Indubond Technologies S.L.

Engineering Leadership Vs Scaling Most scaling problems are not commercial problems. They are engineering problems. When companies struggle to scale, the first reaction is often: - change the sales strategy - add distributors - push harder on volume But in complex manufacturing, scaling usually fails elsewhere. What really breaks first is: - process robustness - repeatability - knowledge transfer - discipline in execution You can sell a machine once with a great pitch. You can’t scale a factory on promises. Scaling forces every weakness to surface: - unclear specs - fragile processes - undocumented know-how And that’s why growth feels painful not because the market is hard, but because physics doesn’t negotiate. The companies that scale successfully don’t move slower. They move with structure. When growth becomes difficult, where do you look first: sales… or engineering?

To all PCB fabricators, high-speed PCB designers, EMS and OEM engineers community !!! Aspect ratio limits in plated through holes and drill: technical limit or assumed limit? In next-generation complex PCBs (AI servers, HPC, high-speed systems), overall board thickness keeps increasing with multiple layers, 84L over 12mm thickness boards, while the pressure to reduce the diameter of plated through holes does not go away. From a design perspective, my understanding is that many designers would prefer plated through holes over segmented vertical interconnections (microvias, blind/buried vias, backdrill), because PTHs allow: -Higher routing density -Simpler stack-ups -A continuous and well-controlled vertical connection, which helps maintain signal integrity However, on the manufacturing side, the aspect ratio of plated through holes quickly becomes a major limiting factor, especially for very thick PCBs. In several industry discussions, I have heard claims that plated through holes with aspect ratios in the range of 25:1 to even 30:1 have been successfully metallized, although fabricators trends to avoid higher than 15:1. What I would really like to understand is whether this is: 1. A one-off or lab-level achievement 2. Something that can be industrialized with reasonable yield Or only feasible under very specific and constrained conditions I would like to open this discussion to practitioners with real production experience: 1. What is today the maximum aspect ratio for small plated through holes that is truly industrial and reliable? 2. When pushing to extreme ratios (≥20:1), where does the main technical limitation appear: copper distribution in the barrel, hole activation and chemistry, debris evacuation, geometrical control, long-term reliability… or a combination of all of them? 3. Do you see real room to further push this limit with process innovation, or are we already close to a physical ceiling? 4. At what point does cost and yield make vertical segmentation unavoidable? I am not looking for theoretical answers, but for real-world experience from manufacturing and design. I would especially value input from PCB fabricators, high-speed PCB designers, EMS and OEM engineers working in AI / HPC environments. Here we go! Who would like to jump in first? Thanks in advance for sharing your technical perspective.

After years of designing industrial equipment, one lesson keeps repeating itself. Specs don’t make machines reliable. Process understanding does. In critical industrial processes, hitting a spec once proves almost nothing. Repeating it, every day, on every machine, is the real challenge. Too many projects still chase: - higher speeds - higher pressures - better-looking datasheets While quietly ignoring the hard problems: - thermal behavior over time - process drift - real tolerances under continuous production Pushing limits is easy. Controlling physics is not. The gap between a “good-looking machine” and an industrial one is usually found where marketing stops and engineering begins. Peak specs sell machines. Long-term stability builds factories. Which one do you design for?

In my recent video I showed the lamination of a 100-layer stack. The purpose wasn’t to talk about registration, but to highlight a key point: if the internal bonding of the stack isn’t solid, uniform and stable, no alignment concept , pinless or not, can guarantee consistent results after lamination. For very high-layer builds (80L, 100L+), what truly defines success is the combination of a tight layer to layer registration with: • polymerization quality • thermal uniformity • mechanical stability • and repeatability of the full process. This is why at InduBond we have spent more than a decade developing our own induction-heating technology to achieve a robust and homogeneous internal bond, even in the most demanding multilayers. As the inventor of this technology, my experience is clear: alignment only matters when the bonding process is capable of keeping that precision through the whole lamination cycle. Registration discussions are important, of course, but they must always be supported by complete process data. Otherwise, theory doesn’t translate into real yield. If you want to see real data from super high layer count production builds, I’m happy to share results and field experience.

Pushing the limits of PCB technology. Today I’m excited to share a short video of an ultra high layer-count bonding process for 100-layer PCB stacks a key step for the new generation of AI servers and data-center infrastructure. At InduBond Technologies SL, we’ve been working for years to develop lamination solutions capable of delivering: • Extreme precision in layer registration • Uniform and ultra-fast induction heating • Stable bonding for very thick sequential builds • High reliability for advanced HDI and AI-driven architectures This 100-layer bonding sequence is a great example of what our technology can achieve when performance and consistency matter most. #InduBond #PCB #AI #Lamination #AdvancedManufacturing #InductionTechnology #HDI #DataCenters