AI 3D AOI + SPI for the largest boards
Horus MAX
Bigger boards.
Same intelligence.
Horus MAX inspects a 50 x 50 mm field of view in 0.6 s/FoV, exactly as Horus does, over a working area sized for boards up to 815 x 715 mm.
Horus MAX
Overview Configurations What it inspects Technology Programming
3 stages
SPI, pre- and post-reflow AOI
815 x 715 mm
board envelope, up to 15 kg
0.6 s
per FoV, 50 x 50 mm
Configurations
Built for heavy, thick assemblies
Available
Single Lane
Boards up to 15 kg and 10 mm thick, with 60 mm clearance above and below for power, industrial and server boards that outgrow standard formats.
- Board size up to 815 x 715 mm
- Thickness 0.5 to 10 mm
- Weight up to 15 kg
Specifications at a glance
| Single Lane | |
|---|---|
| Inspection | SPI, pre- and post-reflow AOI |
| Optics | 25 MP top camera, 4 RGB projectors, custom telecentric lens |
| Field of view | 50 x 50 mm in 0.6 s/FoV |
| Height resolution | 2 µm |
| Height measurement range | Up to 30 mm |
| Max. board (X, Y)* | 815 x 715 mm |
| Thickness (Min-Max) | 0.5-10 mm |
| Max. weight | 15 kg |
| Clearance top / bottom | 60 / 60 mm · 3 mm edge |
| Machine dimensions (W, D, H) | 1,620 x 1,780 x 1,776 mm |
*X=Length, Y=Width.
What it inspects
Inspect virtually every component
From paste deposit to finished joint, read by neural networks on high-resolution 2D and true-3D data together.
Body and placement
Component location, size and position, tilt and lift, with IPC-oriented acceptance criteria.
Neith readout Locations
Leads, solder and marking
Pin and lead presence and position, solder joint condition, bridges, planarity, solder balls and detection of display scratches, with IPC-oriented acceptance criteria.
Neith readout Pin solder joint 3D / Tall white connector 3D
Difficult geometries
Reflective surfaces and metallic pins. High-density fine-pitch. Deep shadows and tall connectors. Macro and non-standard applications.
Marking and orientation
Polarity, OCR and OCV, including rotated text and supplier variants, plus resistor colour-band identification.
Neith readout OCR / Polarity
Beyond the standard inspectors
Macro builds custom tests for geometries no standard inspector covers: heights, distances, coplanarity. Anomaly detection flags deviations nobody wrote a rule for, at component level and across the whole board.
Neith readout Anomaly detector
Step stencils
Automated inspection support.
Deposit consistency
2D/3D comparison across identical components, focusing on height.
Full-board 3D inspection
Detect unexpected paste deposits.
Board control
Pass/fail, quality flags, pass-through to AOI and repair-station inspection locks.
Neith readout SOIC smearing / Missing BGA and planarity / QFP peaking planarity
joint 3D
connector 3D
detector
smearing
and planarity
planarity
See what others miss
Underfill
UV-reactive underfill inspected with UV illumination, enabling 3D reconstruction to verify glue presence, distribution and applied amount.
Encapsulation
Encapsulant volume, height and coverage inspected in true 3D, so a component is confirmed protected, not just present.
LED
Position, coplanarity and die alignment inspection, even on reflective packages, with reliable identification of the LED’s true colour without interference from die fluorescence.
Pin-in-paste
Inspection of through-hole leads inserted into solder paste before reflow, verifying lead position and paste coverage during the same inspection cycle as SMT components.
Send us a difficult assembly.
We run it and send back what the platform has found.
The technology inside
The technology that makes it possible
True 3D. Patented optical head.
A 25 MP telecentric top camera, four RGB projectors and multi-level RGBW lighting, with optional polarised and side cameras. Built for reflective metals, shadowed zones, tall connectors and fine pitch.
AI in the Inspection Cycle
It decides during the inspection cycle, not after it. Proprietary neural networks recognise and classify board features in real time, with FPGA processing keeping pace with the line.
Neith Software Platform
One browser-based software environment for programming, data and workflows, across every inspection stage and every Delvitech platform.
Programming and learning
AI Programming and Learning
Training Manager
Your team can train the neural networks directly on the platform, using your own boards and production images. The machine learns to recognise components beyond its initial training while preserving existing knowledge, allowing your team to adapt inspection independently and keep know-how in-house.
Self-programming
Accelerates component recognition and recipe creation, reducing the engineering effort required to bring a new product to stable inspection.
Coming Soon
Predictivity
Proactive process control through neural pattern prediction, recognising what precedes a defect before one occurs.
Coming Soon
Automated Validation Station
The system autonomously analyses components and defects, validates recurring calls and uses AI to optimise inspection parameters, verifying each change against a pass/fail reference set.
Where it fits
Horus MAX is built for large, long and heavy assemblies across automotive and EV, Electronics Manufacturing Services (EMS), industrial and power electronics, aerospace and defence, medical devices, telecommunications, AI computing and data centres and consumer electronics, wherever several inspection stages run on one line and the team would rather learn one platform than three.
Brochure
Download the Horus MAX brochure
Enter your email and we will send you a link to download the brochure.
FAQ
What is the maximum board size and weight for Horus MAX?
Horus MAX supports boards up to 815 x 715 mm and 15 kg. The Single Lane configuration supports board thicknesses from 0.5 to 10 mm.
Which inspection stages does Horus MAX support?
It supports solder paste inspection (SPI) and pre- and post-reflow automated optical inspection (AOI).
What AOI defects and features can Horus MAX inspect?
It checks component location, size and position, tilt and lift, as well as lead presence and position, solder-joint condition, bridges, planarity and solder balls.
How does AI work during inspection?
Proprietary neural networks recognise and classify board features in real time during the inspection cycle, while FPGA processing keeps pace with the line.
Can a team train the neural networks using its own production images?
Yes. Teams can train the neural networks directly on the platform using their own boards and production images, and adapt inspection while keeping know-how in-house.
