MES states that quotes for the same motherboard layout drawings can differ by up to 3 times, AI Data center power demand increases, attracting attention
PR Newswire
47m ago
Ai Focus
MES releases engineering guidance stating that for the same motherboard layout drawing, quotes from different suppliers can vary by up to three times, with the main differences stemming from the various manufacturing processes assumed by each supplier. This guidance is aimed at data center engineers and procurement teams, discussing the selection of processes, tolerances, edge quality, and alloy choices for copper and aluminum motherboards in high-density power supply applications.
Helpful
No.Help

Global manufacturing and procurement company issues guidance on batch demarcation points, tolerances, edge quality, and alloy selection for copper and aluminum busbars

Columbus, Ohio, October 27th / PRNewswire / -- MES, a global manufacturing, procurement, and supply chain company that serves Fortune 500 original equipment manufacturers, Inc. Today, it released engineering guidelines indicating that supplier quotes for the same circuit board design can vary by up to three times, depending on the manufacturing processes assumed by each supplier. These guidelines are aimed at data center engineers and procurement teams responsible for specifying copper and aluminum circuit boards for high-density power supply systems.

As the infrastructure requirements for AI increase, so do the associated risks. A liquid-cooled AI rack can have a power demand of up to about 120 kilowatts. At the 48 to 54 volt direct current voltage used inside many such racks, the current exceeds 2,000 amperes, which makes the busbars a critical component both within the rack and in the switchgear and busbar trays that supply power to it.

MES, President and CEO, Hiten Shah stated: "When we send the same master layout drawing to multiple suppliers, the quotes can vary by three times, and materials and profits cannot explain this difference. One supplier quotes based on stamping and bending, another based on stamping only, and yet another based on machining. Before these assumptions were considered, the buyer was not comparing the same part at all."

Batch decision-making on the starting point. MES provides an overview of the approximate demarcation points for four common manufacturing processes. CNC Machining usually results in the lowest total cost when the annual production volume is below about 200 units. CNC Stamping and bending are more advantageous for volumes between about 200 and 5,000 units. As the cost of molds is spread over larger batches, stamping becomes increasingly competitive beyond about 10,000 units. Extrusion is suitable for long aluminum strips with a constant cross-section. The complexity of parts, the number of variations, and the stability of the design can all affect these demarcation points.

The precision should be focused on the joint areas. This guidance suggests that strict control should be exercised in areas that affect electrical connections; for example, the flatness of the bolt contact area should be about 0.05 millimeters, and the key bolt hole positions should be within ±0.10 millimeters. For other parts, standard tolerances can be used. Applying strict tolerances to the entire part may force the manufacturing process to rely more on machining, but it will not improve performance.

For insulated busbars, the edge condition is very important. Deburring can remove burrs from the edges of stamping and punching, but it cannot eliminate flanging, broken areas, or hardened material from processing. Powder coatings and epoxy coatings may thin at sharp edges, and the electric field will also concentrate in these areas.

MES Project Manager Brad Layne stated: "The edge radius is one of the most common missing items on our insulation busbar drawings. If not specified, the edge condition depends on the process chosen by the supplier. Marking the radius on the drawing ensures that the coating performance remains consistent regardless of who manufactures it."

The choice of alloy can affect electrical conductivity. For extruded busbars that carry current, it is recommended to use 6101 aluminum alloy instead of the 6063 alloy used for structural purposes; 6101 can meet the minimum conductivity requirements as indicated by ASTM B317.

MES supplies copper and aluminum busbars by covering manufacturing networks in eight countries, and possesses CNC capabilities in stamping and bending, stamping, machining, and extrusion. Its engineering team reviews the drawings before mold opening to identify the most feasible manufacturing route.

Complete guidance, including the main circuit board drawing and the RFQ list, can be found at https :// mesinc.net / why-the-same-busbar-can-cost-3-more-a-manufacturing-process-comparison.

Regarding MES, Inc.

MES, Inc. Headquartered in Lewis Center, Ohio, it is a global manufacturing, procurement, and supply chain company that provides services for OEM in the fields of electrical and energy, automotive, robotics, lighting, as well as construction and agriculture. MES manages procurement, engineering, quality, and delivery within a global network of suppliers, with operations spanning the United States, Mexico, Poland, India, Vietnam, China, and other regions. MES has been certified according to ISO 9001:2015 since 2009 and has also obtained NMSDC certification. For more information, please visit https :// mesinc.net.

Tip
$0
Like
0
Save
0
Views 23
HQYC reminds readers to view blockchain rationally, stay aware of risks, and beware of virtual token issuance and speculation. All content on this site represents market information or related viewpoints only and does not constitute any form of investment advice. If you find sensitive content, please click“Report”,and we will handle it promptly。
Submit
Comment 0
Hot
Latest
No comments yet. Be the first!
Related
NVIDIA News: After investing in Meta, the chip manufacturer secures an even larger hardware order for AI
The Wall Street Journal reported that NVIDIA will invest $3 billion in the upcoming stock issuance of Meta. As part of the deal, Meta will also expand its use of NVIDIA hardware to support the next generation of agents behind its Muse AI products, including the Vera Rubin system and Vera CPU. Meta plans to raise up to $95 billion by selling stocks, which will be used for AI infrastructure expenditures; CNBC stated that AMD and Broadcom have also expressed non-binding interest in purchasing a total of up to $5 billion worth of Meta shares.
Coinpedia
·2026-10-27 18:50:53
6
BlackRock Collaborates with SBI and Startale to Explore Connecting Tokenized Funds to Japanese Yen Stablecoins
BlackRock signed a memorandum of understanding with SBI Holdings and Startale Group on October 27th to explore the connection between tokenized money market funds and the Japanese yen stablecoin JPYSC. All parties stated that this is still in the exploratory stage, and no actual product or launch date has been announced yet.
Coinpedia
·2026-10-27 18:50:51
5
The Flagship Institute Launches Website and Introduces AI Training Services for Public Sector
The Flagship Institute Announces the Launch of Its Website flagshipinstitute.org, Providing Artificial Intelligence and Applied Technology Training for Government Employees, Institutions, Contractors, and Organizations Supporting Public Sector Tasks. The offerings include online courses, live lectures, training camps, certification pathways, subscription-based training, and customized projects. The company states that these courses cover a range of topics such as large language models, prompt engineering, machine learning, retrieval-enhanced generation, agent workflows, responsible AI, cybersecurity, data engineering, and applied data.
PR Newswire
·2026-10-27 18:50:49
6
Ripple receives final approval from OCC to establish a trust bank, serving 260 million US dollars of RLUSD stablecoins
Ripple indicates that final approval has been obtained from the US Federal Reserve (OCC) to establish Ripple National Trust Bank. This bank will manage the reserves that support the stability of the RLUSD stablecoin, act as the custodian for related collateral, and provide digital asset custody services to enterprises.
Decrypt
·2026-10-27 18:40:33
13
Haut.AI Launches SkinFM Skin Basic Model, New Benchmark; SkinBench Reveals AI Skin Analysis Defects
Haut.AI releases its skin base model SkinFM and introduces a public benchmark SkinBench. The company states that tests have shown that six leading AI models, including Google, OpenAI, Anthropic, etc., often misjudge non-existent skin issues or fail to detect real problems, which may affect the accuracy of skincare advice and product recommendations.
PR Newswire
·2026-10-27 18:29:40
14
View More