©Lin Yangchen

Since almost the earliest days, I have worshipped the Berek compensator as the Tomahawk cruise missile of polarizing microscopy. A marvelous invention with miniature gears that rotate a small precisely cut crystal of calcite, magnesium fluoride or aluminium oxide on a precise axis, it can do everything the quarter and full waveplates and quartz wedges can do, plus measure the precise optical path difference even if the interference colours are masked by intrinsic colour.

Being, however, a cult accessory deployed only by fanatics of retardation quantification, there aren't many of them in the world, especially the original part for my Soviet-era ‘Special Forces’ Olympus BHSP. An exhaustive eBay search drew a complete blank.

Although the current Berek compensators for the BX series are of the correct DIN slot dimensions, they look overly complicated, have no vernier scale (an elegance unto itself) and are possibly mostly plastic like the BX fixed-wavelength plates.

I turned to other manufacturers. The Leica Bereks looked too complicated and fragile—what's with modern engineering and design? The Zeiss ones were nowhere to be found. Then I discovered something interesting—Nikon’s current polarizing microscopes use a compensator made by Nichika that looks virtually identical to the BH-2 one from the 1980s. Did someone copy someone? One Nikon supplier quoted me more than $2000 for the Nichika compensator. I tried to bypass Nikon. I dug up the Japanese listing on Nichika’s own website and contacted them. They bounced me back to Nikon.

You could make your own compensator, of course. Ian Walker has demonstrated a home-made substage Berek compensator, while Jay Phillips made a rotating substage filter that emulated the action of the Victorian selenite stage. But I already had neutral density filters taking up the space above the field lens and wanted a compensator that slid neatly into its slot above the objectives.

I was being driven to desperation when the big break came. I found the original BH-2 Berek compensator from a supplier in the United States for a fraction of the Nikon quote and with free international courier service.

It took three days to get from their shop to my doorstep, most of it first-class on a Boeing 747 (tracking by Flightradar24). What a far cry from the costly eBay shipments that took weeks crawling through obscure warehouses!

But this wasn't a happy ending yet. In fact, it was just the beginning of trouble.

This and all other Berek compensators I know of have a fundamental design flaw. There is no circular bright-field slot in the slider. This forces you to remove the compensator entirely when not in use, as it is too dangerous to leave it in the microscope with most of the slider sticking out. It violates my industrial design doctrine that accessories should remain attached to the microscope at all times. It drives me crazy to be exponentially inconvenienced by the absence of a simple design feature that would have been so easy to implement. And a big headache to fix.

I started by dismantling the compensator and isolating the base plate, hoping I could just send the plate to the workshop to cut the hole and alignment notch. However, the engineer said it would be too risky given the irreplaceable original, and advised machining a new plate from scratch.

Tedious measurements using a vernier caliper.

Using the microscope’s calibrated eyepiece micrometer to measure the thread pitch of the cover plate screws. A screw was leveled on stage using Bostik White Tack. Recorded thread dimensions: outer diameter 1.65 mm, inner diameter 1.2 mm, pitch 0.35 mm.

CAD drawing of my modified design for the Olympus AH-CTP Berek compensator base plate, special designation BH2-YCTP. The big difference is the big fat hole and notch (upper left). The most critical part of the plate, however, is the hole for inserting the rotating shaft (highlighted in green) connecting the gear mechanism to the mounting stem of the calcite crystal. This has to have a locational clearance fit—the closest possible fit that requires no force for assembly and moves smoothly with lubrication—to ensure stable and accurate alignment of the crystal. You can ask the shop to either check fit with the original shaft or machine a new shaft to fit. Recorded thread dimensions of shaft tip: outer diameter 1.9 mm, inner diameter 1.39 mm, pitch 0.4 mm.

The original (lower) and modified (upper) base plates. The latter was machined in aluminium 6061 and given a black anodized finish. I made a larger depression in the plate at the base of the crystal mounting stem so one can hold the stem with the fingers and a rubber band for friction while screwing or unscrewing the gear shaft, instead of having to use needle-nosed pliers. Perhaps Olympus made theirs small to discourage people from meddling with it. The hole near the middle of the original plate is for the screw whose head protrudes from the underside to stop the slider in the correct position. To avoid threading another hole, I simply drew a protruding knob of identical dimensions to the screw head.

Zero the compensator during reassembly. First assemble the slider portion without the vernier drum. If you are transferring the gear wheel from the old shaft to a new one, just make sure the gears are orthogonal to the crystal so they can transmit an equal amount of rotation from the drum in both directions. Then insert the slider into the microscope slot, observe through crossed polarizers and turn the calcite crystal until the polarization cross is centered in the view through the eyepieces.

Turn the drum knob until it reads 30.0°, position the drum atop the slider, and interlock the gear teeth. Insert and tighten the two screws joining the drum to the slider. The crystal shaft may have had to be rotated slightly for the teeth to interlock. In this case slightly loosen the two set screws on the gear wheel and use a slotted screwdriver to turn the shaft head until the polarization cross is again centered, while holding the drum stationary at 30.0°. Tighten the set screws.

The zeroing doesn't have to be terribly precise. The inclination measurements are made by turning the drum in both directions and taking the difference, so the end result is insensitive to the alignment. But it should be as close as possible to 30.0° to give the same amount of rotation in both directions.

Hunting down sub-1.5 millimetre hex wrenches for the set screws securing the gear wheel to the shaft. Only one shop had the microscopic sizes I needed. These are old but real industrial wrenches, not the low-quality household ones sold in most shops nowadays.

The retardation calculation also depends on a machine constant unique to each individual compensator, that depends on the optical behaviour and exact positioning of its particular crystal and on the illumination wavelength. This constant was measured for each compensator and recorded alongside its serial number in the instruction manual accompanying it. The constant may be altered in the process of disassembly, reassembly and switching of parts; it can be recalibrated by taking inclination measurements from a reference sample of known retardation and calculating backwards. I used Olympus' 530 nm tint plate for this purpose.

One screw for the cover plate couldn't go in all the way so I left it out. It doesn't affect operation or accuracy and can't be seen when the slider is in the slot. The elegant but serious industrial design of the 1980s—in solid metal—goes beautifully with the Soviet-era BHSP. All the technical lettering, numbering and scales are engraved, unlike the printed Microsoft Word-style font on Olympus’ BX compensators.

Determining the sign of elongation of a fibre. Beware—the directions of the slow and fast rays are opposite to those on a typical waveplate. This is to give the maximal change in retardation within the limits of rotation of the crystal.

For the physics and mathematics of Berek compensators see Burri (1950), Durey (2021) and the instruction manual.


Before I got the Berek compensator, I performed surgery on the λ/4 mica plate from my now-defunct Radical microscope to convert the plate to a de Sénarmont compensator. I used a toothpick to rotate the crystal until its slow direction was parallel to the polarizer with the slider at 45°. The slider was securely clamped with the mechanical slide holder, while the calibrated stage goniometer and eyepiece crosshairs ensured accuracy.

Olympus glued their crystals down so people couldn't tamper with them; thankfully this one was adjustable. But you can still get de Sénarmont compensation with a regular λ/4 waveplate by rotating the polarizer and analyzer to the 45° mark as the starting point. With this hack you can’t measure retardation all the way to the full wave, as the analyzer on the BHSP rotates only 180°.

Tightening the snake-eye clamping ring with a fine screwdriver tip while holding the crystal steady with a toothpick wrapped in a lens cloth.

My converted de Sénarmont compensator with a new handwritten label.

The microscope in emergency "night vision" mode, with interference filter affixed to the bottom of the condenser with 3M Magic Tape.


Berek, M. 1913. Zbl. Miner. Geol. Paläont.

Burri, C. 1950. Das Polarisationsmikroskop: eine Einführung in die Mikroskopische Untersuchungsmethodik Durchsichtiger Kristalliner Stoffe für Mineralogen, Petrographen, Chemiker und Naturwissenschafter im Allgemeinen. Lehrbücher und Monographien aus dem Gebiete der exakten Wissenschaften.

Durey, G. 2021. Polarized microscopy with the Berek compensator: a comprehensive tutorial for the modern reader. The European Physical Journal Plus 136:866.

Naidu, M. G. C. 1965. Berek compensator. Mineralogical Magazine 35(270):431–432.
Powered by SmugMug Owner Log In