Supermicro PWS-2K20A-1R 2200W Power Supply

Overview

PWS-2K20A-1R is a Supermicro 2200W redundant power-supply module that appears in this set as a replaceable subcomponent inside Exxact barebones and GPU servers. The tickets are overwhelmingly break-fix or RMA cases, with this PSU usually handled as a barebone-manufacturer part rather than stocked as a standalone Exxact field-replaceable item. (#3481, #5639, #9177, #11064, #17107)

Known Issues

• power-supply-failure , dominant issue in this set, 13 of 13 tickets. Recurring symptoms include solid amber fault LEDs, no LED activity at all, no fan spin, no power-on, rebooting under load, or needing multiple plug-in attempts before startup. Failures often follow the PSU when cross-tested in another chassis or slot. (#17087, #17107, #19129, #22092, #37377, and 8 more)
• system-boot-failure , secondary symptom pattern. Several tickets begin as a server that will not boot or post, but troubleshooting narrows the root cause to one or more failed PWS-2K20A-1R units. (#3481, #5639, #5849, #11064)
• os-boot-failure , occasional downstream effect rather than direct PSU fault. In one case the PSU replacement restored power, but the customer still had a separate Linux boot issue that needed another ticket. (#22092)

Common Questions

• What failure pattern is most common for this PSU? Solid amber LEDs, no LED at all, no fan spin, or a system that will not power on are the most consistent symptoms in the evidence. (#17087, #17107, #19129, #20374, #22092)
• How is the PSU usually isolated as the root cause? Customers or agents commonly swap in known-good PSUs, move the suspect PSU to another identical chassis or slot, or compare behavior across outlets and power cables to show the failure follows the unit. (#17087, #17107, #18780, #19129, #5849)
• Can the system still partly function with a bad PWS-2K20A-1R? Yes. Some systems still expose IPMI or run temporarily with reduced redundancy, which can make diagnosis less obvious. (#5849, #9177)
• Is replacement usually advanced or only after return? Most tickets were return-first manufacturer workflows because the PSU is treated as a barebone subcomponent, but Exxact sometimes approved faster handling or advance replacement when uptime pressure was high. (#9177, #22092, #19129)
• Does Exxact repair these directly? Usually not. The common workflow is RMA intake, return to Exxact, and then manufacturer repair/replacement or RTV processing through Supermicro. (#11064, #17107, #19129, #20374, #37377)
• Can load trigger the failure even if the machine sometimes powers on? Yes. One ticket tied the issue to reboots under full 4-GPU load, and another described intermittent startup where power had to be plugged in multiple times. (#9177, #37377)

Related Products

• Supermicro redundant PSU families used in Exxact barebones, especially where the support path depends on barebone-vendor processing rather than direct Exxact spare stock. (#9177, #11064)
• Host systems using this PSU, including dense GPU and 4U rack platforms where a PSU fault may first appear as no-boot, reboot-under-load, or loss of redundancy rather than a clearly named PSU event. (#4135, #5849, #9177, #22092)
• Power distribution board or chassis power path as a common confusion pair, since at least one ticket considered PDB failure before choosing PSU replacement as the safer first step. (#5849)

Referenced by

• Power Supply Failure — issue affecting this product (×13)
• RMA Workflow — issue affecting this product (×13)
• System Boot Failure — issue affecting this product (×1)