Other Thunderbolt, USB 3.x & Early Intel Macs

[TheShortTimer]

Not wishing to continue taking over the eBay bargains thread with an off-topic discussion of specialised technical issues, I thought it best to spin it off here. I have a couple of Belkin TB docks - one of which is TB3 that I use with my Macs and I also have a HP TB3 dock but it doesn't work or register in System Information when connected directly to a Mac using the Thunderbolt 3 (USB-C) to Thunderbolt 2 adapter and a TB2 cable.

It only works when daisy chained to the Belkin TB3 dock downstream TB3 (USB-C) port and then appears in System Information as follows, with the HP dock listed at the bottom in reflection of its daisy chained status:

Thunderbolt 3 Express Dock HD:

Vendor Name: Belkin International, Inc.
Device Name: Thunderbolt 3 Express Dock HD
Mode: Thunderbolt 3
Device ID: 0x95
Vendor ID: 0x13
Device Revision: 0x1
UID: __________________
Route String: 3
Firmware Version: 19.1
Port (Upstream):
Status: Device connected
Link Status: 0x2
Speed: Up to 20Gb/s x1
Current Link Width: 0x2
Link Controller Firmware Version: 0.22.0
Port:
Status: Device connected
Link Status: 0x2
Speed: Up to 40Gb/s x1
Current Link Width: 0x2
Link Controller Firmware Version: 0.22.0

HP Thunderbolt Dock G2:

Vendor Name: HP Inc.
Device Name: HP Thunderbolt Dock G2
Mode: Thunderbolt 3
Device ID: 0x467
Vendor ID: 0xF0
Device Revision: 0x1
UID: __________________
Route String: 103
Firmware Version: 19.2
Port (Upstream):
Status: Device connected
Link Status: 0x2
Speed: Up to 40Gb/s x1
Current Link Width: 0x2
Link Controller Firmware Version: 0.80.0
Port:
Status: No device connected
Link Status: 0x7
Speed: Up to 40Gb/s x1
Current Link Width: 0x1
Link Controller Firmware Version: 0.80.0

I'm not familiar enough with the Thunderbolt standard to understand why the HP dock requires this arrangement to work. The fact that it's recognised by macOS within System Information when used in this manner would indicate it is compatible somehow but not when connected directly. @joevt and @Amethyst1 I've tagged you both in as you're the resident Thunderbolt experts.

 

[joevt]

My HP Thunderbolt Dock G2 has a Firmware Version of 58.1. Have you tried applying the HP Thunderbolt Dock G2 firmware update? I think you need a PC with Thunderbolt port. Maybe BootCamp might work.

What Mac are you connecting to? What macOS version? I have a iMac 27 inch Late 2013 with Thunderbolt 1 ports running Sonoma 14.2.1. It won't connect directly to my HP Thunderbolt Dock G2. It will connect directly to a OWC Mercury Helios 3S which is a Titan Ridge based Thunderbolt 3 device like the HP Thunderbolt Dock G2 so that's not the issue.

 

[TheShortTimer]

Thanks for replying.

My HP Thunderbolt Dock G2 has a Firmware Version of 58.1. Have you tried applying the HP Thunderbolt Dock G2 firmware update?

No, I hadn't, because it's Windows only. On that note...

I think you need a PC with Thunderbolt port. Maybe BootCamp might work.

I do have Windows 11 on a PC with the GA-H81M-DS2V motherboard and three PCIe slots:

1. 1 x PCI Express x16 slot, running at x16
2. 2 x PCI Express x1 slots
   (All PCI Express slots conform to PCI Express 2.0 standard.)

Plus, I have an ASUS ThunderboltEX II/Dual TB2 PCIe card. Would this work on the PC using Apple's TB3 (USB-C) to TB 2 adapter? Or would I need this TB3 card instead?

Google is my friend, I found the answer:

Yes, it works, on a thunderbolt 3 host or a thunderbolt 4 host with old NVM firmware. It does not work on newer thunderbolt 4 hosts.

I use Apple TB3->TB2 adapter and then an Apple TB2->firewire adapter for a professional focusrite audio interface. It works. On Windows 10 and on Linux. I'm not totally sure how it is today with Windows 11, because they update and change the whole Kernel Protection or whatever options like every week. I think you can still disable those options and then it should still work.

It's all about DMA access and 'security'.

What Mac are you connecting to? What macOS version? I have a iMac 27 inch Late 2013 with Thunderbolt 1 ports running Sonoma 14.2.1. It won't connect directly to my HP Thunderbolt Dock G2. It will connect directly to a OWC Mercury Helios 3S which is a Titan Ridge based Thunderbolt 3 device like the HP Thunderbolt Dock G2 so that's not the issue.

I've connected the dock to the following Macs:

• 13" 2011 MBP w/ High Sierra
• 13" 2012 MBP w/ Catalina and Ventura
• 2013 Mac Pro (Late 2013) w/ High Sierra and Ventura

The outcome is the same with all three machines, I have to daisy chain the G2 to the Belkin TB3 in order to activate it.

 

[TheShortTimer]

The FL1100 USB 3.0 is a PCIe gen 2 x1 controller. USB 3.0 is limited to 4 Gbps (after 8b/10b encoding). PCIe gen 2 x1 is also limited to 4 Gbps (after 8b/10b encoding) but it maybe has more overhead than USB so you won't get full USB 3.0 speed from the FL1100. If you have a Thunderbolt 3 dock, than a USB port that supports 10 Gbps (such as the downstream Thunderbolt port) will probably get you better USB 3.0 performance than a USB 3.0 port.

Got it. I've ordered some adapters and will begin testing with the USB-C port when they arrive.

The adapters arrived and I tested the downstream port on my Belkin TB3 by connecting the EVO 870 via the USB 3.0 JMicron chipset based enclosure to the dock's USB-C port with a USB-C to USB-A adapter using the 2013 Mac Pro and this was the outcome:

That's the best result so far but the drive is ejected from macOS once the speeds reach 4xx MB/s and this happens every time the test is run.

 

[joevt]

The adapters arrived and I tested the downstream port on my Belkin TB3 by connecting the EVO 870 via the USB 3.0 JMicron chipset based enclosure to the dock's USB-C port with a USB-C to USB-A adapter using the 2013 Mac Pro and this was the outcome:

That's the best result so far but the drive is ejected from macOS once the speeds reach 4xx MB/s and this happens every time the test is run.

Try AmorphousDiskMark.app for benchmarks. It will probably give numbers closer to the theoretical maximum. ATTO Disk Benchmark.app is good for testing multiple drives at once.

Does AmorphousDiskMark.app benchmark cause the drive to be disconnected? Does normal usage in the Finder cause the drive to be disconnected?

Instead of using a USB-C to USB-A adapter, can you replace the cable of the USB 3.0 enclosure? Which enclosure is it?

 

[TheShortTimer]

Try AmorphousDiskMark.app for benchmarks. It will probably give numbers closer to the theoretical maximum. ATTO Disk Benchmark.app is good for testing multiple drives at once.

Ok, I'll check that out...

Does AmorphousDiskMark.app benchmark cause the drive to be disconnected?

...and provide an answer to this.

Does normal usage in the Finder cause the drive to be disconnected?

No but I'll leave it connected for an hour to be certain.

Instead of using a USB-C to USB-A adapter, can you replace the cable of the USB 3.0 enclosure?

I've ordered a USB-C to USB-B cable.

Which enclosure is it?

This one and here's the details from System Information:

jmicron:

Product ID: 0x1337
Vendor ID: 0x152d (JMicron Technology Corp.)
Version: 5.08
Serial Number: 7F833EEF5DC0
Speed: Up to 5 Gb/s
Manufacturer: jmicron
Location ID: 0x14300000 / 3
Current Available (mA): 900
Current Required (mA): 896
Extra Operating Current (mA): 0

 

[TheShortTimer]

Does AmorphousDiskMark.app benchmark cause the drive to be disconnected?

Yes, it does. The SSD is disconnected within seconds when trying to run AmorphousDiskMark using the USB-C to USB-A setup. Here's the results that are shown before this happens:

As a comparison, these are the results for the 2013 Mac Pro's USB 3.0 ports:

As you can see, the Belkin's USB 3.0 ports provide the slowest results of all.

Does normal usage in the Finder cause the drive to be disconnected?

Yes - I watched a film from the SSD and Finder disconnected after 50 minutes. Everything was rock solid when I switched to the dock's USB ports instead and continued watching the film.

 

[joevt]

You have sleep disabled? Or disk sleep? Check power management settings.
Could be a bad enclosure. I would consider buying only USB 3.1 gen 2 or better enclosures these days.

 

[Amethyst1]

Plus, I have an ASUS ThunderboltEX II/Dual TB2 PCIe card. Would this work on the PC using Apple's TB3 (USB-C) to TB 2 adapter?

The thing with Thunderbolt add-in cards is that they're only officially compatible with a few specific mainboards from the same manufacturer and usually rely on code in the UEFI to bring them up. The ASUS card may not do much in a Gigabyte board unless a way is found to get it to play ball on "incompatible" systems, as has been with the GC-TITAN RIDGE.

This one and here's the details from System Information:

I am not quite confident which JMicron chip this is. The PCI ID (152b:1337) isn't yielding much in terms of definitive results; this suggests it may be a JMS578. Thus opening the enclosure and checking would be worth it. Then you can look if there's a firmware update to improve its behaviour (like this).

 

[TheShortTimer]

You have sleep disabled? Or disk sleep? Check power management settings.

The layout and options for power management settings were better in earlier macOS versions IMHO.

I've changed the settings below to never turn off display when inactive and never require the password after display is turned off.

If I've missed anything, please let me know.

Could be a bad enclosure. I would consider buying only USB 3.1 gen 2 or better enclosures these days.

Noted.

The thing with Thunderbolt add-in cards is that they're only officially compatible with a few specific mainboards from the same manufacturer and usually rely on code in the UEFI to bring them up. The ASUS card may not do much in a Gigabyte board unless a way is found to get it to play ball on "incompatible" systems, as has been with the GC-TITAN RIDGE.

View attachment 2427274

So much for the industry working towards standardisation!

I suppose it's worth giving it a try nonetheless and seeing what happens. All I have to do is install the card onto the motherboard - which is easy enough. If it doesn't work then I'll have to either find someone with a Windows computer that has TB or see if I can update the firmware using one of the Macs under Bootcamp. Maybe it the firmware could be updated on my Windows 11 PC via USB-C.

I am not quite confident which JMicron chip this is. The PCI ID (152b:1337) isn't yielding much in terms of definitive results; this suggests it may be a JMS578. Thus opening the enclosure and checking would be worth it. Then you can look if there's a firmware update to improve its behaviour (like this).

Let's see if I can take some photos of the PCB.

 

[TheShortTimer]

Instead of using a USB-C to USB-A adapter, can you replace the cable of the USB 3.0 enclosure?

The USB-C to USB-A cable arrived and the outcome is exactly the same - the SSD is ejected from Ventura.

Let's see if I can take some photos of the PCB.

No joy. The text on the chips is too indistinct, even with ample lighting. I'll have to contact Combrite and ask if they can provide me with the information.

 

[Amethyst1]

The USB-C to USB-A cable arrived and the outcome is exactly the same - the SSD is ejected from Ventura.

Does this also happen on the Haswell PC running Windows?

I'll have to contact Combrite and ask if they can provide me with the information.

It’s possible that the manufacturer uses several bridges in the same enclosure. Fun to see the device ID is 1337 of all things. I‘d hope the firmware update for a 578 would only proceed on one of these.

 

[TheShortTimer]

Does this also happen on the Haswell PC running Windows?

No, it doesn't. An ATTO Disk Benchmark test runs in its entirety.

The results aren't great but it didn't disconnect during the test unlike Ventura. (Yes, the enclosure was connected to a USB 3.0 port.)

It’s possible that the manufacturer uses several bridges in the same enclosure. Fun to see the device ID is 1337 of all things. I‘d hope the firmware update for a 578 would only proceed on one of these.

I might take one for the team and see what happens.

 

[TheShortTimer]

I might take one for the team and see what happens.

Well, I took one for the team and lived to tell the tale.

I ran the firmware updater on my Haswell PC under Win 11 and felt cocky enough that I didn't bother to backup the existing firmware. It worked - both the enclosure and the SSD were recognised afterwards in Windows so I switched over to the Trash Can and tried AmorphousDiskMark again with my re-programmed enclosure connected to the TB3 dock's downstream port.

It made it all the way to the end. Disconnections no more!

Note the enclosure's name now replaces that of the drive in the bottom bar.

The read/write speeds have taken a noticeable hit. Here's the read/write stats using the TC's internal USB 3.0 ports:

Down from 385.78, 394.92 for SEQ1M QD8.

Now for the TB3 dock's USB 3.0 ports:

Typically, the slowest of the three but even slower than before. Down from 376.82 and 336.52 for SEQ1M QD8.

Perhaps I should've backed up the original firmware beforehand! Ah well, I can always restore it using a copy from one of my other enclosures. Interestingly, it sill refuses to play with the Kanex dongle - connecting the pair together still triggers constant ejects.

 

[Amethyst1]

Disconnections no more!

That's awesome. See, my suggestion was helpful for a change

Note the enclosure's name now replaces that of the drive in the bottom bar.

Has the PCI device ID also changed?

The read/write speeds have taken a noticeable hit.

I'd value no disconnects (and, hopefully, no other surprises) more than a ≈10% reduction in (mostly irrelevant) sequential speeds.

 

[TheShortTimer]

That's awesome. See, my suggestion was helpful for a change

Such humility, your suggestions are always helpful!

Has the PCI device ID also changed?

Yes it has!

This was the original…

jmicron:

Product ID: 0x1337
Vendor ID: 0x152d (JMicron Technology Corp.)
Version: 5.08
Serial Number: 7F833EEF5DC0
Speed: Up to 5 Gb/s
Manufacturer: jmicron
Location ID: 0x14300000 / 3
Current Available (mA): 900
Current Required (mA): 896
Extra Operating Current (mA): 0

Here's the updated version with the new firmware:

USB to ATA/ATAPI Bridge:

Product ID: 0xa37a
Vendor ID: 0x125f (ADATA Technology Co., Ltd.)
Version: 94.02
Serial Number: 0123456789ABCDEF
Speed: Up to 5 Gb/s
Manufacturer: JMicron
Location ID: 0x14300000 / 25
Current Available (mA): 900
Current Required (mA): 8
Extra Operating Current (mA): 0

Quite a difference.

I'd value no disconnects (and, hopefully, no other surprises) more than a ≈10% reduction in (mostly irrelevant) sequential speeds.

Who I am to disagree with the infallible @Amethyst1? I'll go ahead and flash the other Combrite enclosure. Thanks for the suggestion.

Which of those results are relevant?

 

[TheShortTimer]

I'm hiving off the conversation from this thread so as not to derail it and because of the relevance to the general discussion here.

The only "dumb" solution I'm aware of is a TB3 add-in card, such as this:

You'd need a riser or enclosure to put the card in and a way to supply it with power.

Could this be accomplished on the cheap with an Alpine Ridge using something obtained from eBay etc that could be appropriately repurposed - like you did with the AKITIO ThunderLink box?

 

[Amethyst1]

Could this be accomplished on the cheap with an Alpine Ridge using something obtained from eBay etc that could be appropriately repurposed - like you did with the AKITIO ThunderLink box?

I’ve actually purchased an Asus ThunderboltEX 3 card (Alpine Ridge) and a HP Thunderbolt Dock G2 (Titan Ridge) to test if DP-to-TB conversion works in an AKiTiO Thunder2. The card and dock also do USB-C DP Alt Mode so I‘ll have to check that four DP lanes are going into the dock to be sure the input is TB.

 

[Amethyst1]

Could this be accomplished on the cheap with an Alpine Ridge using something obtained from eBay etc that could be appropriately repurposed - like you did with the AKITIO ThunderLink box?

The Asus TB3 card has arrived. First experiment (replicating this) is successful: it does DP > USB-C DP Alt Mode conversion in ghost mode.

MBP <> TB <> TB2 box <> DP+PCIe <> TB3 card <> USB-C <> Dock <> DP <> Display

This is probably the most unnecessarily complex way of connecting a display to a computer

 

[TheShortTimer]

This is probably the most unnecessarily complex way of connecting a display to a computer

Let's not forget, we're special people and anything less just wouldn't be us.

 

[f54da]

@Wowfunhappy based on the result above it seems like this card should work for your purposes? Although considering you have to lug that thing around it might just be easier to fiddle with the docking station itself

 

[Amethyst1]

@Wowfunhappy based on the result above it seems like this card should work for your purposes?

I’ve only confirmed USB-C DP Alt Mode works so far (the Wacom Link Plus does the same thing). We need TB video output — that will be tested once my TB3 dock arrives.

 

[Wowfunhappy]

@Wowfunhappy based on the result above it seems like this card should work for your purposes? Although considering you have to lug that thing around it might just be easier to fiddle with the docking station itself

Yeah, unfortunately I don't think it's worth it for my use case due to the fact that it has to be powered externally combined with the size.

 

[weckart]

I use a Mac Mini 6,2 (2012) as a browsing machine and am currently writing this post on it. I was wondering if I could get a Mac Mini doc/hub - the type which sits under the Mini and blends in with it. All those I have seen are TB3 or USB-C. I was wondering whether Thunderbolt is forwards compatible as well as backwards - at least up to the speed of its slowest connection. I was thinking of using Apple's TB3 to TB2 adapter to hook my Mini up to the dock. My understanding is that the adapter was primarily intended for those with newer Macs who had old TB peripherals. I'm not sure if the Mini would recognise a TB3 connection which postdated it.

 

[TheShortTimer]

I use a Mac Mini 6,2 (2012) as a browsing machine and am currently writing this post on it. I was wondering if I could get a Mac Mini doc/hub - the type which sits under the Mini and blends in with it. All those I have seen are TB3 or USB-C. I was wondering whether Thunderbolt is forwards compatible as well as backwards - at least up to the speed of its slowest connection. I was thinking of using Apple's TB3 to TB2 adapter to hook my Mini up to the dock. My understanding is that the adapter was primarily intended for those with newer Macs who had old TB peripherals. I'm not sure if the Mini would recognise a TB3 connection which postdated it.

I can assure you from personal experience that it's bi-directional and intended to also connect older Macs to newer TB hardware.

In this thread, @Amethyst1 and I connected our respective MacBooks to TB3 eGPU enclosures that postdated our machines using the Apple TB3 to TB2 adapter. I use TB3 docks via the adapter on TB1 Macs. It's just a shame that the adapter has become so expensive. I purchased my first one brand new from Amazon for £18 in 2021 and the price has soared since.

Which dock(s) brands do you have in mind?