IIhaI have jfjjjfjjfjjjfjjjfjfjfj
I started my PhD research in 1958.
I inherited a working Soft-XRay spectrometer which was beautifully engineered both by his predecessor.
Sevrin Crisp had inherited the basic experimental system from ???????.
Sev added in some further enhancements, and in particular, electronic enhancements that allowed the measured photons to be counted and presented as graphical intensities at increasing wave lengths.
However, the task of measuring 3 or 4 hundred positions, each about one millimeter apart, so that the intensity at that wavelength could be estimated from a variable graph. In particular, the user usually created a few (3-6) such graphs. Measurements from each point on the graph were added together in order to improve the accuracy of the resulting curves. This was a lot of tedious work.
Fortune was with me, because when I took it over, I was able to make use of another teams 400 channel analyser. Basically, as the specrometer slowy changed the detected wavelengths, the number of photons detected for that wave length was added successively to one of the 400 channels. This captured the graphs in an electronic form. It also allowed repeat scans, where the counts were added to previous scans. When there was a danger of an channel exceeding its maximum count, the scanning was stopped. However, if 6 scans could be used, there was 1/6 of the number to scans that needed to added. Further, its output was a print out of counts of the photon received. No measuring was needed.
Fortune occured a second time. The University installed a digital computer. The Professor of Physics, who was just taking on the role of Vice Chancellor, was against the purchase. He said that it would never be fully used. He didn't manage to stop its purchase. Suddenly, I could produce results using 100 scans, which gives an averaged value that is ~10 times as accurate the previous values.
However, the technology is not like todays. It was based on punch cards, one for each recorded value; the recorded channels were printed out as 400 five digit numbers on a sheet of paper. Hence, one run, containg results from about 6 scans, required 100 cards. This was achieved by having a typist enter the 400 numbers on a keyboard. The another 'girl' (Yes, 'girls' were used for menial tasks in those days). Another girl then loaded the resulting cards into a reader, and re-entered the same numbers. Their card machine compared the second set of values with the first and reported any that differed.
Then, I could use the computer to add the ~16 sets of cards, which then effectively produce the result of 100 runs. With is data I could make many refinements; such as extracting and the removing the inevitable back ground to the graphs, or look for kinks and other characteristic features in detail. Effectively, I was getting 10 times the accuracy by effectively summing together a 100 recorded spectra. I was able to record the satelite structure, that was barely visible 15eV below th Aluminium main spectrum. Hence, by using even more runs, I got some idea of the shape and position of the graph.
The computer runs invoved in this took several hours, and I was allowed to have that time provided I accepted the night shift. I brought in an old army folding bed and tied one leg of it up to the large clunky teletype unit used for controlling the computer. Then I set it up tp rint a message every hour and when ever packs of cards needed replacing.
This was not like computing today!!!