MSI RTX 5090 RMA Report: Per-Pin PSU Monitoring Flags Broken 12V-2x6 Sense Pin

What the owner reported​

According to the Reddit post, the first RTX 5090 RMA did not fix the owner's original problem. After a second return, MSI sent a different RTX 5090 model. The owner said the replacement initially showed none of the earlier card's errors.

They then bought an MSI MPG Ai1300TS PCIE5 power supply, hoping that matching MSI parts would give some peace of mind. The monitoring software then showed trouble:

• Zero current on one pin: Wccftech's write-up says the Redditor saw one pin carrying no current at all as soon as he started monitoring. Neowin's account identifies it as pin 2.
• An imbalance alarm under load: The owner says the software raised an alarm within seconds of putting the GPU under load. A screenshot they shared still read "Normal" because it was taken just before the alarm.
• Cable and port swaps changed nothing: A different cable and a different PSU port gave the same result.
• A raised connector: The GPU's power port looked slightly lifted from the board when the cable was plugged in.
• A broken sense pin: When the owner unplugged the cable to put an older RTX 3080 back in, they found a broken sense pin. They say this explained the raised connector.

The owner now expects a third RMA. They say the repeated returns will leave them without a GPU for about nine weeks in total. They also said that without the monitoring they would probably have assumed the card was fine and blamed their own cable seating.

What the evidence does and doesn't show​

This is a single, uncontrolled user report. MSI has not confirmed it, and nobody independent has inspected the card. The post doesn't say when or how the sense pin broke. It also doesn't prove that the zero-current reading came from that broken pin.

The cautious reading is that per-pin monitoring drew the owner's attention to something abnormal that they say they had not noticed. The report does not show that the PSU stopped a connector from melting. It doesn't show that the card would have overheated if left running, and it doesn't show that Safeguard+ repairs anything. Neowin's own conclusion is similar: such features don't fix a faulty connector or remove the underlying risk of the 12V-2x6 interface, but they can help users spot a problem sooner.

How MSI's Safeguard feature works​

The feature is built into the power supply, not the graphics card. MSI says GPU Safeguard measures current per pin inside the PSU, with an independent shunt resistor on each 12V-2x6 pin group. The small voltage drop across each resistor is amplified and digitized, so the PSU can evaluate each pin's current continuously.

MSI sells two tiers:

• GPU Safeguard (basic): On abnormal current the PSU sounds a loud buzzer. If the user doesn't intervene within three minutes, it cuts power to the 12V-2x6 rail.
• GPU Safeguard+: It is available on the MPG Ai1600TS PCIE5 and MPG Ai1300TS PCIE5 and adds enhanced real-time monitoring. If an imbalance is detected, the PSU beeps and MSI Center shows a warning to check the connector seating or the cable.

The Afterburner integration goes further. When abnormal 12V-2x6 current is detected, Safeguard+ tells Afterburner to show a warning and lower the graphics card's power limit to 75%. MSI says this needs Afterburner v4.6.7 or newer. HWiNFO's beta builds also show per-pin 12V-2x6 current on these MSI PSUs and support abnormal-current alerts.

What to do if you see an alert​

MSI's guidance is clear. If current is abnormal or imbalanced, shut down immediately, then check the 12V-2x6 connector installation and inspect the cables for faults. The alert is a reason to stop, not a sign the system is safe to keep running.

A sensible troubleshooting order, based on that guidance and on what the owner tried:

1. Power down and unplug the PSU.
2. Reseat the 12V-2x6 cable fully at both the GPU and the PSU end.
3. Inspect the cable and both connectors for bent, recessed or broken pins, including the small sense pins.
4. Try another cable or PSU port. The owner did, and the fault stayed, which pointed toward the card.
5. If the fault follows the GPU, stop using it and contact the vendor about a warranty claim.

Limits and caveats​

Safeguard+ is a diagnostic and mitigation layer. MSI describes it as an addition to the conventional PSU protections. It is not a cure for the 12V-2x6 design's thin safety margins.

The monitoring also depends on hardware. The Ai1300TS product page says Safeguard+ monitors the dual 12V-2x6 connectors and gives real-time alerts. MSI lists support for Nvidia GeForce RTX 50 series cards. If you run a different PSU, you won't get this per-pin visibility unless it has comparable sensing. Neowin also mentions other mitigations, such as an Asus current-balancing cable, but those weren't verified here.

Neowin's wider claims about DLSS 5 stressing GPUs and a user-built monitoring app were not corroborated here, so treat them as the outlet's framing. The practical takeaway is narrower: a card that behaves normally in games can still have a connector fault. Seeing per-pin current can turn a vague worry into something you can act on before heat damage sets in.