NF5688M6 OCP Retimer and Midplane Connector: Bent-Pin Case and Mitigation

1. Reported fault and affected parts

The source reports bent pins at the high-density connection between the OCP retimer board and midplane after inserting/removing an NF5688M6 motherboard node. The resulting fault can include no display at power-on. Where site conditions permit, its handling method below is intended to reduce connector misalignment during node insertion. This is a historical case, not evidence that every no-display fault has this cause.

PartSource part number
OCP retimer boardYZRI-02202-102
Midplane boardYZTC-02206-101

Overall location and the mating connection:

Localized overview showing the OCP retimer and bent-pin connection location; faint original author watermark retained
Original photograph of the retimer-to-midplane connection
Original close-up showing visibly bent connector pins

Midplane board and damaged-connector detail:

Localized midplane photo identifying the same damaged connector with a red circle
Original close-up of the midplane pin damage

OCP retimer location and side view:

Localized motherboard photo identifying the OCP retimer board location
Original side view of the raised OCP retimer connector

2. Source handling method

Prepare and isolate the server using the matching service procedure before handling internal boards. The source’s temporary method opens the top and middle covers and checks whether the two mating connectors are at the same height. If misaligned, a trained service engineer adjusts the retimer connector height by gently pressing or lifting the board as applicable, then slowly inserts the motherboard node while observing alignment. Do not force mismatched or already damaged connectors together; this alignment method does not straighten or repair bent pins.

Original chassis top-cover view from the source handling example
Original opened middle-cover view exposing the board interface
Localized manual-alignment callout with original connector boxes, arrow and timestamp retained

For a persistent recurrence, the source says a 1U-base rework had been initiated and a new spare-part number would be requested; replacement of the motherboard tray and 1U base was its longer-term remedy. It does not supply an approved new part number, current stock status or a torque specification. Confirm the applicable revised assembly with the manufacturer rather than treating the two original board PNs above as the revised tray/base PNs.

The source design slide describes the mechanical change: add a fixing stud on the 1U base and clearance in the motherboard tray; slide the OCP retimer card into its guide in the pictured direction; secure it with the SMT spring screw. The card’s original underside includes a copper-clad sliding area. Fixing the card to the 1U node base is intended to reduce the influence of tray deformation. The source reports production-line validation; this translation has not independently verified the rework’s field outcome:

Localized five-view retimer fixing-design slide preserving the stud, guide arrow, SMT spring screw and underside copper strip

3. Reported root cause

Transport or repeated motherboard-node removal/insertion can slightly deform the motherboard tray. The OCP retimer can then lift enough to misalign with the midplane connector, allowing insertion to bend the midplane pins. Check both the mechanical alignment and connector condition rather than repeatedly pushing the node into place.

4. Recommendations and summary

The source’s final heading has no additional prose. Its actionable distinctions are the temporary alignment observation and the manufacturer-confirmed revised tray/base remedy for recurrence. Retain the exact server configuration and service authorization when applying the case; no physical work was performed during translation.

References

Can't find an answer?

Upload a log for AI diagnosis, or ask the AI assistant directly in the console.