fringes that originally existed. From the angles that the 
sandwich hologram had been tilted to transform Fig. 7 
into Fig. 8, it was also possible to calculate magnitude, 
direction, and sign of the knee tilt, in spite of the fringes 
originally being so closely spaced that they could not be 
properly counted. The speckles of Fig. 8 are larger than 
those of Fig. 7 for the following reason. When the 
sandwich was tilted, the fringes started to move when 
the point of observation was changed. Therefore the 
aperture of the camera that was used to make the pho- 
tograph had to be so small that the whole used area of 
the lens sees the fringes in the same position. Because 
of the small aperture the speckles are large in Fig. 8. 
There exist effective means to get around this problem, 
but in our experiments we were satisfied if the quality 
was such that we could count the fringes. 
It was, of course, also possible to study the deforma- 
tion of the head, but then it was necessary to tilt the 
sandwich hologram in the opposite direction (Fig. 9). 
The fringes on the reference surface could, however, not 
be counted because of the large speckle size. 
Trying to find the limits of the sandwich method we 
doubled the force difference applied to the machine and 
made a third sandwich hologram. This time we used 
a thicker extra glass plate to separate the hologram 
plates so that the distance between the emulsions was 
some 9 mm. The quality of this hologram and the res- 
olution of the fringes were not reduced by this large 
separation. The result without any sandwich tilt is seen 
in Fig. 10. No fringes are seen on the knee, but it was 
still possible to eliminate the about 250 fringes that had 
theoretically originally existed there to study the de- 
formation which had resulted in about twenty fringes 
(Fig. 11). In later experiments we have managed to 
eliminate up to 500 fringes, and we have not yet found 
the limit. 
VI. Results of the Second Experiment 
When I presented some of my holographic studies of 
a milling machine at the 25th General Assembly of the 
International Institution for Production Engineering 
Research in Germany, there was doubt that holography 
really could be used so simply directly on the floor of a 
workshop. It was mentioned that these methods could 
perhaps work but only on the stable Swedish granite 
rock. To test the statements in my paper I was invited 
by E. Matthias to repeat my experiments during a 1- 
month stay at the Department of Machine Tools and 
Production Engineering at the Swiss Federal Institute 
of Technology (ETH) in Zürich, Switzerland. It was 
a great pleasure for me to accept the invitation, and in 
January 1976 I arrived in Zürich with just one piece of 
holographic equipment, our special hologram holder. 
For the experiment a Swiss plain horizontal kneetype 
milling machine, Oerlikon model MN2H, was selected. 
It was about 1.7 m high and stood directly on the ground 
floor in a workshop which was only about 7 X 10 m. 
Because of the limited space we could not without dif- 
ficulty get further from the machine than 4 m, and the 
spatial filter had to be placed within half a meter of a 
window. ETH is situated in the middle of Zürich, and 
the stability problems appeared to be rather severe 
2528 APPLIED OPTICS / Vol. 16, No. 9 / September 1977 
because of the heavy traffic of lorries, busses, and trams 
at the street corner outside our working place. The 
floor was covered with cork tiles, a material we found 
rather difficult to work on because it retained previous 
weight depressions and dimensions for a long time after 
a load had been changed. Thus it was necessary to wait 
a few hours after changes in the holographic setup be- 
fore a hologram could be exposed successfully. 
For the illumination we used a Spectra-Physics argon 
laser model 165 which was kindly lent to us for 1 week 
by Brown Boveri Research Center in Baden, Switzer- 
land. We used an output of 0.5 W in the green 
514.4-nm wavelength. After some adjustments of the 
etalon we could prove with a simple interferometric test 
that the coherence length was more than 2 m and 
therefore did not put any tight restrictions on our ho- 
lographic configuration. The laser was placed along the 
wall on an ordinary table, and the beam was deflected 
90° by a mirror and directed onto a spatial filter which 
was fixed to a very stable, heavy steel table (weighing 
some 300 kg) onto which we also fixed the plate holder. 
The reference mirror was placed about half a meter 
from the milling machine on a stable stand that weighed 
around 100 kg. All the heavy equipment used were 
such pieces that could be found in any well equipped 
heavy duty machine shop. 
The milling machine itself had a green shiny finish, 
and a definite desire was that it not be repainted or 
treated in any special way. Therefore that those parts 
of the machine that produced a direct reflex got much 
brighter in the hologram than other parts was un- 
avoidable. 
The holograms were exposed during the night be- 
cause the activity in the institute was much lower and 
most of the traffic had stopped. The force deforming 
the machine was produced simply by turning the screw 
  
  
Fig. 12. The deformation of a horizontal milling machine was studied 
using double exposure holography. Temperature gradients in the 
air produced fringes on the reference surface at the lower left.