Thunderbolt 3 M.2 NVMe SSD Enclosures

[AAPLGeek]

Re JEYI Support :

" By consulting with our engineers, the ASM2464 chip is indeed incompatible with the Thunderbolt chip in some Mac devices. We are working on upgrading the firmware to resolve this compatibility issue. ASM2464 is more suitable for mac products after 2022. Because the previous apple device chip is JHL7XXX, ASM2464 is more suitable for JHL8XXX thunderbolt chip "



I tried with a MacBook Pro M1 Pro and mac mini M2 from 2023 and it's exactly the same, i don't exceed 1500mo/s

Sounds like a load of BS to me. My 2020 M1 MBA is from the earliest possible batches and it never had any speed anomalies on any of the firmware versions.

According to iFixit teardown of the M1 MBP from Nov 2020, it contains the "Intel JHL8040R Thunderbolt 4 Retimer".

M1 MacBook Pro and Air Teardowns

Past, meet future—it’s a mix of old and new in our dual teardown of the newest M1-powered MacBooks.

www.ifixit.com

 

[PaulD-UK]

There was talk in the early M1 days of the TB/USB-C ports ‘reserving’ bandwidth for a Video data stream even if only the HDMI port was being used. This reduced data bandwidth significantly.
If a USB-C video monitor was connected all the ports data bandwidth increased to the expected transfer rate.
I never heard if the OS now corrects this.

 

[Earl Urley]

I was finally able to flash my JEYI 2464-FAN to 10_05 by connecting it with a cable I had bought what seems like ages ago, an IOGear USB 3.2 USB-C to USB-C SS-10 Gb cable.

So it looks like the JEYI firmware flasher requires a pure USB-C connection, not a Thunderbolt connection.

Thanks to masotime for the idea!

BTW I was able to do this on a 2018 MacBook Pro.. with Windows 10 installed via BootCamp..

Hmm, weird, I'm getting slower speed tests on the JEYI than I did with the Hyper with both the Hynix P31 and the Inland Performance Plus on an M1..

 

[occamsrazor]

Bought the Maiwo K1695 USB4 NVMe enclosure (ASM2464) and hitting these speeds with the Samsung 990 Pro.

View attachment 2319043

How do you like it so far? Is it this one?

MAIWO 40Gbps NVME Enclosure with Fan, M.2 Enclosure for MacBook Pro/Air, Type C SSD Enclosure Compatible with Thurderbolt4.0&USB4/3.2/3.1, 8TB Storage Expansion,Tool-Free

[Support up to 8TB M.2 NVME SSD]:This M.2 NVME enclosure supports M key and B+M key M.2 NVME SSD with size 2280, ensuring compatibility with a wider range of devices,and this enclosure can support up to 8TB storage capacity expansion. [40Gbps High Speed] : The NVME reader compatible with Thurderb...

www.amazon.com

I was looking to buy another enclosure to use with my 2019 MBP 16" and was going to get another Acasis TBU401 (which I have already) but saw these newer enclosures at around the same price.

 

[chars1ub0w]

It seems the Zike drive is actually a Satechi. From my M1 Max MacBook Pro.
So one could buy that instead.

USB4 NVMe SSD Pro Enclosure | Satechi - Apple and PC Accessories

satechi.net

 

[masotime]

It looks fine. I see you formatted it to exfat. Have you tried apsf? I get the write speed in AmorphousDiskMark completely wrong (far too low: 1000 mb/s).
I have the feeling that AmorphousDiskMark has a problem with e.g. SN850x and other SSDs from WD.
I get this with a SN850x 4tb (ssd little over half full):

Surprisingly reformatting APFS has different results - not by too much, but enough that it's significant. Interesting...

I didn't have problems with write speed for AmorphousDiskMark though

 

[sam85]

It seems the Zike drive is actually a Satechi. From my M1 Max MacBook Pro.
So one could buy that instead.

USB4 NVMe SSD Pro Enclosure | Satechi - Apple and PC Accessories

satechi.net

Is ASM2464PD used inside that Satechi Pro enclosure?
Interestingly, they don't want to publish that information...

 

[mogens]

Surprisingly reformatting APFS has different results - not by too much, but enough that it's significant. Interesting...

View attachment 2320739 View attachment 2320740
I didn't have problems with write speed for AmorphousDiskMark though

View attachment 2320744

Thanks! Could you post the firmware version of your sn850x? Mine is revision: 624331WD (see under NVMEExpress in system report). Maybe that’s the problem with mine?

 

[masotime]

Thanks! Could you post the firmware version of your sn850x? Mine is revision: 624331WD (see under NVMEExpress in system report). Maybe that’s the problem with mine?

Looks identical to mine. Maybe something else?

 

[mogens]

Looks identical to mine. Maybe something else?

View attachment 2320827

Thanks! The behavior and naming of ssd firmware is pure woodo.
maybe it’s because the iMac M1 is only thunderbolt 3?
it looks like there’s a newer firmware (620361WD), but I have no idea how to update. I tried WD dashboard on my windows PC, but it wouldn’t load. This has been reported all over WD’s support forum:
https://funfunhan.com/2477941

 

[chars1ub0w]

Is ASM2464PD used inside that Satechi Pro enclosure?
Interestingly, they don't want to publish that information...

I'm guessing it is the same as the Zike as the enclosure looks identical. But it's so new, I guess nobody has posted on it.

 

[masotime]

It seems the Zike drive is actually a Satechi. From my M1 Max MacBook Pro.
So one could buy that instead.

USB4 NVMe SSD Pro Enclosure | Satechi - Apple and PC Accessories

satechi.net

Just a caution that I'm getting very different results when I scan my own ZikeDrive. Not sure why there's a discrepancy.

 

[pransu]

I'm guessing it is the same as the Zike as the enclosure looks identical. But it's so new, I guess nobody has posted on it.

I found this blog stating both Zike and Satechi are using ASM2464PD chip.

I just got the Satechi enclosure yesterday. R/W is pretty impressive at 3,100MB, but it runs pretty hot. At idle, it sits at 120F/48C. Not sure if it's considered normal. This is my first SSD enclosure, so not sure what to expect.

 

[masotime]

Both Satechi and ZikeDrive (ZikeTech) are manufactured by Gopod Group Holding Limited.

Excellent find! I guess it’s best to just go with whichever’s cheaper then, thermal aspects notwithstanding.

 

[macphoto861]

I found this blog stating both Zike and Satechi are using ASM2464PD chip.

I just got the Satechi enclosure yesterday. R/W is pretty impressive at 3,100MB, but it runs pretty hot. At idle, it sits at 120F/48C. Not sure if it's considered normal. This is my first SSD enclosure, so not sure what to expect.

I ordered the Zike, haven't received it yet though. I do have some concerns over the temperature though... apparently all the ASM2464PD enclosures run hot?

 

[chars1ub0w]

I found this blog stating both Zike and Satechi are using ASM2464PD chip.

I just got the Satechi enclosure yesterday. R/W is pretty impressive at 3,100MB, but it runs pretty hot. At idle, it sits at 120F/48C. Not sure if it's considered normal. This is my first SSD enclosure, so not sure what to expect.

I use a Lexar. It has no DRAM cache and possibly takes less power

 

[Loco Sock]

When I had my Acasis enclosure with a Samsung 990 Pro, it ran at 120F with no problem. Wasn't close to get throttled or anything.

 

[richpjr]

My Hyper enclosure with a 4TB WDSN850X in it, idles about 107F/42C.

 

[masotime]

I ordered the Zike, haven't received it yet though. I do have some concerns over the temperature though... apparently all the ASM2464PD enclosures run hot?

Yeah it can get pretty warm. I got the fan version of the JEYI for that reason. It works pretty well, but if you have a sustained high workload it'll eventually get pretty hot anyway.

 

[decat]

I have two TBU405's in use. One with a 980 PRO 2TB and one with a 990 PRO 4TB.
The 980 is connected to a Caldigit Element USB4 hub, the 990 is connected to a Mac Studio.

Idling they both hover around 35 C / 95 F.
Under load like for example copying 600GB temps of both drives rise to +- 55 C / 131 F.
Both drives have a 0.5mm 13.3w/m.k thermal pad and a copper shim between the pad and the casing.
Temperatures with the stock pads were a tad higher under load, around 65 C / 150 F if I remember correctly and around 42 C / 107 F while idling.

Idling / while browsing the Mac Studio consumes around 14w.
When copying files to the 990 Pro connected to the Studio power consumption rises to 26w max.

 

[chars1ub0w]

A bit worrying then. My Zike drive with Lexar 4TB is at 60ºC doing absolutely no work according to Mac Fans Control

 

[AAPLGeek]

A bit worrying then. My Zike drive with Lexar 4TB is at 60ºC doing absolutely no work according to Mac Fans Control

That's way too hot for idle. Are you sure there isn't some indexing activity going on in the background?

It also might be a good idea to try a different NVMe. I've had this issue with two Phison E18 drives on ASM2464, where the idle temp would rise steadily around 58-64ºC for no reason at all.

 

[macphoto861]

Yeah it can get pretty warm. I got the fan version of the JEYI for that reason. It works pretty well, but if you have a sustained high workload it'll eventually get pretty hot anyway.

My current enclosure has a fan, but what I don't like about it is that the write speed is substantially lower when connected through my TS4 dock... with a Sabrent Rocket, my write speed is about 2172 when connected directly to my MBP, but when connected to one of the TB ports on the TS4, write speed is 1517 (read speed stays roughly the same either way).

I'm taking a shot in the dark with the Zike, hoping that perhaps the new chipset will get around this issue (since it's reported that it doesn't unnecessarily reserve bandwidth for a display). But I don't have a lot of confidence that this will be the case, since it's probably the dock itself that is causing the slowdown, regardless of what kind of enclosure is used.

Still, seeing these reports of high temperatures, I kinda wish I'd gone with another enclosure with a fan.

 

[masotime]

but when connected to one of the TB ports on the TS4, write speed is 1517 (read speed stays roughly the same either way).

since it's probably the dock itself that is causing the slowdown, regardless of what kind of enclosure is used.

Yeah I think it's really the dock itself - looking at https://www.caldigit.com/ts4-interface-bandwidth-allocation-and-diagram/

My read is that the 40Gps downstream ports in group 3 are "shared" - so I would expect it to reserve some of the bandwidth for the display (since it's a primary function as a dock)

 

[macphoto861]

My read is that the 40Gps downstream ports in group 3 are "shared" - so I would expect it to reserve some of the bandwidth for the display (since it's a primary function as a dock)

Yeah, you're probably right. I wish that was not the case though. I understand that using a display would consume some upstream bandwidth (and needs to be prioritized), but it seems wasteful to simply make a big chunk of bandwidth permanently unavailable... this should be a dynamic thing, allocating the full 40gbps bandwidth to whatever devices are connected, only setting aside bandwidth for a display when one is actually connected.