It was a preorder. If you purchase now I'm pretty sure you will get 1.0, but then at this point I wouldn't even bother getting OWC Envoy Express. Like I said I haven't tested the 1.0 revision with a 4TB SSD of one of those previously failed models. This is just too risky.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
[OWC Envoy Express TB3 Enclosure] Sabrent 2TB Rocket issues. Help?
- Thread starter amartinez1660
- Start date
- Sort by reaction score
I have a new hypothesis. The thermal pad of Envoy Express is pretty thick. On paper this sounds like a good thing, but what might be happening is that it applies excessive pressure onto the drive especially when the chips on the drive are thick as well. This itself isn't problematic until when the drive is double-sided, in which case when the drive heats up to a certain degree, connectors on the other side of the drive might get "desoldered" if the drive is bent, killing the drive instantly.
The only safe options are probably single-sided drives. For example I used a Samsung MLC drive which is single-sided with the same enclosure without hitting any problem whatsoever.
The only safe options are probably single-sided drives. For example I used a Samsung MLC drive which is single-sided with the same enclosure without hitting any problem whatsoever.
There are both TLC and QLC versions of Sabrent Rocket, and I think I can draw the conclusion that the problem is related to if the drive is single or double sided, rather than if the drive is QLC, TLC or if it has DRAM, so ironically since a single sided QLC high volume SSD is much easier to find compared to TLC, you will have better chance using a QLC drive with this enclosure. In my case why the Sabrent Rocket Q failed it was because mine is the 4TB version which has chips densely packed on both sides. My other Sabrent Rocket Q4 drive (PCIe 4 version of Rocket Q), which is also 4TB, has better luck because large portion of the back of the drive is empty.
I will post a detailed explanation soon.
So at this point I think it's safe to draw the conclusion that Envoy Express' overly thick thermal pad might have caused the problems.
TL;DR version:
The thermal pad might have caused SSD to bend and if there are chips on the back of the SSD, they might get "desoldered" as the SSD heats up. If you have a double-sided SSD, either avoid Envoy Express or use without the lid.
In my case the SSDs that worked:
WD Blue SN550 1TB (single-sided)
Intel 660p 2TB (single-sided)
Samsung SM961 (single-sided)
Sabrent Rocket Q4 4TB (double-sided with empty space on the back)
What died from the enclosure:
Sabrent Rocket Q 4TB (double-sided)
Corsair MP440 4TB (double-sided)
The author of the original post had a Sabrent Rocket 2TB, assuming it was the TLC version, it's double-sided with no empty space on the back.
(Hypothesis) How an SSD dies,
1. The thermal pad of the enclosure applies excessive amount of pressure onto the front of the SSD, causing it to bend
2. The SSD heats up when it is in use, slightly softening the solder on the SSD
3. Some connectors on the edges of the chips on the back of the SSD get desoldered due to the force caused by the bending and the softened solder
It was a great mystery that two 4TB SSDs were killed by the enclosure while the Sabrent Rocket Q4 4TB works just fine, until I realized it has a large empty space on the back. Looking at what worked and what have been killed, all of this started making sense to me.
If you are using a double-sided SSD with Envoy Express, remove the lid immediately. Your drive might encounter thermal throttling more frequently, but it's better than a dead drive. I personally don't recommend using this enclosure with a double-sided SSD with sparsely packed back side. It might work today but who knows at some point the metal fatigue might throw the last straw. What really scares if that data can easily be recovered by reflowing the SSD, so you return a drive thinking it's totally dead and there is nothing you can do, but other people can just get your data.
As I mentioned the OWC Envoy Express has at least two revisions, but the thermal pads are of the same thickness. Don't expect the 1.0 revision can eliminate the issue. I will try to contact OWC to signal the issue and hope they come up with a newer revision to either have a thinner thermal pad or provide swappable thermal pads with different thickness.
Here are some photos to illustrate what drives are safe to use (green), what might work in short term (orange) and what should be completely avoided.
TL;DR version:
The thermal pad might have caused SSD to bend and if there are chips on the back of the SSD, they might get "desoldered" as the SSD heats up. If you have a double-sided SSD, either avoid Envoy Express or use without the lid.
In my case the SSDs that worked:
WD Blue SN550 1TB (single-sided)
Intel 660p 2TB (single-sided)
Samsung SM961 (single-sided)
Sabrent Rocket Q4 4TB (double-sided with empty space on the back)
What died from the enclosure:
Sabrent Rocket Q 4TB (double-sided)
Corsair MP440 4TB (double-sided)
The author of the original post had a Sabrent Rocket 2TB, assuming it was the TLC version, it's double-sided with no empty space on the back.
(Hypothesis) How an SSD dies,
1. The thermal pad of the enclosure applies excessive amount of pressure onto the front of the SSD, causing it to bend
2. The SSD heats up when it is in use, slightly softening the solder on the SSD
3. Some connectors on the edges of the chips on the back of the SSD get desoldered due to the force caused by the bending and the softened solder
It was a great mystery that two 4TB SSDs were killed by the enclosure while the Sabrent Rocket Q4 4TB works just fine, until I realized it has a large empty space on the back. Looking at what worked and what have been killed, all of this started making sense to me.
If you are using a double-sided SSD with Envoy Express, remove the lid immediately. Your drive might encounter thermal throttling more frequently, but it's better than a dead drive. I personally don't recommend using this enclosure with a double-sided SSD with sparsely packed back side. It might work today but who knows at some point the metal fatigue might throw the last straw. What really scares if that data can easily be recovered by reflowing the SSD, so you return a drive thinking it's totally dead and there is nothing you can do, but other people can just get your data.
As I mentioned the OWC Envoy Express has at least two revisions, but the thermal pads are of the same thickness. Don't expect the 1.0 revision can eliminate the issue. I will try to contact OWC to signal the issue and hope they come up with a newer revision to either have a thinner thermal pad or provide swappable thermal pads with different thickness.
Here are some photos to illustrate what drives are safe to use (green), what might work in short term (orange) and what should be completely avoided.
i returned the OWC and just purchased two of the Sabrent drives that come with the enclosure. So far the 8TB and 4TB are working fine with no problems. Talked to tech support to understand the difference between QLC and TLC-- the simple answer is use QLC if you reading a lot of information and use the TLC if you are writing a lot of data. Also a good tip was to remove the drive from the spotlight search on a mac which will reduce wear and tear if you leave the drive connect all the time -- which I do.
Just to confirm, the two inlands SSDs were both double-sided, right? (TLC drive with Phison controller)
Okay, 4TB is always double-sided regardless if it's TLC or QLC. This further proves my theory...
I like your theory. Is the thermal pad on the Envoy Express removable and replaceable with a thinner one?
Unfortunately it's glued to the lid. I tried to peel it off, but that doesn't seem to be possible without damaging the thermal pad.
I'm probably going to get a single sided SSD (Either the Samsung 970EVO or the WD SN750) just to be safe. As far as damaging the thermal pad, I wouldn't mind doing that, since I'd want to replace it anyway, right?
Well, you can just leave the thermal pad there for single-sided SSDs otherwise I assume the removal of a glued thermal pad would void the warranty, but at this point I don't actually recommend Envoy Express. I did a stability test yesterday and my Mac mini ejected my SSD after overnight. The SSD didn't die and it's single-sided so the reason must be something else. The culprit could be the Mac mini, also. I have another two enclosures on the way, one is a Sabrent thunderbolt case, which I will be testing tonight. Another is a JEYI thunderbolt case, which is also sold at 80 bucks but has JHL6340, which enables 4 lane PCIe connection. Envoy Express and the Sabrent case are both 2-lane PCIe. If you are going to use high-end SSDs like 970 or SN750, 4-lane PCIe makes more sense (if you are willing to wait for a few weeks until I get the enclosure and do some stability tests).
The Trebleet enclosure is $130, and uses JHL6540 and is supposedly 4 lanes as well. The "off-brand" enclosures seem to use a standard 21A-25029 board. The Sabrent and the Envoy Express have different boards.
I think I'll buy the SSD now while it's on sale, and then wait for your testing before I decide on an enclosure.
I was under the impression that OWC and Sabrent went with 2-lane PCIe because of the need for thermal performance and to keep the enclosure bus powered with a low power chipset. I'm curious to see how the JEYI performs.
I think I'll buy the SSD now while it's on sale, and then wait for your testing before I decide on an enclosure.
I was under the impression that OWC and Sabrent went with 2-lane PCIe because of the need for thermal performance and to keep the enclosure bus powered with a low power chipset. I'm curious to see how the JEYI performs.
I have 2 of the pre-order (0.8 revision?) OWC Envoy TB3 enclosures, and I bought an A-Data XPG 8200 Pro for it, after seeing a YouTube review of a guy using that very same combo. Well, the A data ssd done not work in that enclosure, but fine in everything else. I have an Intel P660 2TB drive in the other, and have used that for nearly for 3-4 months without a single issue. Spoke to OWC about the A-Data and they knew nothing at the time.
Gonna buy a Samsung 970 EVO Plus 2TB for the extra OWC TB3 enclosure, since it's a much faster (especially with sustained loads) SSD that the Intel.
As my M1 MBA is incoming, and there are so many issues with enclosures only connecting at 5 GB/s, I need to ramp on on my TB3 storage anyway ?
Gonna buy a Samsung 970 EVO Plus 2TB for the extra OWC TB3 enclosure, since it's a much faster (especially with sustained loads) SSD that the Intel.
As my M1 MBA is incoming, and there are so many issues with enclosures only connecting at 5 GB/s, I need to ramp on on my TB3 storage anyway ?
Last edited:
I wish I had seen this thread a long time ago. My original goal was to put an SSD in the OWC Envoy Express enclosure and connect it through a Thunderbolt 3 dock to my late 2015 iMac, which only has Thunderbolt 2 and a non-upgradeable internal SSD. Skipping several weeks waiting for parts to arrive, many failures, and a few successes, this is what happened in my setup.
SSDs tried: two 2TB ADATA XPF SX8200 Pro (chips on both sides of PCB) and one 1TB Inland Premiere (chips on one side of PCB)
Docks tried: used HP Elite Thunderbolt 3 Dock (P/N P5Q58UT#ABA), new Caldigit TS3+
Computers involved: late 2015 iMac (internal 250GB SSD) and M1 Mac Mini (internal 512GB SSD)
Enclosure: OWC Envoy Express ordered directly from OWC in late November 2020 (v0.8)
2TB ADATA + OWC Envoy Express (closed with thermal pad directly on chips) + HP dock + iMac: 50GB into a 100GB write the ADATA was no longer recognized by macOS. The enclosure had no time to heat up beyond room temperature. The enclosure smelled of fried electronics. Visual inspection revealed the thermal pad of the enclosure contacted only half of two chips on either end of the PCB, as described earlier in this thread
1TB Inland + OWC Envoy Express (closed with thermal pad directly on chips) + M1 Mini: enclosure reached 140F but a write of 700GB finished successfully
1TB Inland + OWC Envoy Express (open with electrostatic fan on high directly blowing on chips) + M1 Mini: surface temperature of the Inland reached 140F but a write of 700GB finished successfully
1TB Inland + OWC Envoy Express (open with electrostatic fan on high directly blowing on chips) + Caldigit dock + iMac: surface temperature of the Inland reached 140F but a write of 700GB finished successfully
1TB Inland + OWC Envoy Express (closed with thermal pad directly on chips) + Caldigit dock + iMac: enclosure reached 140F and a write failed ~75GB into a 700GB write operation. After both the Inland and the enclosure cooled to room temperature, the Inland was again operational through the enclosure
2TB ADATA + OWC Envoy Express (open with electrostatic fan on high directly blowing on chips) + M1 Mini: surface temperature of the ADATA never rose above 80F but the unit still died ~450GB into a 700GB write operation. The ADATA SSD is still unresponsive
My advice: do not use the 2TB ADATA XPF SX8200 Pro with the OWC Envoy Express v0.8 enclosure under any circumstances. After my experience, I have no idea how the ADATA survived in any of the tests I saw on YouTube, let alone being used on a daily basis. I will continue to use the 1TB Inland with the open enclosure and a fan, and if it dies, I'll update this post.
SSDs tried: two 2TB ADATA XPF SX8200 Pro (chips on both sides of PCB) and one 1TB Inland Premiere (chips on one side of PCB)
Docks tried: used HP Elite Thunderbolt 3 Dock (P/N P5Q58UT#ABA), new Caldigit TS3+
Computers involved: late 2015 iMac (internal 250GB SSD) and M1 Mac Mini (internal 512GB SSD)
Enclosure: OWC Envoy Express ordered directly from OWC in late November 2020 (v0.8)
2TB ADATA + OWC Envoy Express (closed with thermal pad directly on chips) + HP dock + iMac: 50GB into a 100GB write the ADATA was no longer recognized by macOS. The enclosure had no time to heat up beyond room temperature. The enclosure smelled of fried electronics. Visual inspection revealed the thermal pad of the enclosure contacted only half of two chips on either end of the PCB, as described earlier in this thread
1TB Inland + OWC Envoy Express (closed with thermal pad directly on chips) + M1 Mini: enclosure reached 140F but a write of 700GB finished successfully
1TB Inland + OWC Envoy Express (open with electrostatic fan on high directly blowing on chips) + M1 Mini: surface temperature of the Inland reached 140F but a write of 700GB finished successfully
1TB Inland + OWC Envoy Express (open with electrostatic fan on high directly blowing on chips) + Caldigit dock + iMac: surface temperature of the Inland reached 140F but a write of 700GB finished successfully
1TB Inland + OWC Envoy Express (closed with thermal pad directly on chips) + Caldigit dock + iMac: enclosure reached 140F and a write failed ~75GB into a 700GB write operation. After both the Inland and the enclosure cooled to room temperature, the Inland was again operational through the enclosure
2TB ADATA + OWC Envoy Express (open with electrostatic fan on high directly blowing on chips) + M1 Mini: surface temperature of the ADATA never rose above 80F but the unit still died ~450GB into a 700GB write operation. The ADATA SSD is still unresponsive
My advice: do not use the 2TB ADATA XPF SX8200 Pro with the OWC Envoy Express v0.8 enclosure under any circumstances. After my experience, I have no idea how the ADATA survived in any of the tests I saw on YouTube, let alone being used on a daily basis. I will continue to use the 1TB Inland with the open enclosure and a fan, and if it dies, I'll update this post.
I am using MP400 8TB, with trebleet 2 in 1 Thunderbolt 3 & USB3.2 SSD Enclosure
it is quite unstable on mac, I formated to APFS, copy about 88GB files, then do some read and write test, then formatted it to Mac OS Extended (Journaled) Then the speed become super low, then I eject, and replug again, the drive never recognize again
the same thing happens when I doing file copy meanwhile reading a big folder, it will eject as well
the case is really unstable on mac, sometimes recognize sometimes not
I am using Sabrent USB-C case, also but sometimes recognize sometimes not
I have reset my Mac NVRAM and SMC, and try to use two different Mac, still the same issue.
One super wired thing is with the Sabrent case, when I use USB TYPE-A 3.0 cable connect to a HUB, it will always recognize, but the speed is slow like sata drive
this happens on RMBP 2018 and Mac Pro 2019, can anyone explain the reason？
I feel that all Nvme Enclosure I test on a mac Sabent(JMS583 Chip set) and trebleet(JHL7440+JMS583 Chipset) is quite unstable
it is quite unstable on mac, I formated to APFS, copy about 88GB files, then do some read and write test, then formatted it to Mac OS Extended (Journaled) Then the speed become super low, then I eject, and replug again, the drive never recognize again
the same thing happens when I doing file copy meanwhile reading a big folder, it will eject as well
the case is really unstable on mac, sometimes recognize sometimes not
I am using Sabrent USB-C case, also but sometimes recognize sometimes not
I have reset my Mac NVRAM and SMC, and try to use two different Mac, still the same issue.
One super wired thing is with the Sabrent case, when I use USB TYPE-A 3.0 cable connect to a HUB, it will always recognize, but the speed is slow like sata drive
this happens on RMBP 2018 and Mac Pro 2019, can anyone explain the reason？
I feel that all Nvme Enclosure I test on a mac Sabent(JMS583 Chip set) and trebleet(JHL7440+JMS583 Chipset) is quite unstable
My experiences have been mixed. I started in August with a 1TB SN750 in a Pluggable USB-C enclosure (while waiting on my OWC pre-order to ship) dedicated to my Parallels VM files. It arrived in mid-September and I actually bought a new drive for it (1TB Intel 660p) as it was a good price and I wanted to do some performance comparisons. I moved the 750 to the OWC and put the 660p in the Pluggable. Everything worked great until roughly November when I started getting "unexpected disconnection" errors. I swapped the 750 for the 660p and that worked for a couple of months until it threw a "unexpected disconnect" as well. I did an RMA on the enclosure in early February and got a 1.0 enclosure. That worked fine until the beginning of March when my Mac Mini 2018 started repeatedly crashing after installing Big Sur 11.2.2. I talked to Apple Support and reinstalled Big Sur at their direction, but 9 hours later it crashed again. Since the only hardware that has given me any problems as been the OWC, I disconnected it, switched the drive to yet another USB-C enclosure, and the system has not crashed for a couple of days.
Both drives have been used in two different USB-C enclosures without any issues hosting my VMs for weeks at a time both before I got the OWC and during long test periods so if it's a drive issue, it's drive compatibility. Both are single sided boards.
I sent a note to OWC support to find out if there were any known issues but they have not gotten back to me. Their responsiveness used to be great but it's really fallen off in the last few months to written inquiries (although they were great with the RMA when I called in Feb).
I think I'm putting the enclosure aside until I either get the OWC TB3 hub in May or the upgrade to 11.3 gets released. Other than my nightly backups running about 15% faster, I really can't actually see any difference with my workload.
Anybody else have issues with it and 11.2.x?
Both drives have been used in two different USB-C enclosures without any issues hosting my VMs for weeks at a time both before I got the OWC and during long test periods so if it's a drive issue, it's drive compatibility. Both are single sided boards.
I sent a note to OWC support to find out if there were any known issues but they have not gotten back to me. Their responsiveness used to be great but it's really fallen off in the last few months to written inquiries (although they were great with the RMA when I called in Feb).
I think I'm putting the enclosure aside until I either get the OWC TB3 hub in May or the upgrade to 11.3 gets released. Other than my nightly backups running about 15% faster, I really can't actually see any difference with my workload.
Anybody else have issues with it and 11.2.x?
I have a positive update on this. Shortly after I had major problems with the enclosure, I decided to start the RMA process for the OWC enclosure and along with it sent one of the ADATA SSDs that appeared dead. About a week or so later, I received both components. The serial number of the returned enclosure matched the one I sent for RMA. Nothing changed with macOS on either of my Macs during this time. Both previously inoperable ADATA SSDs were immediately recognized by macOS under all configurations I mentioned in a previous post. I have been using both ADATA SSDs and the Inland SSD without issue through both the HP and Caldigit docks for the past 1.5 months. Additionally, the SSDs no longer get ridiculously hot under load. I now recommend this setup with reservations -- if users of the enclosures experience issues they should contact the helpful support team at OWC.
I bought a pre-order Envoy Expreas that failed multiple times on an OSX clone to a 2TB ADATA 8200 Pro at the end of last year. It was within a week of receiving the thing in the mail, so I did an immediate RMA. The RMA replacement had the same problem, so I just wrote off the product and shelved it. In April I saw forum chatter and the vendor FAQ about a hardware fix that resolved the problem with my specific drive, so I did a 2nd RMA which OWC kindly paid for return shipping on.
I finally got my second replacement today (same serial# on the drive case, so they must have either swapped or repaired the main board) and I was able to perform a successful clone using CCC on a 2019 iMac. I have since rendered a huge Davinci Resolve project using this drive with no problems at all. The case seems much cooler with this ADATA than the original one I received.
By the way, I had 2 different ADATA 2TB drives fail a disk clone and OSX install in the original enclosure, but neither was actually damaged and I am not convinced of other anecdotal reports of permanently damaged drives due to flexing, solder, etc. OSX detection of the drive completely crapped out in both cases after ~3 minutes of writes, but I was able to successfully revive each time by installing the NVMe in a Windows machine, running diskpart, selecting the drive, and doing a “CLEAN ALL”. This returned the drive to an OSX “uninitialized drive” state and my current good drive had been through at least 3 cycles of this process since December.
OWC did not say what was done to the enclosure for the RMA.
Last edited:
According to Tom's Hardware, the EX950 does have DRAM : https://www.tomshardware.com/reviews/hp-ex950-nvme-ssd-2tb,5306.html#specifications
Thanks very much for this old but still relevant thread which I have just found after getting a 4TB Rocket blade (TLC version) to put in an Envoy Express I have had for some time.
I immediately noticed the problem of the extra thickness as the Envoy would not close snuggly like it has done with single sided blades. I squeezed it while tightening the screws, thinking I was squeezing the pad, and perhaps enhancing thermal contact. I did not let the screws pull it down. It seemed OK and I loaded my 3TB of data. (It did spontaneously eject once, but as this has started happening with both my M1 Macs this week with 12.1 I didn't put too much emphasis on this).
Then I found this thread, and promptly removed the thermal pad, which came away cleanly and will be reusable.
I seem to have got lucky and not damaged the blade.
Not sure what to do now. I only use the Envoy for data (not boot) and not heavily, so it might be OK, especially as the heat generated is less than for faster enclosures.
Trebleet seems a recommended name, may have a close look at that. I already have a Orico (metal ribbed type) which Orico says is limited to 2TB. Many enclosures are limited to 2TB...... I wonder if blade thickness is the main reason.
The OWC Envoy Express page says:
"USES TODAY’S AND TOMORROW’S DRIVES
Whatever drive you need to work on, the Envoy Express is ready for action. It’s designed to support any NVMe M.2 2280 SSD and gives you the freedom to use any recently made drive — including OWC Aura SSDs with up to 4.0TB capacity. It’s also compatible with tomorrow’s 8.0TB, 16TB, and future capacities… without limit"So either they expect you squeeze the thermal pad, or there has been a revision since I got mine.