OTHER WORLD COMPUTING OWCTB3DK13PSL OWC 13 Port Thunderbolt 3 Dock - Disconnects when READING from External HD at high speeds

[KYBOSH]

Weird one and way out of my league but I thought i drop it here in case anyone might have an idea on where to start looking.

I have a Thunderbolt Dock that has been behaving fine for a long time.
Admittedly I have only used it for menial data transfer tasks.

A few days ago I connected a USB-C external hard drive to it and wanted to test the transfer speeds.
I connected the proper high speed cables (USB-C to USB-C) and started the test with BlackMagic Disk Speed Test.
It wrote to the disk just fine. Got nearly 900MB/s+ write speeds but when it came time to READ the disk just disconnected from the computer.
I tested it with another drive (portable SSD) and it did the exact same thing. I think connected the drive to the dock via a USB-C to USB-A cable and there was no issue.
The read and write speed was abymsmally low (~40MB/s) but the drives did not disconnect when reading or writing.

I have eliminated the cables, the drives and even the computers as the source of the issue.
Issue persists regardless of 2 thunderbolt ports you use.
The computer is connected to the other thunderbolt port via the other TB port.

If I were crazy enough to open this box up what might i want to look for?
Outside of the high speed data transfer/disconnection issue... the unit works well.

Here's the reveiw for a sister model...
The top picture Im posting is my actual unit.
The bottom pic is from the sister model (mostly identical but fewer ports)
https://m.eprice.com.tw/tech/talk/11...12/1/m/5653487

[KYBOSH]

Here's an image to illustrate what's happening here. Writing to the disk is okay but as soon as it tries to read from it there is an instant disconnection.
It is never able to actually read from the disk in any significant manner. I tried to open a small pdf off of the drive and it would only partially load before the drive disconnected.
The pdf was only a KB is size.

[mon2]

Are you using the round 20V 6.75A power supply? If not, suggest that you do. Guessing that the external power adapter is required to make this widget functional.

Confirm you have the latest driver package for your OS.

In using an external USB Type C HD / SSD for your testing - the interconnect cabling is critical. Be sure that the USB Type C cable is rated for USB 3.x SS (ideally, Gen 2 = 10 Gbps) AND support for power delivery so it should be marked with 100W or more. Only then will you have an internal e-marker IC inside of the cable to permit the passing through of the required power (voltage and higher current) to the target HD / SSD being tested. If the cabling is not as referenced, then you will most likely be only USB 2.0 HS (480 Mbps) at best and limited on the current draw for the target devices. Only a cable marked with PD support can offer the higher voltage and higher current (5A or more) support that is desirable for HD / SSD that are external.

Dock Ejector (macsales.com)

[KYBOSH]

Are you using the round 20V 6.75A power supply? If not, suggest that you do. Guessing that the external power adapter is required to make this widget functional.

Confirm you have the latest driver package for your OS.

In using an external USB Type C HD / SSD for your testing - the interconnect cabling is critical. Be sure that the USB Type C cable is rated for USB 3.x SS (ideally, Gen 2 = 10 Gbps) AND support for power delivery so it should be marked with 100W or more. Only then will you have an internal e-marker IC inside of the cable to permit the passing through of the required power (voltage and higher current) to the target HD / SSD being tested. If the cabling is not as referenced, then you will most likely be only USB 2.0 HS (480 Mbps) at best and limited on the current draw for the target devices. Only a cable marked with PD support can offer the higher voltage and higher current (5A or more) support that is desirable for HD / SSD that are external.

Dock Ejector (macsales.com)

Thank you for the reply!
The power supply I am using is the original OEM one.

I have the latest Mac OS along with dock ejector installed and everything else works just fine.

The same behavior was exhibited when trying to use the OWC enclosure with this dock as well.
The enclosure has its own power supply and does not (should not) rely on power delivery protocols.
That sort of eliminated the cables, the power supply and the drives as the source of the problem.

I have another dock (Caldigit TS3+ and one of the thunderbolt docks exibits the EXACT same issue but thankfully this devise has 2 other thrunderbolt ports (which works just fine) so I can limp along with 1 port missing.
Something happens with these ports/controllers... IDK if its due to heavy usage, some sort of electrical surge or what... they get fried or something. I would love to know as it maybe possible to replace an IC easily.

Regarding the cables used... for the test I provided screenshots for I was using a no-name brand cable but before this I tested with an OWC supplied 1.0M USB-C to USB-C cable and was getting 900+Mb/s.
So I want to say this cable would be rated for 3.x.

[mon2]

For the working ports, are you using the same cabling? Each port may and should have esd TVS diodes to shunt transients. After such an event the tvs diodes usually need to be replaced. These shunted tvs diodes will impact the signal integrity.

Consider to test each in of the working and non-working ports in diode mode. Post each measurement.

Meter in diode mode. Remove all power. Red meter lead, yes red to ground. Black meter lead to the point to test.

[KYBOSH]

Thank you for this insght. I havent opened up the Caldigit yet but I wasnt aware these diodes where common on these things.
I'm looking at the PCB of the OWC (pic provided in poast #1) and Im not seeing where those diode packs Im used to seeing on TV panel PCBs.
Would they be located very near the thunderbolt port?

[mon2]

Personally have never seen one of these widgets. The TVS diodes are sold in SMD arrays and that is what we use for our many designs. They tend to be compact and get the job done. We build for assorted markets and one of the targets is the very noisy CNC cutters. We have everything including the kitchen sink onboard to protect against transients and can say we sleep at night well knowing that we have had zero failures due to the same TVS devices. Earlier models would have a few phantom kills to the board. Best if you can carefully open the unit and take close up clear pics near the connectors. By PCB layout rules, the ESD devices should be nearby each connector. This is only an educated stab at the issue that you are facing. Regardless, then you can perform the diode mode of testing on each port.

[KYBOSH]

Personally have never seen one of these widgets. The TVS diodes are sold in SMD arrays and that is what we use for our many designs. They tend to be compact and get the job done. We build for assorted markets and one of the targets is the very noisy CNC cutters. We have everything including the kitchen sink onboard to protect against transients and can say we sleep at night well knowing that we have had zero failures due to the same TVS devices. Earlier models would have a few phantom kills to the board. Best if you can carefully open the unit and take close up clear pics near the connectors. By PCB layout rules, the ESD devices should be nearby each connector. This is only an educated stab at the issue that you are facing. Regardless, then you can perform the diode mode of testing on each port.

THanks for your input mon2. I have to admit this is/was above my skill level to diagnose and even rectify. I threw in the towel on the OCW and sold it to someone who may want to give it a shot or use it as is.
If I ever get an opportunity to open up my caldigit i will be on the lookout for these diodes and test them to see if they are out of spec. I maybe able to revive that dead port.