3. The Prototype
In 2011 I discovered the IT department at my workplace was doing an upgrade and they kindly agreed to donate an unwanted 19" HP server rack to my home project.
I now proceded to assemble everything into the rack. Naturally, I put the heavy power supplies at the bottom, followed by the equally heavy Acton Research Spectrapro Sp-275 spectrometer.
Aware of the dangers of an invisible 6MW laser, I bought a 2U height drawer and installed the MK367 and circuitry inside it, the intention being to never run it unless the drawer was shut.
Above that I put a sliding keyboard/mouse tray (more protection from the laser drawer below it) and above that the PCM401 PC and finally the monitor.
The PCM401 software also functions as a highly accurate spectrometer, the photograph to the right shows it verifying the wavelength of a Chinese 905nm laser diode from eBay that turned out to be 891nm.
Above, Stellarnet fits their entire PortaLIBS into a tiny suitcase - mine occupies a huge 19" rack!
http://www.stellarnet.us/systems/porta-libs-2000-and-plasma-monitor-configurations
Having everything conveniently arranged in the rack meant I could store it in a corner of my lab
(the smallest bedroom of my house), and wheel it out for experimentation. The laser drawer is removable, and it was easier initially to test the laser with the drawer on my lab bench where I could fine-tune the circuitry and verify its operation on my 100MHz Rigol DS1100CD DSO [I13].
The rack drawer is made of (black painted) steel, which is difficult to cut and drill, and I needed an easy way of assembling the MK367 and components in it that would allow adjustment, but a conventional optical breadboard was too bulky (and too expensive). The solution was to cut a square of MDF the size of the drawer and spray it matt black, insert it in the bottom and mount everything on the wood, drilling holes where necessary. I anticipated I might try using both MK367s, and initially made cutouts to accomodate two of the 20µF 1kV capacitors, but I eventually settled for just one MK367:
Above and below right - the MK367 is enclosed within a ceramic body. It's specified at 0.3Hz but will fire up to 1Hz on a heatsink. The body outline is difficult to accomodate without special machining, so I wrapped the centre in compressible Bergqist thermal interface material and added a further small heatsink to the top. This assembly helps dissipate heat when it is running in multi-pulse mode. The high voltage cables run across the metal heatsink through insulating plastic tubes, mating with the MK367 pins using crimped sockets hidden inside nylon covers that were once the protective covers on a UK mains plug.
Top left are two blue safety warning relays that are activated when two tubular super magnets are moved over them by a solenoid. Out of view, the other end of this solenoid is attached to a DIY razorblade beam dump that moves out of the way to permit the laser beam to exit at the rear of the tray.
The aluminium heatsink is secured to the inside of the steel drawer using six super strength rare earth magnets (the thick steel drawer acts as a safety cover and I wanted to avoid drilling holes in its front in case I changed the design) and adhesive mounts onto the MDF below. Immediately behind the MK367 is the small Chinese red laser diode (LD) module and above the heatsink is its TO-3 78P05 +5V regulator (way over spec, but in a convenient metal package), and beside it, its switch. The LD body is live, so it's insulated and mounted to its heatsink with nylon screws, similar to the cotton reel which forms the 'air' core for its 22µH inductor.
Above: the initial MK367 firing components used a manual pushswt
to +12V instead of the CCS 4001.
Above: The laser's aluminium heatsink is secured to the inside of the steel drawer using six super strength rare earth magnets
(the round silver discs to the left).
Once I had confirmed the laser flash lamp fired, I needed to check the curve of the current pulse, to confirm the flash lamp was being correctly driven: not ringing or over-damped, either of which could damage it. The current curve needs to be critically damped, illustrated 40% down this page:
[E16] Module 3-2 Pulsed laser flash lamps and power supplies
http://pe2bz.philpem.me.uk/Lights/-%20Laser/Info-999-LaserCourse/C03-M02-PulsedLaserFlashLamps+PowerSupplies/mod03_02.htm
High current flows when the flash lamp ionises and to view this I used a Pearson 410 current transformer which has a risetime of 20ns and produces 100mV output for every amp through it, up to 5,000A with a 50Ω terminator, datasheet here: [E3], see also [Research: High Speed Current Probe].
http://www.pearsonelectronics.com/products/wideband-current-monitors
The drive circuit is as follows. The gate of a Fairchild FQA8N100C 1kV 8A N-channel MOSFET [D32] is fed with a manual single shot CMOS pulse from a Global Specialities 4001 pulse generator, connecting the MOSFET drain to the Pacific Designs PSU through a 120kΩ limiting resistor, and the source through a WIMA 250Vac 100nF capacitor into the [D28,D29] EG&G TR-123C trigger transformer primary, the other end to 0V. Behind the photograph, the rack with the Wilmore 1515 capacitor charging PSU (+673V) and 2K-10 PSU (+142V) feed into the drawer via HV coaxial leads. The Rigol oscilloscope top trace is the start pulse from the 4001, and the bottom trace is the current pulse into the flash lamp from the huge Arcotronics C20A2GR52007BSK 1kV 20µF 20milli-Ohm ESR capacitor, measured at 50A per division via a Pearson 410 current monitor.
The drive circuit, from my project logbook.
The complete test setup:
Expanded below, the critically damped 160A current pulse from the 20µF capacitor:
Expanded below, the 10kV output of the trigger transformer, ringing due to a lack of load except the 500M Ohm 15kV probe. I selected the series capacitor empirically; 100nF produced the required 10kV:
Next, I needed to see if the laser was actually firing - 1064nm is invisible, so I set up the Stellarnet EPP2000 SR spectrometer with a 1µm fibre optic (FO) cable pointing at the wooden base of the drawer, the reflection of the laser giving plenty of light; pointing it at the laser would have destroyed it instantly. The remaining spectra are from the flash lamp.
Success! I had a working MK367 Nd:YAG laser.