Tech

ASRock Z690 Taichi & Z690 PG Velocita Motherboards Review – Active Cooling Goes Brrrr!

Intel’s Alder Lake-S desktop CPU platform will feature support on LGA 1700 socket which made is a brand new socket design. The Alder Rocket Lake-S processors will be launching alongside the 600-series motherboards and won’t be compatible with existing LGA 1200 socket motherboards such as Z590 and Z490. In addition to the flagship Z690 motherboards, B660 motherboards will support memory overclocking which has been highly anticipated by budget PC builders.

Intel Z690 Platform Features:

The 11th gen desktop platform has a range of new features that mainly include:

  • Support for LGA 1700 Intel Core / Pentium Gold / Celeron Processors
  • TDP Support Up To 125W
  • DDR5-4800 / DDR4-3200 Support For Alder Lake Desktop Processors
  • Support For Intel Wireless-AX (CNVi)
  • Up To 16 PCIe Gen 5.0 Lanes (Discrete GPU via CPU Only)
  • Up To 12 PCIe Gen 4.0 Lanes
  • Up To 16 PCIe Gen 3.0 Lanes
  • 8x SATA III ports (6 Gbps)
  • 4x USB 3.2 Gen 2×2 20 Gbps
  • 10 USB 3.2 Gen 2×1 10 Gbps
  • 10 USB 3.2 Gen 1×1 5 Gbps
  • 14 USB 2.0
  • x8 DMI Gen 4.0 (CPU-PCH Interconnect)
  • IRST 17 for PCIe storage RAID on CPU & PCH Support
  • Hardware-Level PCIe Gen 4 Integration from various board vendors
  • Optane Memory Ready
  • Thunderbolt 4.0
  • Integrated Intel 2.5G Base-T Ethernet
  • Integrated Intel WiFi 6E Support

Another interesting thing to talk about the Z690 PCH is its process node and dimensions. The Z690 PCH is based on the 14nm node and measures at 98mm2 which is slightly larger than the Z590 PCH featured on LGA 1200 socket motherboards.

Intel Desktop Platform Chipset Comparison

Chipset Name Alder Lake-S (ADL-S) PCH / 600 Serie (Z690) Rocket Lake-S (RKL-S) PCH / 500 Series (Z590) Comet Lake-S (CML-S) PCH / 400 Series (Z490) Coffee Lake S (CNL-H) PCH / 300 Series (Z390/H370, B360, Q370, H310) Coffee Lake S (KBL-R) PCH / Z370 Platform
Process Node TBA 14nm 14nm 14nm 22nm
Processor 16C,12C,10C,6C,4C (Full corporate/consumer SKU stack at launch) 8C, 6C (Full corporate/consumer SKU stack at launch) 10C, 8C, 6C, 4C, 2C (Full corporate/consumer SKU stack at launch)
Enhanced IA and Memory Overclocking
Gen 9 Intel Graphics GT2 (Up To 24 EUs)
Corporate/vPro & Consumer
8C, 6C, 4C, 2C (Full corporate/consumer SKU stack at launch)
Enhanced IA and Memory Overclocking
Gen 9 Intel Graphics GT2 (Up To 24 EUs)
Corporate/vPro & Consumer
8C, 6C, 4C (6 Consumer SKUs at Launch)
Enhanced IA and Memory Overclocking
Gen 9 Intel Graphics GT2 (Up To 24 EUs)
Consumer Only
Memory Up To DDR5-4800 (Native)
Up To DDR4-3200 (Native)
Up To DDR4-3200 (Native) Up To DDR4-2933 (Native) Up To DDR4-2666 (Native) Up To DDR4-2666 (Native)
Media, Display & Audio eDP / 4DDI (DP, HDMI) Display Capabilities DP 1.2 & HDMI 2.0, HBR3
HDCP 2.2 (HDMI 2.0a w/LSPCON)
12-bit AV1/HEVC & VP9 10-bit Enc/Dec, HDR, Rec.2020, DX12
Integrated Dual-Core Audio DSP With USB Audio offload
SoundWire Digital Audio Interface
DP 1.2 & HDMI 1.4
HDCP 2.2 (HDMI 2.0a w/LSPCON)
HEVC & VP9 10-bit Enc/Dec, HDR, Rec.2020, DX12
Integrated Dual-Core Audio DSP
SoundWire Digital Audio Interface
DP 1.2 & HDMI 1.4
HDCP 2.2 (HDMI 2.0a w/LSPCON)
HEVC & VP9 10-bit Enc/Dec, HDR, Rec.2020, DX12
Integrated Dual-Core Audio DSP
SoundWire Digital Audio Interface
DP 1.2 & HDMI 1.4
HDCP 2.2 (HDMI 2.0a w/LSPCON)
HEVC & VP9 10-bit Enc/Dec, HDR, Rec.2020, DX12
Integrated Dual-Core Audio DSP
I/O & Connectivity Integrated USB 3.2 Gen 2×2 (20G)
Integrated Intel Wireless-AC (Wi-Fi6E/ 7 BT CNVio) with Gig+
Integrated SDXC 4.0 Controller
Thunderbolt 4.0
Integrated USB 3.2 Gen 2×2 (20G)
Integrated Intel Wireless-AC (Wi-Fi6E/ BT CNVi)
Integrated SDXC 3.0 Controller
Thunderbolt 4.0 (Maple Ridge)
Integrated USB 3.2 Gen 2
Integrated Intel Wireless-AC (Wi-Fi / BT CNVi)
Integrated SDXC 3.0 Controller
Thunderbolt 3.0 (Titan Ridge) w/ DP 1.4
Integrated USB 3.1 Gen 1 (5 Gbps)
Integrated Intel Wireless-AC (Wi-Fi / BT CNVi)
Integrated SDXC 3.0 Controller
Thunderbolt 3.0 (Titan Ridge) w/ DP 1.4
Integrated USB 3.1 Gen 1 (5 Gbps)
Thunderbolt 3.0 (Alpine Ridge)
Storage Next-Gen Intel Optane memory
PCIe 5.0, 6x SATA 3.0
Next-Gen Intel Optane memory
PCIe 4.0, 6x SATA 3.0
Next-Gen Intel Optane memory
PCIe 3.0, 6x SATA 3.0
Next Gen Intel Optane memory
PCIe 3.0, 6x SATA 3.0
Next Gen Intel Optane memory
PCIe 3.0, 6x SATA 3.0
Security N/A N/A Intel SGX 1.0 Intel SGX 1.0 Intel SGX 1.0
Power Management C10 & S0ix Support for Modern Standby C10 & S0ix Support for Modern Standby C10 & S0ix Support for Modern Standby C10 & S0ix Support for Modern Standby C8 Support
Launch 2022 2021 2019 2018 2017

Meet The LGA 1700 Socket – 12th Generation CPU Support

As mentioned earlier, LGA 1200’s reign is finally over and the LGA 1700 socket is here now. The new socket adds more pins to the socket and changes the dimensions entirely. The new LGA 1700 socket offers 500 more pin connections to the CPU, allowing for more communication channels with the board itself and accommodating electrical pin configurations that are required to support 12th Gen CPUs.

So as for the socket details, Intel is going with an asymmetrical design which poses since the Alder Lake CPUs are no longer square-shaped. The Alder Lake desktop CPUs will come in a 37.5×45.0mm package and will be supported by the ‘V0’ socket which we know as LGA 1700. The new socket also changes the mounting positions to a 78x78mm grid rather than a 75x75mm grid. The Z-height has also changed to 6.529mm compared to 7.31mm on the previous LGA 12**/115* sockets.

This would lead to two big changes, first of all, CPU coolers will have to be mounted properly over the CPU which needs to be confirmed with the vendor prior to installation, and secondly, new and refreshed mounting brackets need to be shipped by cooler manufacturers for Intel Alder Lake and LGA 1700 support.
Specifications
Intel LGA1700 Socket details
IHS to MB Height (Z-Stack, validated range): 6.529 – 7,532 mm
Thermal Solution Hole Pattern: 78 x 78 mm
Socket Seating Plane Height: 2.7 mm
Maximum Thermal Solution Center of Gravity Height from IHS: 25.4 mm
Static Total Compressive Minimum: 534N (120 lbf), Beginning of Life 356 N (80 lbf)
End of life maximum: 1068 N (240 lbf)
Socket Loading: 80-240 lbf
Dynamic Compressive Maximum: 489.5 N (110 lbf)
Maximum Thermal Solution Mass: 950 gm
Important Note: A Keep In Zone is introduced for LGA17xx-18xx thermal solutions. Two volumes are provided.
The Asymmetric volume provides the maximum available design space. The Symmetric volume
provides for designs to be rotatable on the board. The thermal solution under load should fit within the volume

What’s interesting is that the Alder Lake CPUs use an asymmetrical design and while we don’t know how the dies will be positioned under the IHS, we do know from AMD Threadripper that CPUs that carry such design require full IHS coverage and that may be the tricky part when it comes to cooling the brand new Alder Lake CPUs. So far, we know that Alder Lake will be a monolithic yet hybrid CPU design so it remains to be seen how cooling is handled for these 12th Generation chips.

Cooler Compatibility With LGA 1700 Socket

To make their existing coolers compatible with Intel’s Alder Lake lineup, many cooling brands have released LGA 1700 upgrade kits which feature mounting hardware for the new socket. But the Intel Alder Lake platform isn’t just featuring a brand new mounting design but the CPU dimensions itself have changed too.

The LGA 1700 (V0) socket not only has an asymmetrical design but also comes with a lower Z-stack height. This means that proper mounting pressure is needed to make full contact with the Intel Alder Lake IHS. Certain cooler manufacturers have already been using larger cold plates for Ryzen and Threadripper CPUs to make proper contact with the IHS but these are mostly higher-end and new cooling designs. Those who are still running older AIOs with round cold plates could have trouble maintaining the required pressure distribution which could lead to inadequate cooling performance.

Cooling will play a major role in determining the performance of Intel’s Alder Lake CPUs, especially the unlocked lineup, which as per the leaked benchmarks, runs really hot. Users will have to utilize the best of the best cooling hardware to maintain an adequate temperature & more so if they are planning on overclocking the chips.

We know from the official announcement that Intel’s 12th Gen Alder Lake Desktop CPUs will be available in two main die configurations. These include an 8+8 (C0) die which will be featured on high-end unlocked and non-K SKUs and a 6+0 (H0) die that will be used for non-K SKUs. All of the dies are based on the 10nm ESF process node (Intel 7) and feature a smaller die size compared to 14nm SKUs.

Intel Alder Lake CPU Dies, Hot Spots & Best Cooling Configurations

The Intel Alder Lake C0 die has a surface area of 215.25mm2 while the H0 die has a surface area of 162.75mm2. This means that as the area shrinks and transistor density increases, heat dissipation will become a major challenge. The CPUs will use a Tin TIM material between the die and IHS along with a soldered design. Just for comparison, the 14nm Rocket Lake CPUs with 8 cores & 16 threads had a surface area of 276mm2.

Since both Alder Lake dies have a different core configuration, their hot spots are also different. As shown in the thermal imagery by MSI, the C0 (8+8) die has its hot spot close towards the center but the H0 (6+0) die has its hot spot slightly towards the left side. So it will be important to use the right cooling configuration based on the SKU you are getting.

For this purpose, MSI has shown some best cooling configurations to use with the Intel Alder Lake CPUs. For starters, it is suggested that copper-based heatsinks are used instead of aluminum ones. The positioning will also play an important role as you are advised to have the heat pipes running parallel to the die and not perpendicular. The parallel configuration will ensure that the heat pipes are making direct contact over the hot spots and not making inefficient contact.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.