» Intel® Core™ Ultra 5-125H/ Ultra 7-155H processor (code name: Meteor Lake-H, TDP 28W)
» Fanless embedded system
» 2 x M.2 2280 PCIe Gen4 x4 (M Key) for SSD
» 1 x M.2 2230 (E Key) slot for Wi-Fi 7
» 1 x MXM 3.1 Type A slot (signal PCIe Gen5 x8)
» 2 x DDR5 5600MT/s SO-DIMM slots, up to 64GB
Compact AI Embedded System with Flexible Expansion Options
Streamlined Integration for Complex Needs
To address the growing complexity of data and the spatial limitations of medical environments, IEI introduces the HTB-300-MTL-H, a next-generation medical-grade AI embedded system. With an ultra-slim 1U design (only 4.4 cm), it is currently the most compact system of its kind in the Taiwan market. Powered by the latest Intel® Core™ Ultra series processor the first to integrate a built-in NPU (Neural Processing Unit), it effectively offloads AI inference workloads from the CPU, fulfilling the dual demands of high performance and compact size in medical applications.
Intel® Core™ Ultra 5-125H/
Ultra 7-155H
MXM Type A Module
Compact 1U Size
Fanless
Dual Power Supplies
Expansion PCIe Gen4 Slot
USB Type-C with DP
Wi-Fi 7
Intel® Core™ Ultra Processor with NPU
-
Time-Optimized AI Workloads
NPUs adopt Multiply-Accumulate (MAC) units hardware blocks designed to efficiently process tensor operations in neural networks. In the chart, the NPU completed 20 iterations in 20.7 seconds, running nearly twice as fast as the CPU (43.3s) and approaching the performance of the GPU (14.5s). -
Power Reduction for AI Performance
With only 10W power consumption, the NPU operates with far lower power than the CPU (40 W) and GPU (37 W). Compared to the two, it requires less power for AI tasks, making it more suitable for embedded systems. -
Efficiency Gains in Inference Tasks
Unlike GPUs, NPUs are purpose-built for inference using pre-trained models. While not optimized for full compute workloads, they still deliver clear advantages in inference tasks, achieving 7.8× higher efficiency than the CPU and about 2.6× greater efficiency than the GPU.
Stable Diffusion v1.5 (20 Iterations)
42 Inferences:
Text Encoder (1) + Unet+ (20) + Unet– (20) + VAE Decoder (1)
All CPU
All GPU
Unet + and - on NPU
20 Iterations (Sec)
| Hardware | CPU | GPU | NPU |
|---|---|---|---|
| Function | Primarily used for executing operating systems, various instructions, and applications | Mainly responsible for image rendering and equipped with computing power for applications such as AI. | Optimized AI accelerator for neural network inference and hardware level algorithm acceleration. |
| Application | Web browsing, code compilation, document editing | AI model training, image processing, graphics rendering | Facial recognition, edge computing (on mobile devices) |
| Hardware | Function | Application |
|---|---|---|
| CPU | Primarily used for executing operating systems, various instructions, and applications | Web browsing, code compilation, document editing |
| GPU | Mainly responsible for image rendering and equipped with computing power for applications such as AI. | AI model training, image processing, graphics rendering |
| NPU | Optimized AI accelerator for neural network inference and hardware level algorithm acceleration. | Facial recognition, edge computing (on mobile devices) |
Unlock AI Scalability with MXM Design
MXM (Mobile PCIe Express Module) is a high-performance GPU module standard that balances compute power and compact size ideal for embedded AI systems where space and reliability are critical.
- Easy
Maintenance
With only 10W power consumption, the NPU operates with far lower power than the CPU (40 W) and GPU (37 W). Compared to the two, it requires less power for AI tasks, making it more suitable for embedded systems. - Flexible GPU
Card Integration
With only 10W power consumption, the NPU operates with far lower power than the CPU (40 W) and GPU (37 W). Compared to the two, it requires less power for AI tasks, making it more suitable for embedded systems. - Efficient
Thermal Management
With only 10W power consumption, the NPU operates with far lower power than the CPU (40 W) and GPU (37 W). Compared to the two, it requires less power for AI tasks, making it more suitable for embedded systems.
Type A
Supported GPU series
|
|
|
| Model Name | MXM-A500 | MXM-A1000 | MXM-A2000 |
| Module Power Consumption | 25W–40W | 35W or 60W | 60W |
Active or Passive Cooling Always Efficient
Optional Fan Module
High Efficiency Fan for MXM
- Airflow
Intel i226IT Dual LAN RJ45 Network ensures stable, high-speed connectivity for industrial use, providing network redundancy, reducing downtime, and improving reliability in mission-critical environments.
NOTE: For ordering information, please contact your sales
representative.
Fanless
Large Area of Heat Fins
- Silently
Without fan, the system can run silently and be a quiet partner with patients. Heat fins are metal fins that increase the surface area for heat dissipation, improving the cooling performance.
Metal Chassis
- Low Failure Rate
The solid construction makes it highly durable.
The housing is entirely made of metal, enhancing cooling, durability and strength.
Versatile USB Interface Integration
The system provides 3 types of USB, delivering flexible, high-speed connectivity for a wide range of peripherals.
- Front USB
Ports for Easy Access
Provides front USB access for peripherals like keyboards and barcode scanners, aligning with healthcare staff habits. - USB 3.2
Gen 2 Type-A Supports 10 Gbps Transfer Speed
Enables high-bandwidth applications such as external storage, high-resolution cameras, and real-time data streaming with reduced latency - USB 3.2
Gen 2 Type-C Port with 4K Output via DP 1.2
A single port integrates high-speed data transfer, DisplayPort 1.2 for 4K video output, 10 Gbps data transfer speed and power charging, streamlining connectivity for both display and power.
|
Data transfer speed |
USB Type A | USB Type C | ||
| USB 2.0 | 480 Mbps | - | ||
| USB 3.2 | Gen 1 5 Gbps |
Gen 2 10 Gbps |
Gen 1 5 Gbps |
Gen 2 10 Gbps |
Dual Power Input offers Installation Flexibility
The system provides both Terminal Block and DC-In Jack as selectable power input interfaces. This design enhances deployment flexibility across varied installation environments.
Medical Devices
System
Cabinet System
System
Note: The terminal block is only available when the system
is configured with an MXM module.
Power Supply Devices
Seamless Connectivity for Complex Environments
Four RS-232 ports can conveniently connect to various physiological measurement and monitoring devices.
Patient monitor
Vital sign measurement station
Mobile vital sign monitor
Potential Equalization Conductor
The Potential Equalization
Conductor minimizes voltage differences between accessible metal parts to
prevent electric shock.
It is used in medical environments where multiple
devices are interconnected, ensuring protection for patients and staff.
PCIe Expansion for More Possibilities
Enables expansion of add-on cards in space-limited systems via full-height, half-length PCIe slot support.
*A riser card is required to use
the PCIe slot.
Full Height, Half-Length
PCIe Card
PCIe Card
PCIe Form Factor
Supports Add-On Cards Expansion
*Not available when MXM is installed due to space limits.
*NOTE: The accessory shown is for illustration purposes only. For ordering information,
please contact your sales representative.
*NOTE: The accessory shown is for illustration purposes only. For ordering information,
please contact your sales representative.
High-Speed DDR5 and M.2 NVMe Storage
- Dual-channel DDR5 SO-DIMM slots
System max capacity
64
GB
Speed
5600
MT/s
- Dual-M.2 2280 PCIe Gen4 x4 SSD slots
High-Speed DDR5 and M.2 NVMe Storage
- Dual-M.2 2280 PCIe Gen4 x4 SSD slots
Supports RAID 0/1 for performance and data redundancy, ensuring backup protection in critical applications.
- Dual-channel DDR5 SO-DIMM slots
The Latest Wi-Fi 7 Module (optional)
Wi-Fi 7 is the next-generation wireless standard that improves upon Wi-Fi 6/6E with significantly faster speeds, low latency, and enhanced reliability making it ideal for connected medical environments where real-time performance is critical.
| Model | AP capacity | Frequency band | Maximum throughput speeds | TWT |
| Wi-Fi 7 (802.11be) |
Multi-link operation (MLO) | 2.4 / 5 / 6 GHz |
Up to 46 Gbps
|
Supported |
Multi-Link Operation
- Extremely Lower Latency
Simultaneous connections across multiple bands improve reliability,latency, and throughput.
Ultra-High Throughput With 320 MHz Channels
- Wi-Fi Band Performance Comparison
* Wi-Fi 7 doubles the channel width to 320 MHz for
improved
throughput
Power Efficiency with TWT
- Extended power life
Model
Wi-Fi 7 (802.11be)
AP
capacity
Multi-link operation (MLO)
Multi-Link Operation
- Extremely Lower Latency
Simultaneous connections across multiple bands improve reliability,latency, and throughput.
Frequency band
2.4 / 5 / 6 GHz
Ultra-High Throughput With 320 MHz Channels
- Wi-Fi Band Performance Comparison
Maximum throughput speeds
Up to 46 Gbps
TWT
Supported
Power Efficiency with TWT
- Extended power life
Ultra-Compact 1U Size
With a 1U (44 mm) height and just 2 kg in weight, it easily installs on medical carts, beside equipment, clinic, or integrates directly into medical devices, ideal for tight spaces without sacrificing computing performance.
Four Mounting Options, Built for Any Setup
Flexible mounting：Supports VESA (75x75mm / 100x100mm), wall,
desktop, and DIN rail
Easy deployment：Speeds up integration in medical and industrial
settings
Desktop – Horizontal
Desktop – Vertical
DIN Rail (Optional)
VESA-Mount (Optional)
I/O Interface
Dimensions
Hardware Spec.
| System | |
|---|---|
| Processor | Intel® Core™ Ultra 5-125H/ Ultra 7-155H processor |
| Supported OS | Windows 11, Linux Ubuntu |
| Network | |
| Ethernet Controller | Intel® I226-V |
| Wi-Fi | IEEE 802.11 be, Intel® Wi-Fi 7 BE200 (optional) |
| Bluetooth | Bluetooth V5.4 (optional) |
| Storage | |
| Storage | 2 x M.2 2280 PCIe Gen4 x4 M-key slots (optional, supoorts RAID) |
| RAM | 2 x 262-pin 5600MT/s dual-channel DDR5 SO-DIMM slots (system max. 64GB) |
| Expansion Slots | |
| Expansion Slots |
2 x M.2 2280 PCIe Gen4 x4 M-key slots for SSD 1 x M.2 2230 E Key slot (for Wi-Fi; PCIe/USB) 1 x MXM 3.1 Type A slot (signal PCIe Gen5 x8) 1 x PCIe Gen4 x4 slot (support Capture card) |
| Audio | |
| Audio | 1 x 3.5mm Mic in & Audio out |
| Speaker | 1 x 3W speaker |
| I/O Interface | |
| Front I/O |
1 x 19V DC-in 1 x 3.5mm Mic in & Audio out 1 x Terminal Block (DC IN) 2 x USB 2.0 Type-A |
| Rear I/O |
1 x HDMI output 2.1 1 x USB 3.2 Gen 2 Type-C (DP1.2 + USB 5V/3A) 2 x 2.5GbE LAN 4 x USB 3.2 Gen 2 Type-A 4 x RS-232 |
| Indicator&Buttons | |
| Buttons |
1 x System power on/off button (with power LED) 1 x Clear CMOS 1 x AT/ATX switch 1 x Reset |
| Physical Characteristics | |
| Color | White |
| Thermal | Fanless |
| Chassis Construction | NCT(N), SECC |
| Construction Material | Aluminum |
| Mounting |
Wall, Stand and Arm VESA 75/100 compliant Din Rail mount |
| Weight (Net) (kg) | 2.04 |
| Weight (Gross) (kg) | 4.26 |
| Dimensions (LxWxH) (mm) |
232 x 202 x 43.5 (Standard) 232 x 202 x 66.5 (MXM or Riser Card) |
| Power | |
| Input | Terminal block 9-28V DC-in |
| Power Adapter | 19V DC-in, 120W medical power adapter |
| Environment | |
| Operating Temperature | 0°C – 40°C |
| Storage Temperature | -20°C – 60°C |
| Humidity | 10% – 95% (non-condensing) |
| Vibration | 1G |
| Operating Shock | 5G peak acceleration (11ms duration) |
| Non-operating shock | 15G peak acceleration (11ms duration) |
| Certifications | |
| Safety & EMC |
CE, FCC Class B Part18 UL 60601-1 IEC/EN 60601-1: 2005+AMD1:2012 (Edition 3.2) IEC/EN 60601-1-2 (Edition 4.1) |
Ordering Information
| HTB-300-MTL-H-ULTRA7-R10 | Fanless embedded system, Intel® Core™ Ultra7-155H 4.8GHz on-board processor, 2 x M.2 2280 PCIe Gen4 x4 M key slot, 1 x M.2 2230 E Key slot for Wi-Fi 7 module, 1 x MXM 3.1 Type A/B expansion, R10, RoHs |
|---|---|
| HTB-300-MTL-H-ULTRA5-R10 | Fanless embedded system, Intel® Core™ Ultra5-125H 4.5GHz on-board processor, 2 x M.2 2280 PCIe Gen4 x4 M key slot, 1 x M.2 2230 E Key slot for Wi-Fi 7 module, 1 x MXM 3.1 Type A/B expansion, R10, RoHs |
Package Contents
| Package Content |
|
|---|
