Review WD SN740 2TB review: The M.2 2230 OEM SSD of choice

HideOut

Distinguished
Dec 24, 2005
564
86
19,070
Shouldnt the endurance go up in a ratio the same as the storage pool size? 2TB should be double of 1, 1 should be double of 512 and such. This makes no sense. Misprint?
 
Shouldnt the endurance go up in a ratio the same as the storage pool size? 2TB should be double of 1, 1 should be double of 512 and such. This makes no sense. Misprint?
You might just try googling before questioning if it's a misprint. A few seconds brought me to the specs pages:
https://www.westerndigital.com/products/internal-drives/pc-sn740-ssd?sku=SDDPNQE-2T00

And sure enough, WD has some curious numbers. 200TBW for 256GB, up to 500TBW for 2TB. Why? Because this is an OEM drive, and so the requirements are different. Apparently, WD doesn't want to give a "reasonable" TBW guarantee on larger models of the SN740. If you took the 200TBW for the 256GB and scaled it, that would be 1600TBW for the 2TB model, which would be more than what you get on the SN770M. 🤷

Realistically, I doubt there's a "drop dead" setting in the firmware, so the drives will work as long as they can, and that likely means the 2TB SN740 can do far more than the listed 500TBW. But it's not guaranteed.
 
The power difference between the WD SN740 (2.5W/GB) and the Corsair MP600 (1.7W/GB) is pretty substantial. We're talking a difference of 800W/TB and that can add up pretty quick.

Having said that, they're SSDs and as such are already remarkably efficient compared to the old platter and optical drives.
 
Last edited:

Aragorn

Distinguished
Feb 17, 2005
528
2
19,015
Thats actually a total power of 2500Ws/TB (that's 2500WattSeconds or .694Wh to write a TB extrapolation from that graph, a non-negligible but still small amount of power. (Edited for bad math, missing a 10^3)
 
  • Like
Reactions: JarredWaltonGPU

Notton

Prominent
Dec 29, 2023
333
294
560
I figured the TBW endurance was low, because it was a 2230 drive.
Sabrent rocket Q4 2230 2TB also only has 450TBW.

When looking at the WD SN770 2TB, it has a similar chip configuration. 1x controller and 1x 2TB density NAND. However, it gets an endurance of 1200TBW.

I wonder if the rating for 2230 versions are so low due to heat?
 
I figured the TBW endurance was low, because it was a 2230 drive.
Sabrent rocket Q4 2230 2TB also only has 450TBW.

When looking at the WD SN770 2TB, it has a similar chip configuration. 1x controller and 1x 2TB density NAND. However, it gets an endurance of 1200TBW.

I wonder if the rating for 2230 versions are so low due to heat?
WD SN770M is virtually the same but has “normal” endurance rating and scaling.
 

cyrusfox

Distinguished
Superb review! Great to see Tom's add this to your 2230 lineup.

My main critique is this review is lacking a stronger discussion around pricing. GB/$ is king for most when looking at purchasing a drive, and this beats every other drive by min of $30 to 2x the price while offering comparable performance. The main drawback is OEM so good luck with a warranty if you need it... But many of us are comfortable with this, evident by us purchasing a drive off Aliexpress, won't lie as I was very worried about it being some fake drive for the price initially. Other point of contention, the 1TB are great for benchmark comparison, but they truly are a separate class, so the suggestion of purchasing BC711 doesn't make sense to me as it has no 2tb entry, would be a fairer fight to compare the sn740 1tb entry as I suspect its performance and efficiency profile may improve(Sacrifices were made to cram 2tb in this form factor). When we are in the 2tb class we are here because we need/desire the space.

It was great to see this drive sometimes beat the SN770M(at a fraction of the price) as well as having equivalent real world performance and battery life (which one would not expect from its poor efficiency showing). Also thank you for the in-depth sustained write performance data, really found the inconsistent performance intriguing as it seems to be capable of flushing the cache, speed returns only to fill it back up again and throttle down. The initial performance of the S91 is amazing on QLC, can't wait to see what WD can do with an update, but this might be the end of TLC, I would expect them to follow QLC for the next release...
tb5oWDgtgmehYsRHc4z5WJ.png

I can't seem to find it anywhere but I would be interested in seeing what max temp was seen during the sustained write performance for each drive, gives a good indication on how the controller is tuned for performance and how important cooling is to maintain peak capabilities.

Thanks again to adding this to your line up of 2230 drives! Will be a great baseline going forward to compare against as we get newer entries into this small form factor space(I sure hope we see 4TB drives or pcie gen 5 entries even though both in this form factor is a bit absurd:cheese:). What we really need is someone brave enough to integrate a solid state fan (airjet) with these solid state drives (Throw power consumption out the window, make it PCIE gen 5 and set some records). I don't expect it to sell well, Optane didn't, as storage is too price sensitive of a space but like drag racing, it is a fun hobby.
 
Last edited:
Superb review! Great to see Tom's add this to your 2230 lineup.

tb5oWDgtgmehYsRHc4z5WJ.png

I can't seem to find it anywhere but I would be interested in seeing what max temp was seen during the sustained write performance for each drive, gives a good indication on how the controller is tuned for performance and how important cooling is to maintain peak capabilities.
The reviews should all have temperature data, though only when running in the desktop system. The SN740 reports three different temperatures in HWiNFO64, and they're all quite high. The first and third numbers appear to be the same and peaked at 91C in our write saturation test, while the second number peaked at 101C. Yeah, that's hot. The good news is that in PCIe 3.0 mode the temps were MUCH lower: 62C on the first and third columns, and 74C on the second column.

The WD SN770M seems to have been tuned to be a bit better on thermals, with the first and third columns at 85C peak and 95C for the second column when operating in PCIe 4.0 mode. Again, PCIe 3.0 greatly reduces thermals and it peaked at 70C / 80C — so hotter in PCIe 3.0 mode than the SN740, but cooler in PCIe 4.0 mode. 🤷

The QLC NAND and Phison E21T combination incidentally tends to hit max temps of around 65~70C in the same write saturation testing. Obviously, being able to write at much higher sustained speeds with TLC is a big part of that.
 
  • Like
Reactions: cyrusfox