Failure Analysis of Mk2 30W Modules

As many in our community know, the LaserMATIC Mk2 has suffered from quite a number of 30W module failures since launch. The failure rate peaked in March of this year, to as high as 8%, due to a particularly problematic batch of modules that were produced at the end of 2024.

To tackle this issue we got around 70 modules back to our factory and analyzed 43 of them. This is a summary of our findings. The upshot is that there are two main issues with the 30W module: 1) Less-than-ideal air flow and lack of filters leading to soot buildup, and necessitating frequent maintenance; and 2) Issue with particulates contamination in the production process. Let's start with a pie chart showing root causes as we get into the details:

Fan and Air Flow-Related Issues

After each module cleaning session, the orange cooling fan has to be put back on top of the module. During this step, the delicate fan wires can often get damaged, pinched or torn from its connector. These failures account for 13 out 15 modules analyzed that have fan-related issues. In each case, the damaged wire caused the fan to stop spinning, either by creating a short and burning out the fuse on the driver board, or by creating a disconnect between the board and the fan. This in turn causes the module to overheat and some of its diodes to fail. To clarify, permanent damage is done to the diodes and the driver board fuse. All the fans, after their wires are repaired, still spin normally.

Putting wire thickness, routing and insulation material choices aside, I believe the best way to solve this issue is to sidestep the need to remove the fan during maintenance altogether. Three potential approaches to accomplish this are:

• Including an inline fan by default with the machine. Exhausting as much of the smoke as quickly as possible is the most straightforward way to reduce soot buildup in the machine.
• Adding a filter to the top of the module, similar to what is done with HTL Custom's filter box mod.
• Adding a removable front panel to the module, so that when the module does need to be cleaned, there is no need to remove the fan and go in from the top.

This way we can avoid the trap of: 1) Lots of soot/dust collects inside module --> 2) Fan needs to be removed during cleaning --> 3) Fan wire damaged and fan stops spinning --> 4) Module damaged. We are also working on a module cleaning and maintenance video which will include tips on how to properly handle the wires when putting the fan back on the module.

There were also 2 modules where the fan itself was physically damaged, with the blade detaching from the orange casing.

Diode Failures

Of the 43 modules we analyzed, 15 had working fans and driver circuit boards, but were still down on power or were missing diode "lines". We deemed these to be diode failures and, in coordination with our laser diode vendor (Shenzhen Lasor Photoelectric), tried to get to the bottom of these failures. In order to identify the root cause, 21 bad diodes from 10 modules were disassembled, cross-sectioned and examined under a microscope.

The images below provide some context on where these diodes reside inside the module, and where the cross section was made:

The following images show the three failure modes uncovered. 8 out of the 10 modules examined exhibited Failure Mode 1. According to Lasor, this failure mode correspond to particulates contaminating the diode output surface during production, before the TO-9 packages are sealed. Similar to dust gathered on protective lenses which you may be familiar with, these tiny particles will cause cracks, but minuscule ones on the diode output facet. There is also one module with damage from reflected light when engraving shiny or reflective materials. We also tested one diode from a module with a non-spinning fan, and this confirmed that the failure is from excess heat.

Lasor has informed us that they've moved the concerned process in house where they have better control over the clean room and particulate levels, so Failure Mode 1 should be greatly reduced going forward. We still need to verify this, and also need to look into whether it's possible to inspect for contaminants at the diode level before putting them into modules. These items are on our to-do list.

Other Causes

There are three other causes that are worth mentioning:

• 3 out of 43 modules analyzed had issues relating to the power selector switch. Two had dust built up in them that caused them to jam; One had an exposed solder joint that came in contact with the metal frame of the laser module, causing the module to emit light as soon as the machine was turned on.
• 2 out of 43 had dirty focus lenses. This is the lens that is meant to be protected by the flat window lens. Both modules performed up to par after the lenses were cleaned.
• 2 out 43 had bad driver boards, with corroded/damaged solder contacts from soot buildup. After cleaning off the soot and re-soldering a few contacts, both modules were back to normal. This again shows the importance of managing dust and soot build up inside the module.

Finally, 5 out of 43 modules appear to have no issues at all, other than perhaps needing a wipe on the exterior window lenses to perform up to standard.

Corrective Actions Going Forward

We now have a good understanding of the root causes behind Mk2 30W module failures, but addressing them will still take some time. The most immediate things we can do are:

• Sourcing a batch of inline fans and including them with LaserMATIC Mk2 purchases, starting from mid July. Customers who place orders starting today will also receive inline fans around that time;
• Continuing to send replacement modules to customers who encounter issues;
• Making a how-to video on module maintenance, including instructions on how to reassemble the fan.

These are not very technical fixes but I believe they'll go a long way in mitigating the issue, and bridge the gap before we roll out more permanent fixes in the form of Mk2 upgrade kits and the LaserMATIC Mk3, which as of now, is still going to be air-cooled.

Thank you for giving this update an read. If you have any questions or suggestions, please feel free to leave a comment below. As always, we can also be reached at support@rolyautomation.com.