PNY GeForce RTX 2080 Ti XLR8 in test – with 300 Watt limit, reference board and sufficient reason

1 - Unboxing, Daten, Testsystem 2 - Teardown und Analyse 3 - Benchmarks und OC 4 - Leistungsaufnahme im Detail 5 - Temperatur, Takt, Infrarot 6 - Lüfter und Lautstärke 7 - Zusammenfassung

Tear Down and Board Analysis

PNY uses the original reference board of the GeForce RTX 2080 Ti FE for the card. With this card, Nvidia has reached deep into the trick box in the board layout and especially in the power supply. We count three phases for memory and also see the appropriate PWM controller on the front. I will discuss the exact components of the entire board in more detail. If you now count the remaining voltage transformer circles, you get to 13 pieces. However, since there is no PWM controller that could realize this (why?), it must have been solved differently.

On the back is the same PWM controller again, but this time for the eight GPU phases. You read correctly, it's eight phases! 5 phases are operated in parallel with two voltage converter circuits per phase and three phases with only one each. This results in exactly what PR likes to sell us as 13 phases, even if it is not at all true. The three additional memory phases are generated by another uP9512P in 3-phase mode.

PCB Front

PCB Back

Let's start with the most interesting part! The fairly new uP9512P is used as an 8-phase PWM controller specifically designed to provide high-precision output voltage systems for the latest generation of GPUs. The uP9512P has programmable output voltage and active voltage positioning functions to adjust the output voltage depending on the load current, so that it is optimally positioned for a good load current transition.

The uP9512 supports NVIDIA Open Voltage Regulator Type 4i+ with PWMVID function. The PWMVID input is buffered and filtered to create a very accurate reference voltage. The output voltage is then precisely controlled on the reference input. The integrated SMBus interface offers enough flexibility to optimize performance and efficiency and also to connect the appropriate software. The controller also supports new Smart Power Stage chips (PLCs). Appropriate PLC then provide very accurate information about e.g. currents (IMON) and temperatures (TMON).

One feature of the uP9512P is the direct parallel connection of several voltage converter circuits without the usual doublers, since due to the necessary direct communication with the PLC no doubler chips can be used. If you would like to find out more details about this type of power supply and the improvements at Turing, please refer to our Investigative Article "Nvidia GeForce RTX 2080 Ti – Internal Details on Power Supply, Deviating Components and Where the Spikes which is always worth reading. There you will also learn more about the new Smart Power Stages, which replace the traditional, individual VRMs. The following table contains the most important components:

GPU Power Supply
PWM Controller	uP9512P UPI Semiconductor 8-phase
Gate Driver	not needed
Vrm	13x FDMF 3170 ON Semiconductor Smart Power Stage
Coils	14x Ferrite Choke 220 mH
Memory and power supply
Modules	MT61K256M32 Micron 11x 8GB GDDR6 SGRAM modules 2 Channels x 256 Meg x 16 I/O 2 Channels x 512 Meg x 8 I/O 14Gb/s
PWM Controller	uP19512P UPI Semiconductor 3 phases
Vrm	3x FDMF 3170 ON Semiconductor Smart Power Stage
	Encapsulated Ferrite Choke 470 mH
Other components
Bios	25WP080 Eeprom Single BIOS
Shunts	1x coil (smoothing) and shunt per supply connection (3)
More details
Other Features	– 2x 8-pin PCI-Express connectors for power supply – Filter coils in the entrance area

Cooler and backplate in detail

The actual radiator structure is very simple and weighs almost ultralight 607 grams! A mounting and stabilization frame on the board is not found, the cooler is held by the four screws on the GPU socket and two more on the left-hand voltage converters, that must be enough. Whether the light weight is sufficient, however. For this purpose, PNY relies on an extra cooler for the right-hand voltage converters and coils.

The cooling fins are narrow, but the available cooling surface is only medium-sized due to the low installation depth. In addition to the copper heat sink for the GPU, memory and the left-sided voltage converters are directly cooled via a circulating heat sink made of light metal.

The baking plate made of blackened aluminium does not cool anything, but in return has been provided with extra air holes. Pure optics, but no contribution to thermal well-being.