I managed to get my hands on an Olympus BH2-MJL microscope (a metallurgic model). A rare pearl.

The Olympus BH2 is an old large Japanese microscope ultra customizable, with infinity corrected optics. It allows to add elements between the eyepieces and the objective.

You can also change the head itself (monocular, binocular, trinocular …) easily, add illumination from below to use it as a conventional microscope.

Vertical illuminator

Illumination is done from above (reflective illumination) and uses an Olympus BH2-KMA.

Turret and objectives

The turret allows to place up to 5 objectives and is motorized (it doesn’t seem to work: wiring to be checked, but you can still turn it manually).


  • 5x MSPlan 0.13 ∞/- f=180
  • 10x MSPlan 0.30 ∞/- f=180
  • 20x MSPlan 0.46 ∞/- f=180
  • 50x MSPlan 0.60 ∞/- f=180
  • 80x Plan 0.95 ∞/- f=180

Here are pictures of the same chip using each objective.

Objective 5x MSPlan

Objective 10x MSPlan

Objective 20x MSPlan

Objective 50x MSPlan

Objective 80x Plan

Nb: photos are taken by hand with a Nikon Coolpix S600, without any adapter.
This camera is a bit old (2009) and the quality is not representative of what we really see.
The image is much blurred on the photographs.

As you can see on the pictures, the 80x Plan lens is so blurred that it is useless and it is better to use the 50x MSPlan lens for high magnifications.

Possible improvements to the microscope

There are other illuminators, such as the Olympus BH2-UMA, which seems to have more options. The lighting could be improved by using LED lighting. The current lighting gives a yellowish color.

One could also consider having darkfield illumination in addition to brightfield illumination, in order to better see the reliefs.

A better lower zoom lens to replace the 80x lens could also be useful.

Setup upgrades


The priority at the moment is to have a better camera to facilitate the shots while not losing quality. It would be possible to make a 3d printed adapter for the small nikon but it has several problems:

  1. It does not have a high enough definition
  2. There is no simple way to start captures directly from the computer.
  3. It cannot be connected to the computer like a camera.


Then, the idea is to make a precise numerically controlled system to move the DIE automatically, and take the pictures between each move.

This becomes really necessary when you want to make a complete reconstruction of a DIE with a strong zoom.

For example, to reconstruct the DIE of a MOS 6502 with a 500x zoom, you would need close to 1000 photos.