USING A FRINGE LOCKER
With a little cleverness, any split beam recording geometry can be electronically stabilized. The requirements are that one path contains the bimorph and that light from both paths is available near or through the recording plane. When the recording is of two plane waves then the combining of waves is a trivial matter and can be done with either a beam splitter or a grating. For any other type of recording a little ingenuity can be applied to arrange for stray light to be present at the appropriate plane.
By far the two most common table setups are the reflection and the transmission transfers, where the image reconstructed from an H1 hologram is being projected onto an H2 recording plane. Examples of each are given in Figs. 2 and 3. Fig. 2 is a basic reflection transfer geometry to which has been added a bimorph with mirror, an optional lens, and a second beam splitter. BS2 has been strategically placed where it can see pieces of undisturbed reference and object light together on one surface. It passes part of one leg and reflects part of the other leg at an angle that makes the two waves travel co-linearly. The combined beams are optionally made to diverge by the lens and are detected at an appropriate distance where the size of the fringes is observed to be about equal to the space between the differential detectors.
The detector head may be placed on a separate table and additional mirrors can be used as needed to get the pattern to it because downstream from the combining beam splitter, the fringe pattern is no longer affected by components. The beam splitter itself can be anything from a simple piece of thick glass to a coated wedge. The clearest fringes will of course produce the best lock so the better the beam splitter the better are your chances of staying locked. Nevertheless, even a noisy, banded output can lock very tightly and will work well if nothing better can be had.
A convenient way to mount the splitter is to fasten it to a stable tilt mount so that it extends above or to the side of the mount. Small adjustments in angle provide precise control of fringe size and position. Locating fringes initially can be tedious but if a cross is first drawn or scratched onto the splitter surface, the task reduces to adjusting the splitter until two crosses overlap. The difference between the reflection and transmission configurations is simply the angle that the beam splitter is placed at.
Again, some cleverness is required to see that the two legs of the system provide some light that intersects near the recording plane. It is not a problem if the light must first pass through H1 or even if you have to add one or two more small, stable mirrors to the system to get pieces of light to intersect at the film plane at useful angles. If the angles they intersect at is very small then combining the two beams with a small grating formed by the intersecting beams is the better way to get useful fringes. The grating replaces the beam splitter and is tilted to adjust fringe spacing on the detector.
If no extra light can be had in any way to intersect near or through the recording plane then an auxiliary laser can still be used to lock the table. The requirements are that both lasers reflect from the same bimorph at similar angles and that the paths of the second laser closely match the paths of the first. This arrangement is useful when recording with a HeCd laser while locking fringes with a small HeNe laser. Keep in mind that the compensation you are making is primarily for thermal drift in the table and in some primary components. Secondarily, the locker is compensating for local vibration of the table or components. Air currents are only compensated for at the second beam splitter so it will be necessary to use curtains or air dams to obtain the required air stability.
The locker will be most appreciated when exposures of 3 to 20 minutes are contemplated with spread out beam paths on 8 to 12 foot tables. The thermal drift in such time frames is certain to be a significant part of one or two fringes for all but the best of tables and components. If your exposure is only a few seconds the locker will still work and may improve uniformity in a production environment but its utility may not equal the difficulty of setting it up. If you find you are stumped by an application then try a call to the manufacturer, fax them a drawing and ask for assistance or for a knowledgeable user referral.