OmniOn Power, a leader in powering mission-critical AI, wireless, and industrial infrastructure, is introducing new high-efficiency, high-powered additions to its Osprey bus converter family.
Designed for AI data centers, high-density computing, 48V distributed power architectures, and regulated 12V intermediate bus applications, the QODN bus converters deliver higher efficiencies than the previous-generation QODE modules while preserving the same industry-standard quarter-brick footprint.
The new QODN Osprey bus converter models include:
- 2,600-watt (W) QODN217, with a 98.5% peak efficiency.
- 2,000W QODN167, with a 98% full-load efficiency at 54V and peak efficiencies exceeding 98.5%.
- 1,600W QODN136, with a 98.1% full-load efficiency at 54V and a 98.5% peak efficiency.
- 1,300W QODN108, with a 98.1% full-load efficiency at 54V and peak efficiencies exceeding 98.4%.
With accelerator servers, GPU baseboards, and high-density compute platforms driving sharp increases in board-level power demand, engineers require higher-power, higher-efficiency power solutions. In addition, there is an increased focus on reducing conversion losses to help preserve thermal margins and ease cooling constraints in dense, always-on AI and networking infrastructure, said the vendor.
“When looking at the power demands of AI data centers, even incremental efficiency gains can deliver significant cost benefits at scale,” said Philip Zuk, SVP and GM of OmniOn Power’s AI and data center business. “Our latest Osprey-series bus converters provide peak efficiencies up to 98.5% in the same quarter-brick footprint as our previous-generation Osprey bus converters, delivering up to a 1% efficiency improvement without increasing board-space requirements.”
The latest Osprey-series bus converters deliver high-current, high-efficiency 48/54V-to-12V conversion, helping address potential power conversion and thermal bottlenecks as AI server racks continue to scale toward higher power densities. In addition, the bus converters achieve high, flat efficiency curves across their operating band, meeting the needs of today’s AI workloads.
For more information about this news, visit https://www.omnionpower.com.