CNC was so precise that nothing untoward would ever occur.
Everything was good until you consider elements that have to rotate around the beam axis, like collimator lenses in diodes or waveplates. So they widened the acceptance angle on the non-linear optics crystals, resulting in the need for fantastic amounts of pump beam power, and massive amounts of realigning the pump every time something rotational moved.
It can be done, but all optics need CNCed mounts, you need to constrain at least one axis of motion, and one good drop in shipping........ The above laser was shipped with ALL optics removed except the fold mirrors. If inverted, all the optics fell a mm or so until caught by the lid. This was never shipped assembled, A Factory engineer had to build the laser on site*.
Net result, great if you don't mind readjusting the axis frequently. Pins were far better then magnets.
To Do this, to align the optics to the mounts, you need a test jig and a autocollimator. Then a multiaxis jig to hold the optics till the glue dries. In some cases you need to order the optics with precise amounts of wedge etc.
All low cost optics have at least some tiny amount of wedge, the question is how much do they really have?
One thing to think about is kinematics. It takes three points to define a plane. Other thing to think about is how much warp occurs in a baseplate that has not been through a Blanchard Grinder for flatness.
*(interferometer guided CNC was used, CNC machine was probably better part of 1 Million Dollars. CNC guy was a wizard, he did pattern artwork using the few toolmarks left in the aluminum.)
The above are widely regarded as some of the best lab lasers in the world. However magnets left the design early. I'm pretty sure the company hobby was Chess, not soccer.
--------------------------------------------------------------------
Commercial optics mounts on switchable Magnetic bases exist. But having used them, you do need to touch up the positioning frequently. They are not used on baseplates that are moved around town. They are used on very flat baseplates that have been surface ground, otherwise they subtly move with time. If you place two of them together they do interact until the lever that turns the internal magnet down into the base is turned.
Usually we use them when aligning, not for day to day work.
---------------------------------------------------------------
Again, its not a bad idea. It has been used commercially. It requires a great deal of thought for the implementation. However most people soon switch back to the classic grid of threaded holes.
Go to
www.thorlabs.com and have a look around.
Steve