Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
NUBM31 - 95W 20 Element array
#1
This post is intended as a short preamble for anyone who is exploring the idea of adding a NUBM31 laser to their build. If you have personal experiences with this hardware you would like to share, feel free to contribute.



Since around mid 2019 these arrays have been available via Ebay/Aliexpress. Originally intended for projection systems, they're one of the more promising candidates for cost effective metal 3D printing. Most of what you find on the market are counterfeit items but are adequate for our purposes here.The main draw backs of this system is the overall mass of the printhead, and the limitations that come with focusing 20 individual beams.



These lasers operate at around 33% efficiency (typical of diode lasers) and therefore output around 160W of waste heat at full power. This means they require a sizable heatsink and high power fan contributing to a bulky printhead. A consideration to bear in mind is that at 95W, the higher end of what is presently 'affordable', print speeds are quite slow (~20mm/s) and therefore this mass doesn't create a huge obstacle. In regard to what is achievable in terms of spot size, I have personally found a 60mm lens with a 60mm focal length will produce a spot size of around 0.4mm is possible. Using a larger lens is recommended to help avoid spherical aberration but will obviously contribute to the overall weight of the print head, therefore it is recommended that either a steel or aluminium bracket is used to support the assembly. A linear rail (e.g. MGN12H) is also useful to help minimize flex.






[Image: nubm31.png]





At 455nm, a visible beam is more practical for someone without a beam profiler and accommodates a wider array of metals given the level of absorption at that wavelength. The draw back being that you have to focus 20 beams as opposed to one leading to a lack of coherence and potentially a larger spot size. Sometimes you may find that an individual or multiple beams aren't aligned well and may contribute to a larger spot size also. From what I have read, genuine units posses better alignment but are only available second hand as the manufacturer will not sell them to the public. That said, it is fairly bad luck to get multiple beams that aren't well aligned with a counterfeit item. The beams themselves are linearly polarized which potentially means that absorption will be more effective along one direction of the metal grain. This is particularly relevant in the context of cutting. Circularly polarized light is optimal but is a topic in itself.






[Image: laser-Absorption.png]









The most practical means of powering these modules is with a constant current NUBM31 driver. In theory they should output around 85V @ 3A, I've observed about 93V unloaded however. They'll require a 12V power supply capable of delivering around 250W. Thanks to 3D printers these are abundant and quite cheap. The advantage these small SMPS units provide is TTL/PWM control which is useful for our sake. One caveat worth mentioning is that the TTL input is active low (despite advertising stating the opposite). This means as soon as the system is powered it'll turn on. This can be dangerous if you're not expecting it.




[Image: SMPS1.png]
Reply
#2
Love the research!!! Would it be possible to somehow combine the beams into a smaller continuous laser that can be transmitted around the kinematics with mirrors? This would be exactly like a CO2 laser cutter system, how the laser is mounted at the back of the machine and the laser beam is bounced off of three mirrors that reach the print head, and the beam is then focused to a much smaller spot.
Reply
#3
(01-29-2021, 03:17 AM)martin Wrote: Love the research!!! Would it be possible to somehow combine the beams into a smaller continuous laser that can be transmitted around the kinematics with mirrors? This would be exactly like a CO2 laser cutter system, how the laser is mounted at the back of the machine and the laser beam is bounced off of three mirrors that reach the print head, and the beam is then focused to a much smaller spot.

It's possible, though challenging. Condensing the beams into a smaller profile will result in added divergence assuming you use a beam expander. An anamorphic prism pair might be suitable but I am not 100% certain on that. There's also the possibility of doing it with a mirror system but that would be entirely custom I believe. Then you have the issue of not all beams being parallel to begin with.

I posted an article here that you might find interesting: https://metalmatters.co/showthread.php?tid=34
Reply
#4
This thread is great, answered the question I had about focusing the beam. At 95W of laser power it should be sufficient to melt many materials at around 300-400mm/sec, provided you can get the spot size down to less than 100um, preferably about 50um. I have experience with an Xact Metal XM200C that uses a 200W laser at 1095nm which has much worse absorption in most materials but is easier to get a tight laser beam profile as it is a single IPG laser unit.
Reply
#5
Isn't the main reason for using lasers to melt metals coming from the coherence of the laser light? With 20 lasers there would be no coherence between them so no way to actually get 95 watts of laser power to melt metal powders. Yet, you ARE melting stainless steel powder! So, if the case is coherence is not an issue then why use lasers at all? Why not simply use something like a 250-watt incandescent light, a fresnel lens, an optical integrator, and then do a final focus using a non-imaging concentrator?
Reply
#6
(07-10-2021, 11:05 AM)James Wrote: Isn't the main reason for using lasers to melt metals coming from the coherence of the laser light?  With 20 lasers there would be no coherence between them so no way to actually get 95 watts of laser power to melt metal powders.  Yet, you ARE melting stainless steel powder!  So, if the case is coherence is not an issue then why use lasers at all?  Why not simply use something like a 250-watt incandescent light, a fresnel lens, an optical integrator, and then do a final focus using a non-imaging concentrator?

I'm not an expert on this matter but will answer the question as best I can. One of the main reasons for using laser diodes is due to the conservation of Etendue. i.e. the projected image (focal spot) from a lens is bound to the size of the emitter and the divergence emanating from it. Diode lasers are able to produce a far greater amount of light per unit area than any conventional source hence with their small profile and relatively small divergence they make an ideal source for creating a high power density / small focal point. The problem with using something like a light bulb / LED is the emitter size is typically large and the divergence unconstrained. Any attempt to focus the light source using optics effectively increases the size of the emitter (think of how a magnifying glass behaves when in focus). You can form a small emitter with low divergence by masking a light source with a pinhole but obviously that drastically reduces power and forms a very lossy system.

On the topic of coherence, most of these high power laser diodes are "multimode" meaning that they lack ideal coherence to begin with. I believe the light still contains wave fronts but not as you would typically imagine. This is due to the cavity size required to produce higher orders of power. To create a single mode laser with perfect coherence I believe the resonant cavity cannot be much larger (if any) than the wavelength produced. This will obviously limit the amount of power a single diode can produce and doesn't yield much benefit in many applications. I suspect that the coherence that exists within a multimode diode laser is how it generates its "stimulated emission" and in turn generates such a high intensity - but don't quote me on that.

In terms of melting metal, all we need to do increase the temperature/raise the kinetic energy of the metal to it's sintering/melting point. I think that has more to do with power density than it does coherence e.g. you could use an incandescent light if there was some way to focus it finely enough. One thing I am not clear on is how destructive interference manifests at the focus point. I suspect it might be easier to disregard that and think about the energy transfer in terms of photons rather than wave fronts.


Attached Files
.pdf   Singlemodeandmultimode.pdf (Size: 190.35 KB / Downloads: 6)
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)