After I was done with my free-running diode lasers, I was focusing on building a robust ECDL as a prototype and reference model. Some preliminary info is here. Concurrently I was playing with the remodelled commercial ECDL described here. Here a pic of the latest prototype:
Both lasers work reasonably well, but the main problem ist that at higher powers, the zones of single mode operation of typical DVD diodes becomes very small and thus the lasers can drift out over time. To combat/control this problem, I had developed a noise detector which works very well for free-running, non-ECDL lasers. As for ECDL lasers, I found that they sometimes run multi-mode without the presence of extra noise in the light output. This would be of course an important problem for day-to-day use, and I am spending most of the time with trying to get arund it. It becomes worse with increasing power, and problematic beyond 70-80mW for common diodes.
It seems that one needs to very carefully select a diode, adjust the operation point and the optical properties of the resonator. Particularly well behaved are the diodes that also work well free running, like the ML101J27 and the HL6385DG. It seems that one can typically achieve up to approx 80 and 100mW, resp., in relatively stable single mode operation. But I need to spend more time investigating; there is a lot of variation between individual diodes and the number of parameters in the game is so large that I don’t yet have a clear picture. Moreover I am running out of the Opnext HL6385DG diodes I got moderately priced via a group buy.
Other things I was busy with were finding better collimators; so far I mostly used the familiar Lens-27 from Roithner, which has an extremely high efficiency but a relatively poor beam quality; some holographers had complained about this. I found that the G-650-1 (obtained from here) has a much better profile while having only slightly less power. Some of my findings are here.