<< Back to projects.bluefan.nl

One of the things I do is scavenge sites like ebay for laser parts, trying to find components for building a laser from scratch. Complete lasers are of course very nice to work with but a homebuild one is a much more fun challenge. High power lasers hardly make it to ebay, if they do they're usually priced accordingly. Interestingly enough laser component find their way easier and much cheaper.

That's how I found laser diode bars, laser diodes consisting of 19 individual emitters in a row. I already have a few 20W 808nm diode bars, one even new. I decided to use a used one for this project, if it hopelessly fails I wouldn't blow up my new one. These diode bars are only used in high powered sidepumped lasers so I went looking for matching crystals.

I found a 2x2x12mm Nd:YAG crystal, matching nicely to the 10mm wide emitter array. No clue about the doping but I'll go with it. I use the same cavity mirrors as the pulsed YAG laser, flat HR and flat 85% OC. I have a 500mm RoC HR, quite a curvature. I didn't choose careful enough when buying it years ago when some passed by on ebay. I'll get some higher RoC mirrors someday. But the first task is to put this randomly acquired hardware together.

Designing the crystal mount

The whole assembly designed in Blender

Designing a sidepumped laser is probably more involved that just putting the diode bar next to the YAG rod but I didn't do much more than that. I didn't want to create a hotspot by putting the diode bar emitter array very close to the YAG rod so I kept a 0.8 mm gap between the emitters and the YAG rod. The idea was to enclose the emitters and the YAG rod in a reflective cavity so that I wouldn't unnecessarily waste pump photons, that's why I milled a triangular snout in the heatsink with at the end the diode bar emitters. The snout should easily capture most of the vertical divergence of the emitters and reflect the rest. My estimate is that the bulk of the power hits the crystal direct on and creates a reasonable even illumination profile. Precise enough considering all the unknowns.

This side view shows the YAG rod in the pumping chamber with the snout on the right

Underneath the aluminum base plate will be a TEC to keep the whole assembly cool. By using plastic bolts the whole assembly is isolated, allowing a positive power supply for the case positive diode bar.

Building the laser

The assembly nearly complete, showing how the diode bar and the crystal holder sit next to eachother

The copper heatsink was milled using my homebuild milling machine. I'm satisfied with the result, I now have a nice tapered heatsink made to the crystal dimensions. I left the heatsink wider than needed to protect the crystal faces a bit when testing in my not-so-dustfree lab.

Showing how the YAG rod sits in the heatsink

I don't know if the unpolished side of the yag rod scatter strongly or are actually ment to diffuse the pump light, my guess and hope is the latter. An indium foil protects the YAG rod against the clamping of the heatsink and improves thermal contact. It may not reflect the pump light very well but I'll take that loss.

Initial testing

The completed laser

I don't have the temperature controller I plan to build for this laser yet, which is a problem. The wavelength of the diode bar is very temperature sensitive and needs to be matched to the absorption of the Nd:YAG. Also running without any cooling means running the risk of overheating the whole thing. Nevertheless I decided to push my luck for this initial test.

The first light measured out of my laser

And succesfully. For a brief moment the power spiked to 70 mW with somewhere around 20A through the diode bar, a truly horrible efficiency. But with no temperature optimalisation, stabilisation or even cooling, no chance to align the cavity properly and plane mirrors. But it's output, it's a laser built from loose components that works. Optimalisation will follow, which is much easier when there's lasing to start with.

Testing the Laser diode bar

Measuring the spectrum of the laser diode bar

One possible reason why the laser would not work properly is a malfunctioning laser diode bar. I checked the spectrum with an Optical Spectum Analyser (OSA).

Spectrum of the laser diode bar

The spectrum is centered properly around 808nm. This depends on the temperature of the laser diode bar. Even at the correct temperature for a 808nm center wavelength the laser did not work properly. The mirror set I use I got from ebay from China. Changing to mirrors I got from a UK seller not only did not work but it stopped any output alltogether.
Thermal lensing is strongly present, the HeNe aligment laser beam is visibly deformed after passing the laser crystal. This may cause an instable resonator.
The laser output only seems to be present when ramping up the power fast, this may indicate thermal issues, but a pulsed current did not work either.
One possibility is that this is not an Nd:YAG crystal but for example Nd:YLF. This explains the horrible efficiency and the disappearing of the output when using a good mirror set (the mirror set already proved itself with the pulsed YAG laser). Next item to check is the fluorescence spectrum of the laser crystal to see if the fluorescence is centered around 1064nm.

Deze site is gemaakt door: Johan Reinink