ECDL module/driver

[gkasprow]

We will need to miniaturize laser rack in our nex project(transportable optical clock). For the moment the Toptica lasers occupy entire 19"rack. We also want to play with ECDL lasers, especially im UV region. Such diodes with AR coating are produced in Warsaw(TopGAN).
So the idea is to pack all lasers and controllers into DI/OT
Some thoughts to discuss:

• use Kirdy analogue circuits as a current / temperature controller but expose them via SPI to the DIOT backplane
• reuse Toptica connectors/pinout to be compatible with their ECDL modules
• add piezo driver to the above controller board
• use STM system board or Kasli to control multiple drivers
• use FastServo with Linien to lock lasers to references
• in the longer term, build DIOT modules with ECDLs based on COTS components from Thorlabs. Of course entire rack would have to be fanless.
• in the longer term build frequency reference based on temperature stabilised Bragg grating. My colleagues from WUT have all the tools to make custom Bragg fibres. I want to explore ideas from AI-ARTIQ project but using bigger optical components.
• prepare all HW for liquid cooling - could be critical for miniaturization and fanless operation

[sbourdeauducq]

pack all lasers and controllers into DI/OT

How to get the beam out (fiber coupling adds substantial optical losses and complexity) and isolate vibrations? The conventional solution of the external free-space-output laser head does have some merit.

My colleagues from WUT have all the tools to make custom Bragg fibres.

Writing fiber Bragg gratings for blue/UV wavelengths sounds quite difficult. Unless you are e.g. exploiting some nonlinearity of the fiber to get a non-sinusoidal grating, you need to write the grating with a laser whose wavelength is the same or shorter (typically much shorter, e.g. use a excimer laser to write a grating for 1550nm) than the wavelength that the grating is designed to reflect.

Thorlabs also has a similar design:
https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=7783

If you want a quick experiment, such 1550nm fiber gratings are available at low prices and fast lead times from China, and so are 1550nm FP diodes - but without AR coating or angled waveguides, so you'll get coupled cavities instead of pure external cavity. It may be interesting to study the coupled-cavity system since, if it works, it automatically takes the price and availability of the diodes out of the quantum gravy train territory.

prepare all HW for liquid cooling

Will it dissipate so much power that convection cooling is not enough?

[sbourdeauducq]

An alternative to ECDLs is frequency doubling/sum-frequency generation of telecom lasers. This technique is exploited commercially by Precilasers. They use very expensive PPLN waveguides for single-pass operation at somewhat reasonable efficiencies. It sounds like the latter can in some cases be replaced with nonlinear optics in a very inefficient and inexpensive conventional crystal (e.g. LBO) followed by injection-locking of a FP diode to get back the optical power. Contrary to conventional injection locking, this doesn't need an expensive circulator either since you can put a dichroic mirror instead. This scheme could lower the cost significantly and remain rather immune to vibrations.