SilverStone's New M.2 SSD Riser Card Fights Overheating

(Image credit: SilverStone)

SilverStone has announced the new ECM23 M.2 PCIe SSD riser card. The company hasn't revealed all details about the device--pricing information still isn't available--but it's likely released just enough to pique the interest of those wanting better storage performance.

Specifications

Swipe to scroll horizontally
Model No.SST-ECM23
LED indicatorPCIe M.2 SSD Read / Write LED: Blue blinking
InterfacePCI Express x4 via full size x16 pinout
SSD InterfaceM.2 (NGFF)
Module KeyPCIe x4 NVMe-based M key
Operating system supportMicrosoft Windows 7 (32 bit/64 bit), 8, 8.1,10 (32 bit/64 bit) or future release versions Mac OS 10.2.8 or above Linux
Support Length of M.2 SSD30mm, 42mm, 60mm, 80mm
Dimensions (WxHxD) and WeightHeatsink with PCB: 4.1 x .4 x 1.7 inches (105 x 11 x 44mm), 52g
Thermal Pad
Thermal Conductivity1.5W/m.k
Operating Temperature-40°C ~ 220°C
Dimensions (WxHxD) and Net WeightCharcoal Pad:2.8 x 0.02 x 0.8 inches (70 x 0.5 x 20mm), 2g   Blue Pad:2.8 x 0.04 x 0.8 inches (70 x 1 mm x 20mm), 3.6g Gray Pad:2.8 x 0.06 x 0.8 inches (70 x 1.5 x 20mm), 5.5g

The ECM23 allows M.2 storage to run via PCIe connections, so performance chasers don't have to buy new motherboards (most probably don't mind buying new components, but the budget-constrained might like to simply buy a riser card and SSD instead of upgrading much of their system). Many motherboards are often limited to just one M.2 slot, but using a riser card like the ECM23 can help get around that limitation.

The ECM23 is compatible with AHCI over PCIe and NVMe but doesn't support plain-Jane M.2 SATA connections, according to SilverStone, which said the AHCI support is "used for PCIe SSDs and interfaced through the AHCI driver and provided PCIe lanes, providing backward compatibility with widespread SATA support in operating systems at the cost of not delivering optimal performance."

SilverStone also tested the ECM23 with several Samsung SSDs: the Samsung SM951 and S961Samsung 960 EVO and Samsung 960 Pro were all said to "work optimally" with this riser card.

In addition to offering PCIe compatibility, the ECM23 offers a thermal pad and aluminum heat sink to improve the connected SSD's performance and help prevent it from overheating. Three pads are available: a small charcoal pad, a medium-sized blue pad and a large gray pad  (good luck telling the difference between "charcoal" and "gray")that are 0.5, 1 and 1.5mm thick, respectively, and vary considerably in weight.

SilverStone's announcement comes after Gigabyte revealed two new riser cards of its own, the CMT4034 and CMT4032, in July, though neither is currently available, and their prices also have yet to be revealed. So it's clear that people interested in M.2 riser cards will soon have more options--the question is how much that extra performance will cost them. 

Nathaniel Mott
Freelance News & Features Writer

Nathaniel Mott is a freelance news and features writer for Tom's Hardware US, covering breaking news, security, and the silliest aspects of the tech industry.

  • bogdi1988
    LOL! Fry's has been selling this for a while now. This is just a rebadge.
    Reply
  • planedrop
    Why? Actively cooling SSDs with heatsinks is a bad idea, there is a reason the companies that make the actual drives don't include heatsinks. The flash is meant to get hot, it's just the controller that has to be kept cold. This is why some SSDs are even designed to move the heat from the controller onto the flash, because flash actually works better at higher temps. Intel and the like stress test their flash by cooling it down lol.
    Reply
  • luckymatt42
    Ya I've been trying to warn others off of those "M.2 Coolerz" for awhile now. The flip side of that though is that us computer geeks have been trained with the "cooler is better" mantra for years now...generally true, but not in all cases.
    Reply
  • stdragon
    I got a cheapie Chinese special (anodized aluminum strips for both sides) held together with two included bands. It does the job well and prevents thermal throttling. And, it doesn't look bad at all.
    Reply
  • BaRoMeTrIc
    21314899 said:
    Why? Actively cooling SSDs with heatsinks is a bad idea, there is a reason the companies that make the actual drives don't include heatsinks. The flash is meant to get hot, it's just the controller that has to be kept cold. This is why some SSDs are even designed to move the heat from the controller onto the flash, because flash actually works better at higher temps. Intel and the like stress test their flash by cooling it down lol.


    Reply
  • luckymatt42
    21314975 said:
    I got a cheapie Chinese special (anodized aluminum strips for both sides) held together with two included bands. It does the job well and prevents thermal throttling. And, it doesn't look bad at all.

    You've had thermal throttling issues with an M.2 drive? You can confirm that the drive performs better after you installed the heatsink? Are you cooling just the controller or the memory modules also?

    I'd be very curious to see your testing results, as everything I have ever seen or read says that M.2 drives do not benefit from cooling the memory modules, and even cooling the controller is a bit pointless UNLESS you are doing very large file transfers, in which case the controller can get pretty warm.



    Reply
  • stdragon
    21315107 said:
    21314975 said:
    I got a cheapie Chinese special (anodized aluminum strips for both sides) held together with two included bands. It does the job well and prevents thermal throttling. And, it doesn't look bad at all.

    You've had thermal throttling issues with an M.2 drive? You can confirm that the drive performs better after you installed the heatsink? Are you cooling just the controller or the memory modules also?

    I'd be very curious to see your testing results, as everything I have ever seen or read says that M.2 drives do not benefit from cooling the memory modules, and even cooling the controller is a bit pointless UNLESS you are doing very large file transfers, in which case the controller can get pretty warm.

    You can read my Amazon review (<-- in the link). Oh, and now on sale for $6.39!

    Because I use my workstation for VM testing and database re-indexing, there are times were I absolutely slam my 950 Pro NVMe. It's rare, but when I do, I do!

    Most of the time, a 3rd party solution is rarely needed. Most top tier SSDs will in fact include a metal strip in the underside of the sticker affix to all the chips on an SSD. The purpose is to spread the heat around so as to not create hot-spots within the chip. But should the temps rise to 75c even for a moment, throttling occurs until the temps drop again. If you were to benchmark, you would see it drop and plateau with flat performance.

    Because of where my M.2 slot is located, I needed a heat-sink to both absorb and radiate the thermals via IR radiation. I'm not sure I've ever heat-soaked it yet, nor do I think it's possible.

    Anyways, there you go. A cheap solution for an extremely rare corner case that not even most gamers would encounter. But if you do experience thermal throttling, it can be solved on the cheap :)

    Note: The one I got is single-sided as I didn't have much clearance between the bottom of the SSD and the MB.
    Reply
  • Christopher1
    Is it true that flash does better when it is (reasonably) hotter? Does that explain why my USB 3.0 flash drives are kinda hot to the touch when I have used them?
    Reply
  • Pompompaihn
    Guys....the side WITHOUT the heatsink is -red-, how much more evidence do you need? I care about my equipment, blue for me, thanks!
    Reply
  • stdragon
    Excessive Heat does NOT make an IC perform better. Now, they could be designed to withstand hotter temperatures so they can be clocked faster, but purposefully heating the IC does not inherently add a boost in performance. In fact, as they get hotter, resistance goes up. This can lead to accelerated electromigration between the chip and the interconnects within the chip package itself. Eventually the connect breaks and you end up with a permanent failure.
    Reply