Question *STABLE* NVMe - USB Adapter?

Page 41 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

cyberjedi

Junior Member
Apr 4, 2010
23
13
81
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!
 
Last edited:

PRSXFENG

Junior Member
Jun 25, 2022
14
3
41
Thank you, and please feel free to share some more hints and experiences, this topic seems to be only very sparsely documented and a lot of people here (including myself) could benefit. (Actually not knowing what all these settings meant kept me from tinkering with it altogether.)
Honestly I dont know much, I just dumped a copy of my enclosure's stock configs
then copy paste stuff like PID/VID, Device name, serial numbers and such
and change some as I like
 

pta

Junior Member
Mar 7, 2023
1
0
6
I have brand new ORICO-BM2C3-G2 which come with RTL9210, FW Ver : 1.24.2, FW Build Date : 2020.12.07.
I could not find new FW for RTL9210 (not A/B). Also I'm looking how to backup old firmware before flashing new. I thought that UTHSB_MPtool_Lite.exe would do the trick, but dump-option gave only config details to me.
140320239560.jpg 140320239564.jpg
 

admanny

Junior Member
Nov 22, 2022
3
0
11
I have brand new ORICO-BM2C3-G2 which come with RTL9210, FW Ver : 1.24.2, FW Build Date : 2020.12.07.
I could not find new FW for RTL9210 (not A/B). Also I'm looking how to backup old firmware before flashing new. I thought that UTHSB_MPtool_Lite.exe would do the trick, but dump-option gave only config details to me.
View attachment 78113 View attachment 78114
Using regular B files worked for me
 

White7561

Junior Member
Apr 17, 2020
20
2
51
the link you posted has UTGDFW_B_v1.25.14.041521.bin but UTGDFW_B_v1.29.21.030422.bin is the newest
But GDFW isn't the actual firmware. That's the golden binary or something. You could just exchange that older GDFW with the newer one you have. The actual firmware being used is the latest, which is 1.31.17 for the B model
 

fsdbgnzn

Junior Member
Feb 12, 2022
3
2
11
enclosure's stock configs then copy paste stuff like PID/VID, Device name, serial numbers and such and change some as I like

Compatibility setting for Compact Flash Express standard (not supported on external enclosures)
CFEXPRESS = 0x0

Enable or disable active power management Open: 0x1 Close: 0x0 - Opened the device will always maintain high power. For WinToGo set it 0x1,for storage 0x0.
ASPMDIS = 0x0

idle time till disable pcie power 0x00 0 min - 0x01 1 min - 0x0A 10 mins - 0x1E 30 mins - 0x3C 60 mins
PCIE_PWRCUT_THRES = 0x1

Responsible for turning off power to PCI bus by timeout for energy saving (it helped at cloning)
DISK_IPS_THRES = 0x1

Enable or disable U1U2 - USB 3.0 Link Power Management - See Microsoft Documentation
EN_U1U2 = 0x1

Enclosure with or without fan On-0x1 Off-0x0
FAN = 0x0

Without drive enclosure is seen by Windows - good for firmware update
DIS_SHOW_EMPTY_DISK = 0x1

Blue light on all the time and flashes when it works, as opposed to no light when not in use
RM_INTERNAL_RD = 0x1

LED setting: When the USB enclosure is connected to the USB port of the machine, the LED is powered to ON. 0x1 value means: 1 (Setting activated), while 0x0 value means: 0 (Setting disabled)
LED = 0x1

CUSTOMIZED_LED is 15 bytes

This line is only a commentary to detail the content value for this setting. It needs 15 bytes for the setting value.

Customize the led color. Setting value is 0a 01 it means blue color, while when the value is 22 00 it means white color
CUSTOMIZED_LED = 0a 01

Activated this setting permits to power off the LED when the Low Power Mode is activated, not when software ejected. 0x0 value means 0 (Setting disabled), while 0x1 value means: 1 (Setting activated)
SUSPEND_LED_OFF = 0x0

This ; character must be placed before the setting (To comment the setting). When commented the default value for this setting is applied by the machine OS/Bios.
;LED = 0x1
 

lopri

Elite Member
Jul 27, 2002
13,209
594
126
I have this which appears to be using realtek chipset.


I am having a problem with it getting recognized in a VM. I want to try a firmware update. Do I just double click on the .exe and will it guide me through?

Also, what is the consensus on the best deal on these things? Price/performance-wise. I am a bit overwhelmed by the sheer length of the thread.
 

lopri

Elite Member
Jul 27, 2002
13,209
594
126
So I have been playing with recent firmwares and I am not seeing an improvement. Not performance, but compatibility-wise. Still having difficulty with connecting it to a VM via USB pass-through.
 
  • Like
Reactions: Ananas

hisamest

Junior Member
Mar 30, 2023
1
0
6
I have exactly the same problem with Acasis CM073.
Installed a WD SN350 for VM using. But in high read+write workload, the enclosure lost reaction, 100% occupied. Rarely it rework after 30s, but oftenly the whole USB enclosure is disconn after 1min., and then reconnected. Aother USB stick on the enclosure will drop too,so it's not SSD issue.
I tried to flash it with RTL9210B FW 1.31 but still failed.
In Chinese forums they also saying don't buy RTL9210 box, even don't buy any Hub with SSD enclosure. But now less and less laptop have enough USB ports....
 

Ananas

Junior Member
Jan 16, 2022
24
6
41
A lot of problems are caused by cables, so I suggest to get a short Thunderbolt 3 or 4 one (at least for testing; I use Thundebolt cables almost exclusively nowadays as cable related problems suck).

And then I get rid of enclosures which don't work reliably with a Thunderbolt cable either, or I use an M.2 SATA SSD instead of an NVMe one as I found the SATA ones to be generally much more reliable, and they don't get as hot either. I prefer slower speeds over unreliability any day.
 

tornado99

Junior Member
Mar 19, 2021
10
4
51
Most enclosure manufacturers will bundle the cheapest cable to try and save money, so best to get a separate one. Also usb-c to usb-c seems more reliable than usb-c to usb-a, if you are able to use that.
 

c0d3z3r0

Junior Member
Apr 4, 2023
1
0
11
Hi, did anybody ever test a ADATA XPG SX6000 Lite with RTL9210(B)? I got two enclosures (Orico and clone), where the NVMe is not detected at all. A Samsung NVMe works fine, so the enclosures are not completely dead at least. Firmware update (of both enclosure and NVMe) didn't change anything.

EDIT: I found two guys reporting the same problem, while for some people the SX6000 works fine with the RTL9210-based enclosures. Weird.

- https://www.jeffgeerling.com/comment/17436#comment-17436
- https://forums.macrumors.com/thread....2034976/page-432?post=31506522#post-31506522

EDIT2: I've given up on this and sent back the two RTL9210 enclosures to Amazon. Got a ICY BOX IB-1817M-C3 (JSM853 A3) working totally fine with the SX6000, even though the performance (300-500 MB/s) isn't the best.
 
Last edited:

Alpaca786

Junior Member
Oct 3, 2021
16
2
41
I have not looked in this thread since over a year. Sad to see that stuff has not changed much and these enclosures are still unreliable as heck.

The new Asus TUF Gaming A1 looks quite nice from a quality perspective, but it still uses the RTL9210 chip and there are not many reviews on Amazon available. Has anyone already tried this one?
 

tornado99

Junior Member
Mar 19, 2021
10
4
51
I have not looked in this thread since over a year. Sad to see that stuff has not changed much and these enclosures are still unreliable as heck.

The new Asus TUF Gaming A1 looks quite nice from a quality perspective, but it still uses the RTL9210 chip and there are not many reviews on Amazon available. Has anyone already tried this one?

My Sabrent RTL9210 has been rock solid. Realtek have released many firmware updates since a year ago. Also, the mistakes people seem to be making are:

1) Buying enclosures with cheap poor quality bundled cables
2) Using enclosures with USB C-->USB A conversion cable
3) Using Gen 4 hot nvme ssds instead of cooler Gen 3 e.g. WDSN570 or P31 Gold
4) Not having a proper heatsink solution - you need a thermal pad which makes good contact with both your ssd board and the metal part of the enclosure. On my Sabrent I took their pad out, and used my own thicker one.
 

Alpaca786

Junior Member
Oct 3, 2021
16
2
41
Thanks for your replies. Nice to see, that there are indeed users with no problems.

I have read that the cables often make problems and also that USB-C to -A is not ideal.

3) Using Gen 4 hot nvme ssds instead of cooler Gen 3 e.g. WDSN570 or P31 Gold
4) Not having a proper heatsink solution - you need a thermal pad which makes good contact with both your ssd board and the metal part of the enclosure.
Are these two points really that important, though?

To 3): Tom’s Hardware tests power consumption and efficiency (MB/s per Watt) and through progress the new PCIe 4.0 SSDs often are more efficient and have less maximum power draw than old PCIe 3.0 ones. E. g. the Samsung 990 Pro has better efficiency than even the mentioned SK Hynix P31 Gold. And as those SSDs are bottlenecked through the enclosure anyway, the same maximum speed of roughly 1 GB/s should leave the more efficient SSDs consume less power, which means less heat.

(I would have used a 990 or 980 Pro, due to their high efficiency like the P31 Gold.)

The respective charts are here under power consumption:


To 4): I see that a thermal pad might be neccessary so that the SSDs does not throttle eventually, but does that affect the bridge chip working properly? The SSD controllers at least have zero problems with the heat, as they throttle at 80 °C to never exceed that threshold and thus are fine, but slower of course.
 

White7561

Junior Member
Apr 17, 2020
20
2
51
Thanks for your replies. Nice to see, that there are indeed users with no problems.

I have read that the cables often make problems and also that USB-C to -A is not ideal.


Are these two points really that important, though?

To 3): Tom’s Hardware tests power consumption and efficiency (MB/s per Watt) and through progress the new PCIe 4.0 SSDs often are more efficient and have less maximum power draw than old PCIe 3.0 ones. E. g. the Samsung 990 Pro has better efficiency than even the mentioned SK Hynix P31 Gold. And as those SSDs are bottlenecked through the enclosure anyway, the same maximum speed of roughly 1 GB/s should leave the more efficient SSDs consume less power, which means less heat.

(I would have used a 990 or 980 Pro, due to their high efficiency like the P31 Gold.)

The respective charts are here under power consumption:


To 4): I see that a thermal pad might be neccessary so that the SSDs does not throttle eventually, but does that affect the bridge chip working properly? The SSD controllers at least have zero problems with the heat, as they throttle at 80 °C to never exceed that threshold and thus are fine, but slower of course.
Usually you can get away without cooling for the bridge chip. But if you can cool it. It's better to do it since I have read in the old days that some bridge chips will disconnect etc and become unstable when too hot. Idk if today's bridge chip acts like that though.
 

delta177

Junior Member
May 8, 2023
1
1
36
As requested RTL9210A firmware 1.29.8A and B and 1.30.21 for B reuploaded and i also added for JMS583 firmware 213 with some modifications done so that name of drive, parameters and stuff are seen by windows programs proper. Inside the package are 2 examples on how the enclosure is seen at eject from windows. Who wants to modify the firmware for it's custom name can do a hex compare , see the name change and do it (see included txt file).

Also as an extra and if you guys are willing to share, i have added also Asmedia and Jmicron firmware dumpers along with some of my tests on my enclosures.

I am looking for ASM2362 firmwares that are generic ones (ending in 00 00) to play with and the only ones i found are very old and on chinese payed websites where i do not have access. So maybe someone is willing to dump some firmwares since it's not distructive and requires only an exe run and drive select.

Code:
RTL9210A and B Reuploaded:
https://disk.yandex.com/d/eIRYQ7-cvJfWng

JMS583 modified firmware 213 to read real drive values.
https://disk.yandex.com/d/lnNYu7PBYpu0bQ

JMS Firmware dumper 0.1a with examples
https://disk.yandex.com/d/ZnfADSHWVIDmAA

Asmedia Firmware dumper with examples
https://disk.yandex.com/d/zZLFx-BhRnpOTg


If you see that you can't download the files, open a free yandex account with three cliks, import the files to your account and then download them from there.
You just saved me from returning a JMS583 enclosure since it wouldn't read my Kingston KC3000 and also not a Kioxia G2.
I downloaded the JMS firmware and installed the sabrent one and after replugging, the drive appeared.
Before, it only showed in the volume manager and it said something like a critical error occured when I tried to assign gpt or mbr boot record.
Thanks!
 
  • Like
Reactions: Ananas