Question *STABLE* NVMe - USB Adapter?

[cyberjedi]

I'm looking to buy a *RELIABLE* NVME-USB external adapter, that doesn't disconnect randomly, can keep up the close to 1GB/s speed.
So far all adapter's reviews contain frighteningly high % of reports of malfunction, random disconnects, sub-par speeds, falling to even under USB 2.0 speeds, etc.
There are 3 main chipsets in the market: JMicron JMS583, Asmedia ASM2362, and Realtek RTL9210 (no product has surfaced with the latter yet, so no REAL reviews, experiences yet UPDATE: it is on the market already).
Also ASUS has a new product, ROG Strix Arion, but there isn't enough feedback about it yet, and I mean real life feedback not just being able to run a few benchmarks, but real stress test to see if the device (adapter) can keep it up, without disconnecting or slowing down hard.
I would really appreciate any feedback from those who have GOOD experiences during stress-tests, like tons of random 4k writes with a stable speed, cloning 100's of GB-s at a stable high speed close to 1GB/s.
I know SSD's very well, so please stick to the point and don't start to educate me on how their speed falls after some writes, etc. I'm aware and this is NOT the point.
You can help if you can point me to a STABLE and RELIABLE NVMe-USB-C adapter that WORKS as expected.
Thank you in advance!

 

[spike_V_sky]

Default for the RTL9210 enclosures I own: Led flashing slow in low power mode.

See post 851 for a config change to have the LED turn off in lower power mode. Feel free to try different values for SUSPEND_LED_OFF... you may find a previously undocumented setting or Slow Flash & Off may be the only options.

hi

is there an option to turn off the power completely to the enclouser when the pc goes to sleep or is turned off? These days I tried with the bios settings (erp, remove power from all usb ports etc ..) but the type-c port continues to be powered.

I thought it could be something related to the type-c cable, in fact I don't use the short one supplied with the enclouser but one a little longer.

 

[peroko]

Hey guys i just received Ugreen case with RTL9210B and 1.29.38 firmware version, anyone have the changelog for this firmware version?

 

[Ananas]

Mine box is Icybox IB-1817M-C31, very nice piece of hardware except the crappy chipset. [..] I have seen some firmware updates posted here but Icybox uses some kind of own firmwares starting at 133.xx.xx.xx and is not updatable to regular ones.

The funny thing about the ICY BOX IB-1817M-C31 is that while it does indeed have a JMS583 chip on the PCB (I own one and just had a look) it identifies via USB as 0x152d:0x0562, i.e. it claims to be a JMS562. This might be the reason for the untypical firmware.

Anyone who might need it here is the 133.01.00.06 version for Icybox. Not as new as generic firwmare, but still might resolve some issues.

JMS583FWUpdater_Win_M2C7_JM1_v1_0_0_0.exe

drive.google.com

Where did you get this from?

 

[Ananas]

However, the big question remains about whether I will be able to initialize the NVMe again without using a PCIe adapter or a motherboard.

Thanks for your help and I would really appreciate some more thoughts on this.

For low level NVMe SSD access a Thunderbolt SSD enclosure should work just as well as an M.2 slot on a mainboard, but one might still need a special tool provided by the manufacturer to fix an erratic SSD.

I just got me an Acasis TBU401E enclosure to hopefully resurrect a failed Transcend 220S with a Lenovo Thinkpad T470 over Thunderbolt. Transcend's "SSD Scope" does in fact see the SSD, but unfortunately seems to lack any real repair features. So I guess I'll have to start the RMA process...

 

[PRSXFENG]

Anyone having problems with 10Gbps? realtek chip, its just super unhappy on my 10gbps ports

 

[Ananas]

Anyone having problems with 10Gbps? realtek chip, its just super unhappy on my 10gbps ports

Make sure it's not the cable. I had it more than once that a cable which came bundled with an enclosure didn't work at 10Gb/s reliably. So I bought me a few true Thunderbolt 3 and 4 cables and use these exclusively now. I haven't had any cable related problems ever since.

 

[Dec0de]

Hello, which one would be better in case of stability and heat dissipation? ORICO-M2PV-C3 or ORICO-M223C3-G4
As I take, the first one uses RTL9210 chip, which is fast, but not super stable if I took the previous posts right.

As for the second one I was unable to find any chip info.
I have Samsung 970 evo 2Tb SSD, so heating issue overcome is important for me.

Or maybe I should stay for Orico M2PF-C3, which has proved its stability according to the previous posts?

 

[Tech Junky]

@Dec0de go on Amazon and search the reviews for 970 and see what people have to say.

 

[PRSXFENG]

Hello, which one would be better in case of stability and heat dissipation? ORICO-M2PV-C3 or ORICO-M223C3-G4
As I take, the first one uses RTL9210 chip, which is fast, but not super stable if I took the previous posts right.

As for the second one I was unable to find any chip info.
I have Samsung 970 evo 2Tb SSD, so heating issue overcome is important for me.

Or maybe I should stay for Orico M2PF-C3, which has proved its stability according to the previous posts?

Searching on Chinese sites, apparently the 20Gbps enclosure uses an AsMedia ASM2364 and apparently has the same "unsafe shutdown" number increasing bug like the realtek on older firmwares

ASM2364|ASMedia Technology Inc.

ASMedia Technology was founded in 2004 and went public in 2012 (stock code: 5269.TW). Headquarter is located in Taiwan. ASMedia specializes in high-speed interface technologies with strong expertise in high-speed SERDES in house development. Company strives for continual innovation and the...

www.asmedia.com.tw

 

[Ananas]

RTL9210B firmware v1.29.391.072822 spotted in the wild! No changelog available though unfortunately.

Graugear is a German company selling "Made in China" products on Amazon throughout Europe and providing good support for them, as I can tell from own experience. I own the G-M202-AC-10G enclosure which this firmware has been published for plus another new RTL9210B docking station, the G-M2DK-AC-10G.

 

[BrallyO]

RTL9210B firmware v1.29.391.072822 spotted in the wild!

Thanks for the tip!

I did a hexdump and looked at it. It's all binary goop so there isn't much to see, but sometimes there's interesting strings and other stuff that can give a hint.

The new firmware has the sting "INTEL SSDSCKGW" in it. This wasn't in the previous 1.29.38 firmware. Looks like that's an old Intel 5 SATA series ID.

sha256 is f09cfc7104776929fd2cdfe3bafbf6671644e74850c2ea0ad84fbad335fad5b3 UTNVME_B_v1.29.391.072822.bin

 

[BrallyO]

My beloved Jeyi i9 GTR adapter has sadly died. I have three other rtl9210b adapters. This was my only rtl9210 adapter, and it had an E1 revision chip.

First it totally died. Then I took it out of the carrier and was able to get it working again, but then I noticed it was corrupting data! Fortunately I didn't lose anything important. It was corrupting new GPT partition tables when I made them, corrupting filesystems, and badblocks write tests would show tons of errors.

I cleaned it with isopropyl alchohol and the errors went away for a day. I thought it was fixed but then the next day it was corrupting things again.

Maybe there's a solder crack on the board. Or maybe the chip has broken somehow. I don't know.

It turns out you can buy just the board on Aliexpress for this model. I'm sure fits in some other models as well: https://www.aliexpress.us/item/3256804436964307.html

 

[Dec0de]

@Dec0de go on Amazon and search the reviews for 970 and see what people have to say.

I've already red this. In general, yep, it's really hot. But I got my 2Tb Samsung Evo 970 heating like a hell in seconds when just playing mp3 or watching low-bitrate HD-ready videos. The enclosure was AGESTAR 31UBNV1C with JMS583 at 2.09 updated firmware.

I couldn't believe that Samsung could make things so crap, so I borrowed ORICO-PCM2 from my friend.
So, the results was astonishing: it gets just a little warm when playing mp3 + videos together. Even after hour of such load.

 

[Ananas]

I borrowed ORICO-PCM2 from my friend.
So, the results was astonishing: it gets just a little warm when playing mp3 + videos together. Even after hour of such load.

Are you saying that the Samsung 2TB 970 EVO was acting differently in two different JMS583 enclosures? Did you measure the temperature through S.M.A.R.T readings or by touching the case? If the latter: Are you sure that the perceived lower temperatures were not caused by insufficient thermal coupling (if an SSD is not connected well to its case thermally than the SSD will become hot but the case will stay cool)?

Having said that, I did have a Samsung 2TB 970 EVO die on me in a JMS583 based enclosure (Icy Box IB-1817M-C31), and while I cannot know for sure that this was temperature related, it might have been. The case always felt warm when left connected to the Thunderbolt dock even without the MacBook present and got very hot when writing to it. I often used a little USB fan to blow on it when transferring larger amounts of data. Eventually the SSD was no longer recognized, not even when moved to an ASM2362 based enclosure, so I returned the Samsung 970 EVO and got my money back.

New SSDs based on 14nm or lower manufacturing processes should run cooler, so I hope this kind of problem will eventually go away.

 

[Dec0de]

Are you saying that the Samsung 2TB 970 EVO was acting differently in two different JMS583 enclosures? Did you measure the temperature through S.M.A.R.T readings or by touching the case? If the latter: Are you sure that the perceived lower temperatures were not caused by insufficient thermal coupling (if an SSD is not connected well to its case thermally than the SSD will become hot but the case will stay cool)?

This is the right questions. I measured it by touching the case and felt significant difference.
I didn't further investigate whether it was Samsung 970 Evo getting hot in this case or its just the case's electronics itself was getting hot.

I tryed to put the drive inside the Orico's case properly, but it was twice times bigger than Agestar's. So in Orico's case there was much more free space than in Agestar's, where all things are packed tightly like sardines. Maybe thats why perceived temperature had so huge difference.

But after almost an hour of testing have passed, even with poor thermal coupling I expected Orico's case to get noticeable hot because of it still has portable size, where air dissipation is slightly possible comparing to big computer's case.
After I started to copy large files, it became hot in few minutes, but the temperatures was dropped as the copy was finished.

 

[PRSXFENG]

Make sure it's not the cable. I had it more than once that a cable which came bundled with an enclosure didn't work at 10Gb/s reliably. So I bought me a few true Thunderbolt 3 and 4 cables and use these exclusively now. I haven't had any cable related problems ever since.

Thank you! I decided to go overkill and got a Verbatim USB 3.2 Gen 2x2 20Gbps cable, now I no longer have problems

Very odd as the included cable worked for 10Gbps on my usbc port, but not when paired with a 10Gbps capable USBA to C adapter to use in a USB A port

EDIT: just did a speedtest, seems like its running at 5Gbps, which is odd but ill still accept it as its better that it being very slow and freezing up

 

[Dec0de]

've already red this. In general, yep, it's really hot. But I got my 2Tb Samsung Evo 970 heating like a hell in seconds when just playing mp3 or watching low-bitrate HD-ready videos. The enclosure was AGESTAR 31UBNV1C with JMS583 at 2.09 updated firmware.

I couldn't believe that Samsung could make things so crap, so I borrowed ORICO-PCM2 from my friend.
So, the results was astonishing: it gets just a little warm when playing mp3 + videos together. Even after hour of such load.

UPD:
Got myself Orico M2PV-C3 with RTL9210. Checked the thermal coupling: its okay, with thermal stripe is closly attached to the upper radiator pad, which I can touch. During the light using of the disk with low bitrate videos or mp3s it's not getting hot.
The only bad thing in this case that back part of this case is bade from plastic.

 

[maleko48]

I'm experiencing the same issue after upgrading to 1.27.24, did you resolve this?

After upgrading to fw v1.27.24 and beyond, I had this same problem and could not get my enclosure to connect beyond USB2.0 speeds resulting in 40MB/s transfers no matter what I tried. Reverting back to the following (newest possible working) config works best for my Xiwai USB-A tipped drive I picked up off Amazon with the RTL9210B-CG bridge chip and brought it back to USB3.0 (5Gbps) speeds, essentially matching SATA performance (around ~500MB/s writes) but in a much more convenient form factor for my purposes:

• RTL9210B_CG.cfg
• UTGDFW_B_v4.22.27.081220.bin
• UTNVME_B_v1.25.18.041921.bin
• UTHSB_MPtool_Lite_x86_v1.12.1.0423.exe

 

[peroko]

UPD:
Got myself Orico M2PV-C3 with RTL9210. Checked the thermal coupling: its okay, with thermal stripe is closly attached to the upper radiator pad, which I can touch. During the light using of the disk with low bitrate videos or mp3s it's not getting hot.
The only bad thing in this case that back part of this case is bade from plastic.

I've had that Orico for a couple of years now and it's ABS plastic so it won't break don't worry👍

 

[spike_V_sky]

Thanks for the tip!

I did a hexdump and looked at it. It's all binary goop so there isn't much to see, but sometimes there's interesting strings and other stuff that can give a hint.

The new firmware has the sting "INTEL SSDSCKGW" in it. This wasn't in the previous 1.29.38 firmware. Looks like that's an old Intel 5 SATA series ID.

sha256 is f09cfc7104776929fd2cdfe3bafbf6671644e74850c2ea0ad84fbad335fad5b3 UTNVME_B_v1.29.391.072822.bin

hi, Is worth update to 1.29.38 or 1.29.391 ? Do you finding any fix or bug ?

 

[igor_kavinski]

Anyone have experience with this one: https://www.amazon.com/dp/B0B1VBQ4K8/

Looks cool.

 

[peroko]

Anyone have experience with this one: https://www.amazon.com/dp/B0B1VBQ4K8/

Looks cool.

No but that is a really chunky thermal pad😁

 

[mclain]

Hello,

I bought the following NVMe Enclosure. It was delivered with FW 1.27.24. https://de.aliexpress.com/item/1005004267770028.html

Unfortunately I was so stupid not to save the original configuration in advance.

Maybe someone here has the original configuration and could post it to me?

Also, what do the settings do?
;EN_UPS = 0x0
;PD = 0x0
;UART_DBG_PIN = 5

After flashing the LED flashes after disconnecting and unfortunately does not go out. In addition, an empty drive is always reported, even if no SSD is installed.
"DIS_SHOW_EMPTY_DISK" fix this.

What is the firmware GD? The bin is clear, but I found nothing about the GD via google.

Thank you!

EDIT:
With CUSTOMIZED_LED = 01 01 the LED "waves" in idle and goes off when I disconnect the drive. But when I read (4k as ssd) it does not work properly. Does anyone here know how to use customize led exactly?

 

[YabbaDabbaDoooo]

Seeing that my SSK NVME enclosure has been limping for the past few weeks, I just ordered my second most expensive enclosure.

Say hello to the Asus TUF Gaming A1 (https://www.asus.com/my/motherboards-components/data-storage/asus-product/tuf-gaming-a1/)

Compared to the Strix Arion Lite and its ASM2362, the TUF Gaming A1 uses an RTL9210. A bit hesitant with the RTL, given my less-than-ideal experiences with it.

Then again, it's Asus we're talking about, not some smaller storage company. Will update with my findings over the next couple of months.

 

[Nixb]

Having said that, I did have a Samsung 2TB 970 EVO die on me in a JMS583 based enclosure (Icy Box IB-1817M-C31), and while I cannot know for sure that this was temperature related, it might have been. The case always felt warm when left connected to the Thunderbolt dock even without the MacBook present and got very hot when writing to it. I often used a little USB fan to blow on it when transferring larger amounts of data. Eventually the SSD was no longer recognized, not even when moved to an ASM2362 based enclosure, so I returned the Samsung 970 EVO and got my money back.

One thing to try next time is power cycling it.
* Basically remove the NVMe drive from the case, and put it in a motherboard slot. Basically we want power applied, but no data activity.
* Then start the computer but go into bios/uefi settings and let it sit there for half an hour. This should hopefully trigger a cleanup/repair process on the drive that is apparently roughly universal.
* Then remove the Drive from power for 30 seconds
* Repeat the sitting powered step for another half an hour

I've done this a few times to recover a 960 EVO that decided it did not like an old enclosure I tried to use and would disappear and wouldn't show up in any system.

Unfortunately I can't say the same for an Samsung PM983 in an the big IcyDock MB861U31-1M2B case (JMS583 A2). I had figured that would be safer to use with having actual PLP. Yeah no.
I'm not sure if the case had anything to do with it or not, but apparently samsung's enterprise drives are known for having a shitshow of firmware issues. In my case the firmware version only reads as "ERRORMOD" (error mode) and the only thing that'll show up is a drive with 1GB of "space" but can't actually be mounted (this might actually be representing the cache space for some technical reasons). It also turns out samsung is for some reason hellbent on not allowing firmware for enterprise drives to be publicly available so I can't even reflash. Their only utility that might allow escaping the error mode wont recognize the drive. Definitely not interested in giving them money ever again.

My dream case would be a dual case with two different drives, backed with battery, and verified to send proper shutdown commands.