Question Need to find a high power NVME 11w or more if possible.

Jan 7, 2025
4
0
10
I know, the request sounds odd but I dont see a way to search the Tom's HW test results in a way that would be easy.

For testing purposes, I am looking for a NVME with the following min specs.
PCIe Gen 4
Any capacity
Any size(length)
11w or more
m.2 M keyed
Speed > 3500MB/s Read 3100MB/s write

Any thoughts or suggestions?
 
Why does it need to be such a high wattage? Most M.2 drives are less than 10W. Even data center drives don't use that much power.
Because unfortunately people do manage to find products that hit 11w. If I recall correctly from the m.2 M key spec I pulled last year the spec calls for 12.5w as a max.

I do some testing on portable/handheld products covered by NDA but basically trying to find retail products that will be problem children as most portable devices may not be built to handle the full potential of a power hungry NVME.

In the past people have managed to come up with DRAM cached NVME that manuf spec show as 8-11w which seem to coincide with BSOD, and when replaced with lower power devices things stabilize.
 
I know, the request sounds odd but I dont see a way to search the Tom's HW test results in a way that would be easy.

For testing purposes, I am looking for a NVME with the following min specs.
PCIe Gen 4
Any capacity
Any size(length)
11w or more
m.2 M keyed
Speed > 3500MB/s Read 3100MB/s write

Any thoughts or suggestions?
If you have specific testing requirements, such as this then you may have to build your test hardware.
 
  • Like
Reactions: drahcir_rahl
Because unfortunately people do manage to find products that hit 11w. If I recall correctly from the m.2 M key spec I pulled last year the spec calls for 12.5w as a max.

I do some testing on portable/handheld products covered by NDA but basically trying to find retail products that will be problem children as most portable devices may not be built to handle the full potential of a power hungry NVME.

In the past people have managed to come up with DRAM cached NVME that manuf spec show as 8-11w which seem to coincide with BSOD, and when replaced with lower power devices things stabilize.
Looks like the Crucial T700 4TB can it the 11W requirement. Otherwise the Optane P4801X 375GB hits the wattage but is only PCIe 3 and doesn't hit the speed requirements.

Alternatively you could use an M.2 > U.2 converter and use a drive like the P5800X series
 
  • Like
Reactions: drahcir_rahl
Do you have metering that will prove the power usage that you desire ?
No, not yet. Mind you, this isn't about having the accuracy that some like Tom's hardware posts, its about finding high drain devices to see if I can cause a device to crash under load. Then working my way down somewhat until I find the cutoff. Mainly its because sometimes manufactures wont share what those limits are. So I will be relying on data from

I am considering working on modding an ADT-Link m.2 extender to see if it can give even mostly reliable readings. Perhaps even adding a way to artificially increase the power draw beyond what the NVME pulls to emulate a more high current model. Other than that I will be depending on tools like smartctl, nvme-cli and whatever else turns up.

I'm doing this in sort of a public service, to help folks who maybe cant really afford many expensive purchases, make better decisions about one line of devices. I've been there in the past, spend money I maybe should have spent elsewhere and wound up with a device that cant use the component I bought, even if it was in line with published specs. Combine that with difficult return processes or restocking fees and... It sucks, and if can save a few folks from the same issues, I will.

I wish that manufacturers did torture test their devices better and be honest and open about the limits but they dont and arent, and I hate to see someone come on and say "I bought X drive for my Y device and whenever I run a disk benchmark the device bluescreens, tried drive in another device and its fine!" Only to find out that they and several others all with similar devices have the exact same issue, and when they change to a different device don't have the issue. I looked at the tools Tom's uses for power testing and no way I could afford to get to that level, nor would it make sense to use it for one product line.

It's possible I could use external thunderbolt enclosures for the second round of tests as I have a couple of good meters for USB power (Power-Z KM003C and Power-Z C240 from ChargerLab) and may be able to work out some data that way.

If I find a particular nvme that seems to be a problem I will purchase additional of the same line and either test them myself or pass them to another just to confirm that the same issues appear to occur.