MSI MEG Ai1300P PCIE5 PSU Review

Practical testing …

Let’s put the PSU through its paces with the following system:

Mainboard & BIOS MSI X670E Carbon WiFi 1.B0 AGESA
CPU AMD Ryzen 9 7950 X3D
Graphics card MSI GeForce RTX 4090 Gaming X Trio 24G
RAM Kingston Fury DDR5-6000 EXPO 2 x 32 GB
SSD WD Black SN750 NVMe 1 TB PCIe 3.0 x4
WD Black SN850X NVMe 2 TB PCIe 4.0 x4
Case Thermaltake The Tower 900
Cooler Custom water cooling powered by Aqua Computer
Airplex Quattro Radiator 480 + 4x BeQuiet Pure Wings 3 120 mm
+ 3 case fans 120 mm
D5 Next pump with Leakshield and RGB ring
CPU water cooler EKWB Velocity2
Quadro fan controller
Monitor LG Ultragear 34GP950G-B 3440 x 1440 @ 180 Hz Overdrive
Operating system Windows 11 23H2 22631.2715

Commissioning went smoothly. As the PSU keeps the fan switched off in idle mode and also in partial load mode, there is nothing to complain about here either. Let’s now turn our attention to the internal values, which should be more interesting for operation.

The voltages of the PSUs are in a very good range in idle mode. The 5 % tolerance is not even remotely utilized here.

However, the situation is different in the partial load range. Here we can clearly see that the voltages drop on the 12 V rail in particular, but also on the 5 V rail. With a load of less than 700 watts, we would have liked to see a little more stability in the voltages. These are the values that are read out with mainboard sensors.

But what does it look like directly in the PSU? The PSU connected via USB delivers the data directly to the MSI Center.

Unfortunately, real-time monitoring also paints a clear picture here, with the voltages of the 12V and 5V rails falling below their nominal values. At 12V it is -1.04 %, at 5V it is still -0.78 %.

The CPU ran at an output of 139 watts, while the graphics card reached a peak value of approx. 500 watts. The remaining power was drawn from various system components and resulted in a peak power of approx. 680 watts.

MSI MEG Ai1300p PCIE5 PSU Result and general impression …