Wavemeter

Some details

1. Firstly, the 25mm square bearing bar was not bearing bar at all, just 25mm square ground steel. Available ready cut and ground. The paper was incorrect - the workshop at Melb Uni told me it was bearing stock but there was some miscommunication between various people there. Cost is about A$58 for 64cm (US$30). It's not critical, as long as the surface can be polished very smooth.
2. We had the ground bar and its supports, and the brass puck, nickel-plated, to prevent corrosion. It was cheap. The Ni plating was extremely thin, not a shiny Ni plate, just enough to stop corrosion. Looks essentially black (well, smokey charcoal perhaps).
3. If I was building another one, I'd use polished marble (or something non-magnetic) for the 25mm bar. Then we could use magnetic pushers to push the puck back and forth. Other people use marble bars. Cheap too, from your local stonemason.
4. We used solder for sealing the tube into the puck.
5. The cart is built from only two pieces of brass. The cylindrical CC holder is made from solid brass rod. The hole is first drilled, and then the top part is honed to provide a smooth surface for the CCs to slide into. One side of the cylindrical mount is milled away to provide a flat mounting surface.
6. The L-shaped base is made from one piece of solid brass, with the majority of it removed to give the L shape. The apex of the L is later removed to provide a flat surface for the cylindrical CC mount to sit against. A channel along the flat section carries the air to the holes that must be drilled through the base.
7. The two sections are actually attached with screws going up through the L-shaped base into the CC mount. (The heads of these are obviously countersunk so as not to come in contact with the track, and the screws are at the ends of the cart so that they don't interfere with the air channels.) This was originally done as a temporary measure during testing, but it has not been necessary to solder the two pieces. Soldering should work equally well.
8. In terms of the air pressure required, we don't know the exact figure since we manually adjust things until the cart is floating (usually done by inclining the track and increasing the air pressure until the cart starts to move). As a guide though, the regulator we use to control the pressure has a working range of 0 to 0.7 bar. We just had the airhose attached at a shelf above the wavemeter. We've also used a post sticking up above the wavemeter, something like 0.5 to 1.0m is fine.
9. The OPT 210 photodetectors are hard to find now. Just use a standard photodetector circuit. In fact, we don't have OPT210 detectors in there right now because we needed the chips elsewhere, we're just using standard silicon PIN photodiodes and a transconductance amplifier. Only needs to be good to a couple of MHz. You can read:
   
   "Photodetector designs for low-noise, broadband, and high-power applications" Malcolm B. Gray,a) Daniel A. Shaddock, Charles C. Harb, and Hans-A. Bachor, Rev Sci Instrum, vol 69 num 11, p3755 (1998)
   
   but the circuits in there are overkill for this applications.
10. Use small photodiodes - to reduce capacitance and therefore increase speed, but also it is much easier to get the wavemeter going if the detector area is small. If the wavemeter is not perfectly aligned, then the overlapping beams on the detector will have several fringes (this is normally the case for us). Let's say there are 10 fringes. If the detector sees all of these, then as the cart moves, the intensity will vary between 0.9 and 1. If the detector is smaller, so that it sees only half the fringes, then as the cart moves, the intensity will vary between 0.4 and 0.5 which is twice the contrast. Thus a smaller detector gives greater contrast.
11. The counter circuit has an error. The CU (count up) inputs on pin 5 of all the 74LS193 counters (U1 through U5) should be tied high, e.g. by a 1k resistor to +5V. Thanks to Thierry Berthier for this.
12. The counter circuit is very cheap, but not really very nice. I'd use more expensive all-in-one chips now, that will do up and down counting and also display the output (e.g., Micrel MIC 50395). Our counter only works when the cart is going from left to right, wasting all the time when it's going back. I'd also like two displays, one showing the last wavelength, and the other showing the current count. Just not worth the effort now that we have something working. Just make sure it will count fast enough (the chips we used could count pulses up to 5MHz) and that they can be preloaded with a number to count down from.
13. All the circuits were on different boards, because we made them at different times. The counter circuit was made by myself on veroboard so it's pretty ugly. The detector/TTL board was made by our electronics workshop.
14. The slotted optical switch is a very very common device used to sense location or presence of something, e.g. whether you have a floppy disc in your disc drive, etc. It's a LED and phototransistor (usually with a Darlington amplifier) in a single package. LED illuminates the phototransistor, making it conductive. There's a slot between the two, and if you put something between, the phototransistor no longer conducts. Hence you can sense the location of the cart, and use the signal to start/stop the counters. Have a look in a Farnell or RS catalogue, or ask the workshop.

Optics Group