Laptop Thunderbolt Port Charges but Has No Data or Video, Dock Not Recognised, and Intermittent Disconnects: An Advanced Diagnosis Guide — After the Basic Checks, Separating Cable, Authorisation, Controller, and Firmware Layers

One USB-C plug, one laptop, one dock that worked yesterday, and today the port only charges while data and video give nothing at all. Or a dock that is not recognised, works briefly after a restart, and drops again after sleep. The first instinct is usually that the port is dead, and the second is to buy another dock to test it. This guide argues that both instincts can be wrong, and wrong expensively: charging proves almost nothing about the port, and a replacement dock proves nothing about the port either.

Start with the scope, because the scope is the boundary of the whole guide. This is the port and controller layer, at advanced depth: the Thunderbolt controller, controller firmware and NVM, Power Delivery negotiation, PCIe and USB tunnelling, security levels, cable types, BIOS and UEFI settings, and the operating-system driver stack. It is not the dock-level guide: dock firmware, dock power, and each port on the dock belong to the dock article (laptop-usb-c-dock-not-recognised-advanced-hk), and this guide uses a dock only as a test device that proves or excludes a fault on the port side. It is not a screen-level guide either: the laptop's own panel brightness, dead pixels and lines, backlight, and internal display cable belong to the screen articles, and this guide mentions an external monitor only to test whether the DisplayPort tunnel works.

The advanced procedure assumes the basic checks are already done: a restart, another cable tried at a basic level, and the dock's power checked, plus the monitor input source checked. If those steps have not been done, do them first, or the advanced results will be polluted by the simplest cause. From here on, change one thing at a time, record what you saw, and only then change the next thing. Changing the cable, the dock, the driver, and the BIOS all at once destroys the evidence.

One safety note matters before anything else: this guide covers diagnosis and observation, not disassembly repair. Any work that opens the machine, inspects an internal port, or touches a controller must be done by a qualified technician with the machine powered off and the battery isolated. Never hot-plug an internal connector while powered, and never treat a firmware flash as a casual routine step. An interrupted firmware update can leave a controller in a state from which it cannot start normally, and that risk is repeated wherever firmware appears below.

1. One USB-C port is doing four jobs

USB-C is only a connector shape, and the shape proves almost nothing by itself. A plain USB-C port may carry only USB 2.0, or USB 3.x, or DisplayPort Alt Mode, or Power Delivery, or, only if the laptop, the controller, and the cable all support it, Thunderbolt or USB4. Treating shape as capability is the single largest source of reports that the port is broken when the real problem is a capability mismatch.

From the bottom up, one Thunderbolt port contains these layers:

1. The physical port and contacts: a 24-pin reversible connector, with retimer or redriver circuitry between the port and the controller on many designs. Bent, loose, dirty, or corroded contacts can break the high-speed lanes while the separate power pins keep working.
2. USB Power Delivery negotiation: a separate negotiation from data. This is why charging can succeed while data and video both fail. Apple states explicitly that on Apple silicon Mac laptops an accessory can still charge the Mac even if the user chooses not to allow data access.
3. The USB4 or Thunderbolt link: the high-speed fabric that carries tunnels. USB4 is based on the Thunderbolt protocol contributed by Intel, and it moves multiple protocols by tunnelling them over one link, with bandwidth shared dynamically rather than fixed per protocol.
4. The tunnels carried on that link: PCIe tunnelling for Thunderbolt storage, docks with PCIe devices, and external GPU enclosures; USB tunnelling for USB devices behind a USB4 hub or dock; DisplayPort tunnelling for video. USB 2.0 runs on separate wires in parallel and is not tunnelled in the same way, which is why a USB 2.0 mouse or keyboard can work on a port whose high-speed lanes are dead. That last point is presented here as a check to try, not as a proven diagnostic rule in that exact form.
5. The Thunderbolt or USB4 controller, the host router: Intel discrete controllers on Windows laptops include Titan Ridge for Thunderbolt 3, Maple Ridge for Thunderbolt 4, and Barlow Ridge for the newer Thunderbolt 4 and USB4 generation. The exact chip must be verified for the exact laptop model and never guessed from the port shape. On Apple silicon there is no separate Intel controller card to reseat or swap in the Intel sense; the controller is integrated into the system-on-chip, so a controller fault points toward board-level work or board replacement, subject to inspection.
6. Controller firmware and NVM: the controller's own non-volatile firmware, distinct from the laptop BIOS and distinct from the Windows or macOS driver. Vendor documentation refers to the NVM Firmware version shown in the Thunderbolt software details page.
7. The operating-system driver and connection-manager stack: the Windows 10 era used an Intel Thunderbolt driver plus the Thunderbolt Control Center app for approval and device listing, while Windows 11 USB4 systems increasingly use Microsoft's USB4 connection manager and driver stack. macOS uses its built-in Thunderbolt and USB4 stack plus the accessory-approval gate described later.
8. BIOS and UEFI settings: Thunderbolt enable and disable, security level, pre-boot and boot support, and on some models a separate Thunderbolt firmware update toggle. Exact menu names vary by vendor and model, so this guide says to check the vendor's manual for the exact name rather than inventing a universal menu path.
9. The security and authorisation layer: a device can be physically linked and yet never enumerated because authorisation was refused, missed, or set to a mode that forbids PCIe. That layer is covered in detail in Section 6, before any hardware conclusion.

Keep this stack in mind. Every table below follows it, and a symptom alone is never a diagnosis; a symptom only tells you which layer to check first.

2. Symptom to layer map: find your row before touching anything

This table is the navigation for the whole guide. Each row names the most likely layer first, then what else must be excluded before a hardware conclusion. No row is a certainty from the symptom alone.

Symptom	Most likely layer first	What must be excluded before a hardware conclusion
Charges but no data and no video	Power Delivery negotiation succeeded, but the high-speed link, a tunnel, authorisation, or the driver did not	Wrong cable, meaning a charge-only or USB 2.0-only cable, or an active Thunderbolt cable used with a USB-only device; approval refused or missed, on Windows or macOS, where charging still works in both cases; BIOS Thunderbolt disabled or set to a no-PCIe mode while Power Delivery still works; driver or controller not loaded; a physical high-speed lane fault in the port or cable while power pins survive
Dock or monitor not recognised at all, with no entry in Device Manager, System Report, Thunderbolt software, or display settings	Cable capability, port capability, authorisation, dock power or firmware, operating-system driver, BIOS setting, controller, in that order	Not recognised is not the same as recognised but no video: the first points to link, enumeration, or authorisation, and the second points to DisplayPort tunnelling, display settings, monitor input, or bandwidth contention when two high-resolution displays plus high-speed storage share one 40 Gbps link
Intermittent disconnects, especially under load, on wake, or when moved	Cable and connector seating, power management, bandwidth contention, dock power, or a marginal link that negotiates down and then drops	A gentle one-time wiggle test at the laptop end versus the dock end, used only to localise, never as a habit or a fix; Windows power management, including the allow the computer to turn off this device to save power option and USB selective suspend; sleep, wake, and Fast Startup state; daisy-chain and power budget; a marginal passive cable at a 20 Gbps fallback that is stable for light use and drops under sustained high-bandwidth use, presented as a test rather than a proven mechanism
Device works on one port but not another, same laptop, same cable, same device	The failing port's physical contacts, its retimer path, its BIOS enablement, or its per-port power or authorisation state	The multi-controller caveat: two ports on one laptop may sit on different controllers or different power and bandwidth budgets, so check the model specification; otherwise this is the strongest port-localising symptom in the guide and the clearest trigger for booking an inspection of that port rather than buying another dock
Works only after a reboot, or only after a full shutdown rather than a restart or wake	Driver, firmware, or connection-manager state, or a device that only enumerates at boot	Fast Startup and hybrid sleep state; a pending driver or firmware update that only takes effect on a cold boot; a dock that must be powered before the laptop is powered, presented as a sequence to try rather than a specification; an authorisation refused in the previous session and only re-prompted after a restart
One function works and another does not, for example video works but storage does not, or USB works but Thunderbolt storage does not	The tunnel layer, and this is the tunnel-level tell the guide teaches explicitly	Video working with PCIe storage missing means the DisplayPort tunnel or Alt Mode is up while the PCIe tunnel or authorisation is not, and the reader should check the security level, approval, and whether the host actually supports PCIe tunnelling, which is optional for a USB4 host; USB working with Thunderbolt missing means the USB path is up while the Thunderbolt link or PCIe tunnel is not, or the cable is a USB-only cable; a monitor failing on a dock while the dock otherwise works points to the DisplayPort tunnel, monitor input or settings, or bandwidth contention, not to the whole port

One recording habit makes every row above more useful: write the four functions separately. Charges, USB device works, monitor or video works, Thunderbolt device works, each as yes or no. A four-way table like that is worth far more than a single sentence saying the port is broken, and it is exactly what a technician wants to see at inspection.

3. Thunderbolt 3, Thunderbolt 4, USB4, and plain USB 3.x: guarantees versus possibilities

Do not headline this section as Thunderbolt 4 being faster than Thunderbolt 3. Both top out at 40 Gbps. The difference is in minimum guarantees: Thunderbolt 4 is a certification with minimums, while USB4 is a specification with options. A USB4 port is not automatically a Thunderbolt 4 port, and a Thunderbolt 3 port is not automatically as capable as a Thunderbolt 4 port.

Label on the box or port	What is guaranteed	What is NOT guaranteed
Thunderbolt 3	Up to 40 Gbps link; PCIe minimum only 16 Gbps; DisplayPort 1.2 or 1.4 era video, model-dependent	Dual 4K on every model; 40 Gbps on a long passive cable; PCIe performance equal to Thunderbolt 4
Thunderbolt 4	40 Gbps; PCIe minimum 32 Gbps; two 4K displays or one 8K display; charging on at least one computer port; docks with up to four Thunderbolt ports; wake from sleep via dock; VT-d-based DMA protection	That a given dock, cable, or monitor uses the full capability; that a USB4-only host is a Thunderbolt 4 host
USB4, 20 Gbps	20 Gbps USB4 link; USB 3.2, USB 2.0, DisplayPort tunnelling, DisplayPort Alt Mode on downward-facing ports, USB Power Delivery	40 Gbps; PCIe tunnelling, which is optional for a USB4 host
USB4, 40 Gbps	40 Gbps USB4 link where the host, hub, and cable all support it	That every USB4 host supports PCIe tunnelling at Thunderbolt 4's 32 Gbps minimum; hub requirements are stricter than host requirements
USB 3.x only, USB-C shape	USB data at the stated USB 3.x speed; often DisplayPort Alt Mode and Power Delivery, model-dependent	Any Thunderbolt or USB4 tunnelling at all

The summary sentence is worth repeating: Thunderbolt 4 is a certification with minimums, USB4 is a specification with options. When checking a model specification, read which name is actually written there, and do not infer guaranteed behaviour from the connector shape or the lightning-bolt mark.

4. Cable science: why a cable can pass power but fail data and video

A USB-C cable is not a single pipe. Power Delivery negotiation, USB 2.0 data, and the high-speed lanes travel on different pin and wire groups, and a cable may be built with only some of them:

• A charge-only or USB 2.0-only cable can negotiate Power Delivery and charge the laptop while having no high-speed lanes at all, so it can never carry SuperSpeed USB, DisplayPort Alt Mode video, or Thunderbolt tunnelling.
• DisplayPort Alt Mode repurposes some or all of the four high-speed lanes for video. In two-lane mode, video shares the lanes with USB data at reduced speed; in four-lane mode, the lanes are dedicated to video and only USB 2.0 data remains. This is why a monitor can work while USB on the monitor is slow, and why USB can work while the monitor is blank, with both outcomes being cable or mode outcomes rather than port faults.
• Thunderbolt and USB4 tunnelling packetises USB, DisplayPort, and PCIe into one high-speed link and shares the bandwidth dynamically, rather than fixing lanes to one protocol. A marginal link therefore degrades or drops tunnelled functions together, while Power Delivery, as a separate negotiation, keeps charging.
• The conclusion is simple: a damaged cable can pass power and fail data and video, and a damaged port can do the same. Only the swap test, same device with a different cable and then a different port, separates the two. Neither it charges nor the cable looks fine settles it.

Cable comparison: active versus passive, length versus speed

Cable type	What the sources say	Diagnostic meaning
Thunderbolt 3 passive, short cable	40 Gbps	A short passive Thunderbolt 3 cable at full speed is a valid control test
Thunderbolt 3 passive, 1 m or 2 m	Drops to 20 Gbps unless active	A long passive Thunderbolt 3 cable is not a valid known-good test for full-speed Thunderbolt
Thunderbolt 3 active, 1 m or 2 m	40 Gbps, using signal-conditioning electronics in the cable	Use an active cable when the setup needs length plus full speed
Thunderbolt 4 universal cable, up to 2 m	40 Gbps without needing active boosters in Intel and vendor wording; Thunderbolt 4 certification requires 40 Gbps over the target length up to 2 m, plus Power Delivery cross-checks and E-Marker validation	If a Thunderbolt 4 setup fails at 2 m with a certified Thunderbolt 4 cable, do not blame length first; check the port, controller, and security branches
USB4 passive, 1 m	40 Gbps in the TechAdvisor length table	A short passive USB4 cable can be a valid control at that length
USB4 passive, 2 m	20 Gbps in the TechAdvisor length table	A 2 m passive USB4 cable is not equivalent to a 2 m Thunderbolt 4 cable; do not substitute one for the other in a test and then call the result a port fault
Thunderbolt 4 active cable, 3 m	40 Gbps in one vendor comparison	This figure comes from one vendor comparison, not from a universal guarantee; check the rating printed on the cable itself

On the passive full-speed length threshold, this guide deliberately states both sourced figures: some sources say 0.8 m and others say 0.5 m, or 18 inches, for full speed. Presenting one figure as an absolute threshold would be false precision. The rule for the reader is that shorter gives more confidence for a full-speed test, and longer raises the suspicion that the link has already fallen back. For a control test, use one short cable, 0.8 m or shorter, certified, with the Thunderbolt logo and the correct number, 3 or 4, plus one single device, direct to the laptop with no dock, hub, or adapter in between. If that combination works, the previous cable or chain is implicated, not the port.

How to read a cable: look for the Thunderbolt lightning-bolt mark on the connector and any printed rating on the cable or its packaging, meaning Thunderbolt 3 versus Thunderbolt 4 versus USB4 versus USB 3.2 versus charging only. No mark and no rating means unproven, not Thunderbolt. Certified Thunderbolt and USB4 cables carry an electronic marker chip that advertises their rating to the host; that concept is enough for the reader, and this guide does not ask readers to buy an E-Marker reader. Keep the original cable that came with the dock or monitor as the control test. One active-cable trap also matters: some active Thunderbolt 3 cables fall back to USB 2.0 speeds with USB-only devices, so a cable that is excellent for a Thunderbolt dock can be a poor cable for a USB-only drive. Test with the actual device class, not just any device.

DisplayPort Alt Mode video path versus Thunderbolt video path

• DisplayPort Alt Mode path: the laptop USB-C port repurposes high-speed lanes for native DisplayPort to the monitor. It requires a port that supports DisplayPort Alt Mode, a cable that carries the lanes, and a monitor input set to USB-C. If any link is USB-only, there is no video.
• Thunderbolt video path: the laptop Thunderbolt controller establishes a Thunderbolt link and tunnels DisplayPort inside it to the dock or monitor. It requires a Thunderbolt link established, security and approval passed, and PCIe and DisplayPort tunnelling not blocked. So direct USB-C to monitor working with through-dock monitor failing points at the dock or link, while a direct Thunderbolt monitor failing on a port whose controller is present points at the port, controller, or security branches.
• The one monitor-side check that belongs in this guide is the input source: if the monitor has both a USB-C input and a Thunderbolt input, those labels are not interchangeable, and the reader should use the input the monitor manual specifies for that cable. Monitor settings beyond that belong to the screen articles.

5. Layered exclusion procedure: in this order, stopping at the first step that changes the result

The order matters: cheapest, least invasive, and most common first, with firmware and hardware last. Record what was seen at each step, meaning device name, link information if shown, and error text, before moving on. A reader who changes the cable, dock, driver, and BIOS all at once has destroyed the evidence.

Step	What to do	How to read the result
Step 0: Confirm the port is supposed to be Thunderbolt at all	Check the lightning-bolt mark beside the port if the vendor marks it, then check the exact model and product number specification on the vendor site under Ports or Connectivity	If the specification says only USB-C, stop and reframe: the expected behaviour is USB-C behaviour, not Thunderbolt behaviour. Record the model, the port position, and whether another port on the same laptop behaves differently
Step 1: Cable, done properly	One short certified Thunderbolt cable, 0.8 m or shorter, plus one single device, direct to the laptop with no dock, hub, or adapter in between; check the cable's printed or packaging rating	If that combination works, the previous cable or chain is implicated, not the port. A cable that works for charging can still be a 20 Gbps cable that starves a dual-display or high-speed storage setup
Step 2: Port, physical and per-port	Visual check with power off and cable out: bent tongue, debris, visibly loose plug seating, a plug that rocks side to side; do not poke inside the port with metal tools; swap ports on the same laptop with the same cable and device; try both orientations once	If the fault follows the port across cables and devices, the port is suspect. Orientation behaviour is included only as a one-time try, not as a diagnosis of a specific contact. A wiggle test is a localisation test only; continued use of a loose port risks further contact damage, stated as workshop prudence rather than a specification
Step 3: Dock or device, only enough to exclude it	Power the dock from its own power supply where it has one; connect the device direct to the laptop, bypassing the dock; if a second Thunderbolt-capable laptop is available, test the same dock or device there with the same cable	Direct works and docked fails implicates the dock, its cable, its firmware, or its power, which is the dock article's territory, and this guide hands off there rather than re-diagnosing the dock. The second-laptop test is the cleanest dock-versus-laptop test available
Step 4: Operating system, driver, and approval state	Windows: Device Manager and Thunderbolt software or USB settings. macOS: System Report and the Accessories setting. Watch for whether an approval prompt appears, was refused, or never appeared	A controller present and working only shows that Windows sees the host hardware; it does not prove the link to a device works. Approval refused means PCIe devices will not appear while video may be normal. Full paths are in Sections 7 and 8
Step 5: Authorisation and security level, BIOS plus operating system together	Note the current BIOS security setting before changing anything, then try a documented reversible test setting, re-test, and restore the recommended setting; on a managed work laptop, ask IT before changing policy, and do not bypass employer device policy	A DisplayPort-only or no-PCIe BIOS mode means PCIe devices will never enumerate. That is a setting, not a fault, and the fix is a setting change with the security trade-off explained, not hidden. A prompt that never appears may mean the manager software is not running, the prompt is hidden behind a locked screen or a minimised window, or a policy blocks the device
Step 6: BIOS and UEFI, settings before updates	Confirm Thunderbolt is enabled, the security level is as intended, and any Thunderbolt boot support or pre-boot setting matches the use case, since a device needed at boot behaves differently from a device needed only after login	Do not update the BIOS as a first step. BIOS updates are a later step, done only from the vendor update for the exact model, on mains power, with no Thunderbolt device update running at the same time. A BIOS update must not be presented as a routine fix for a port fault
Step 7: Controller firmware and NVM, separate from BIOS and driver	Find the current NVM firmware version first, in Thunderbolt software under Settings and Details as NVM Firmware version; on Windows 11 USB4 systems where that app has no function, the vendor's own update tool or support page for the exact model is the source of truth	Update only from the laptop vendor, or where the vendor directs it, the dock vendor for dock firmware, which belongs in the dock article. Vendor guidance requires enabling a BIOS Thunderbolt firmware update setting first, running the vendor tool, not touching keys or shutting down before it finishes, then shutting down and restarting to activate the new firmware. After any firmware or driver change, do a full shutdown, then a cold boot, then reconnect in a known-good order, direct device first and dock second, and re-check Device Manager or System Report before declaring success or failure
Step 8: Controller and hardware, only when every step above has been done and recorded	Remaining patterns that point to hardware: the same device plus the same certified short cable fails on one port but works on another port of the same laptop; or no controller appears in Device Manager or System Report on any port, with the driver correctly installed and Thunderbolt enabled in BIOS, and the port shows physical damage or looseness	Hardware can mean several different things, and they must be kept distinct: the port connector and solder joints; the retimer or redriver in the port path where the design has one; the Thunderbolt controller chip itself, whether an Intel discrete chip or an integrated controller on Apple silicon or a modern Intel system-on-chip; or board traces and power delivery circuitry feeding the port. Symptoms alone cannot distinguish these, and on many modern laptops the controller and port are not separately replaceable parts

If Step 8 still leaves the fault unexplained, the correct next step is not another dock, not another firmware flash, and not opening the laptop. It is inspection, with the checklist in Section 12.

6. Security levels SL0 to SL3: the section most worth reading

Many cases that look like a dead port are in truth an authorisation layer doing its job. The security levels exist because a connected Thunderbolt device can act as a direct-memory-access master and read host memory without the processor or operating system knowing, unless input-output memory management based DMA protection is in place. A lower setting is therefore a diagnostic test, not a permanent fix to leave in place.

Level	Meaning	Effect on symptoms
SL0, None, Legacy mode	Every attached device is authorised without user interaction	Highest convenience, no PCIe authorisation gate; use it to isolate a fault, then restore the vendor-recommended setting
SL1, User authorisation, the default from Thunderbolt 2 onward in the Intel summary	The user is asked when a new PCIe device connects, typically with connect one time, always connect, or do not connect; the host keeps an access-control list of device unique identifiers granted always connect; permissions are per machine, not per user login	A missed, dismissed, or never-shown approval prompt looks exactly like a dead port for PCIe devices, while video may still work
SL2, Secure connection	As SL1, plus cryptographic challenge and response: when a device is first granted always connect, a key is written to the device controller's non-volatile memory and recorded on the host, and later connections are verified against it; this prevents a cloned device identifier from inheriting always-connect status	Not all peripherals support SL2; a failed verification produces another permission prompt rather than silent success, and must not be misread as a dead device
SL3, No PCIe tunnelling, DisplayPort-only and related modes	The port is set to DisplayPort-only, or in kernel documentation variant naming, display-only, USB-only, or no-PCIe modes; DisplayPort still works because it has no direct-memory-access exposure in the same way, while PCIe devices do not enumerate at all	A Thunderbolt storage device or PCIe dock function will never appear no matter how many cables are swapped, while a monitor may be fine; this is a setting, not a fault

BIOS names vary by vendor: Legacy or None, Unique ID or User, One-time saved key or Secure, DisplayPort-only or USB-only, or similar vendor wording, and kernel documentation maps those BIOS names to the SL levels. Dell's primary path is System Configuration, USB and Thunderbolt Configuration; other vendors' paths must be checked in that model's manual, and this guide does not publish any one vendor's path as universal.

Two related settings must not be conflated. Thunderbolt Boot Support, when enabled, helps the system detect Thunderbolt devices during startup and pre-boot, so a device that works after login may not work at the login screen or in pre-boot if boot support is off. Separately, Kernel DMA Protection intentionally prevents Thunderbolt docks from initialising before the operating system boots, so peripherals on a dock may not work at the login screen but work after login, which Dell describes as working as designed under that policy, with a Group Policy or DmaGuard change as the documented route. Both produce not working until after login, but they are different settings with different fixes.

Thunderbolt 4 certification requires Intel VT-d-based DMA protection. On Windows, if Kernel DMA Protection is off, the reader should be pointed to the firmware VT-d or input-output memory management setting as a check, without any claim that turning it on fixes every case. That fix claim is unsourced and is not made here.

7. Windows check path: controller present, warning, or entirely absent

Windows checks split into three states, and each state leads to a different next step.

First, controller present with status working. In Device Manager, under System devices, there is a Thunderbolt Controller, or on newer USB4 systems, under Universal Serial Bus controllers, a USB4 Host Router. Status working means Windows sees the hardware; it does not prove the link to a device works. In this state the next checks are authorisation, cable, and BIOS security level, not another driver reinstall.

Second, controller present with a warning or error code. This points to the driver layer, and the driver must come from the laptop vendor's page for the exact model, not from a generic driver site. This guide does not print a catalogue of Device Manager error codes, because the only concrete example found in this research is error code 28, meaning no driver found, in a single Microsoft question-and-answer report of a dead Thunderbolt port. That is a single user report, not a code table. The correct action is to record the exact code text shown on the reader's machine and search it against the vendor's support site.

Third, controller entirely absent. Split this further: if a Base System Device appears under Other devices, that pattern indicates the Thunderbolt driver is not installed, and installing the correct vendor driver should replace the Base System Device with the controller entry. If the driver is correct and Thunderbolt is enabled in BIOS and the controller is still absent, the suspicion moves to the firmware or controller-hardware layer. One Windows 11 trap must be stated plainly: absence from the old Thunderbolt Control Center does not prove absence of a controller, because the app may simply have no function on that stack. Check Settings, Bluetooth and devices, USB, which is the Windows 11 USB4 information location in secondary guides, and Device Manager for USB4 Host Router before concluding anything. A community thread quotes an Intel reply saying there are no Thunderbolt 4 drivers for Windows 11 planned and the Microsoft USB4 drivers are sufficient; this guide presents that only as a reported vendor position from a community thread, not as a primary Intel document, and the correct action remains the vendor's exact-model driver and firmware page.

Power management is a further Windows branch for intermittent disconnects. In Device Manager, on USB Root Hub or Thunderbolt entries, the Power Management tab option allow the computer to turn off this device to save power, together with USB selective suspend in the power plan, is documented in secondary Windows guides as a cause of seemingly random disconnects and black screens. Unchecking it is a reversible test for intermittent disconnects, and the original setting should be noted.

A final Windows branch is a Group Policy or custom image block. Dell documents that a Group Policy setting in a custom Windows image can prevent USB-C and Thunderbolt devices and drivers from being installed at all, with a yellow exclamation mark in Device Manager, as a leftover mitigation for a Thunderbolt 1 and 2 direct-memory-access vulnerability that does not apply to Thunderbolt 3 devices. This matters for ex-company or second-hand laptops, and the article's question is whether the laptop was ever on a company image. If so, the fix is policy, not hardware, and the reader should ask IT rather than bypassing employer policy.

8. macOS check path: System Information first, Finder second

The macOS order differs from Windows, but the logic is the same: check whether the hardware layer sees the device before discussing mounting or Finder.

First, System Information. From the Apple menu, About This Mac, More Info, System Report, under Hardware there is a Thunderbolt or USB4 entry. A connected device listed there means macOS recognises it at the bus level; an entry reading no device connected with a device plugged in means the bus never saw it, which points to cable, port, or device rather than a Finder or mount problem. Do not conclude that a drive is dead merely because Finder does not show it.

Second, the accessory-approval gate. Apple Support documents that on a Mac laptop with Apple silicon, a new or unknown USB accessory, Thunderbolt accessory, or memory card triggers an alert asking the user to allow it. If the user clicks Do not Allow, the Mac will not recognise the accessory or give it data access, and will ask again next time. Charging still works. The settings path published by Apple is Apple menu, System Settings, Privacy and Security, Accessories, Allow accessories to connect, with options Always ask, Ask for new accessories as the default, Automatically allow when unlocked, and Always allow. The Mac must be unlocked before approval is possible, so the test is to unlock, disconnect, reconnect, and watch for the prompt.

Third, the lid-closed trap. Apple's own note states that a lid-closed or docked Mac needs a display, mouse, and keyboard allowed first before other accessories can be allowed without opening the lid. A reader who tests only with the lid closed can misread a missing prompt as a port fault. There is also a time trap: if the Mac has been locked for three or more days, a previously allowed accessory may need approval again.

Fourth, the boundary on secondary macOS claims. Secondary material describes a message of the form cannot use the Thunderbolt accessory, please connect the accessory to a Thunderbolt port, as triggered by using a non-Thunderbolt cable in a Thunderbolt port. This guide presents that only as a secondary report, and the action remains the Step 1 cable check, not the message text itself. No Terminal command for resetting the Thunderbolt bus, and no Apple silicon reset key combination, is published here, because this research did not retrieve an Apple primary page for those steps. The safe sourced macOS steps are: unplug the chain and test the device directly in the port with no daisy-chain and no adapter in between; test another known-good Thunderbolt device on the same port; check System Information before touching Finder or Disk Utility; update macOS; and check the Accessories allow setting.

The Apple silicon structural point bears repeating: the M1 was Apple's first custom USB4 and Thunderbolt controller implementation, and later Apple silicon integrates the USB4 and Thunderbolt host controller into the system-on-chip rather than using a separate Intel controller chip. On Apple silicon there is therefore no separate Thunderbolt card to reseat or swap in the Intel sense, and a controller fault points toward board-level work or board replacement, subject to inspection.

9. Controller firmware and NVM: why firmware is not a driver, and why a failed update means stop

Three things are routinely confused: the BIOS, the driver, and controller firmware. The BIOS is the board firmware that governs boot and hardware settings. The driver is operating-system software that talks to the hardware. Controller firmware, or NVM, is the firmware inside the Thunderbolt controller chip itself, shown in Thunderbolt software under Settings and Details as the NVM Firmware version. The three update differently, carry different risks, and must never be done all at once or treated as substitutes for each other.

The vendor firmware procedure, from the MSI guide, is: enable the BIOS Thunderbolt firmware update setting first, run the vendor tool, do not touch any key or shut the system down before it finishes, then shut down Windows and restart to activate the new firmware. The vendor troubleshooting messages should be read literally: no Thunderbolt device in the tool points first to the BIOS toggle; service does not exist points to an old driver, meaning remove the old driver and install the vendor's latest; controller does not support device firmware update points to the wrong item selected in the tool; and a repeated update reminder after entering Windows points to disabling the Thunderbolt firmware update setting in BIOS after the update finishes. These are vendor-documented tool messages, not a general diagnosis.

Safe mode is a specific state that must be understood before any update is attempted. Kernel documentation states that if the existing NVM is not properly authenticated or is missing, the host controller goes into safe mode, in which the only available functionality is flashing a new NVM image; reading the NVM version fails in that state and device identification information is missing. Lenovo's primary-source symptom list for the named ThinkPad models on its critical Thunderbolt update page includes: USB-C port not working; Intel Thunderbolt controller not visible in the operating system or Device Manager; USB-C or Thunderbolt docking stations not visible or having connectivity problems; HDMI output not available; system battery not charging with a USB-C power adapter; Intel Thunderbolt pop-up error message; Intel Thunderbolt safe-mode error message; and BIOS Thunderbolt communication error or hang during POST. That list is Lenovo's wording for those models, not a general rule for all laptops, and the page's note that symptoms may occur after six to twelve months of typical usage is likewise Lenovo's wording for those models and must not be generalised. A single forum report of a controller disappearing after a Windows update and returning after a BIOS non-volatile memory clear is also not published here as a procedure.

The pre-update checklist that is safe to publish is: laptop on mains power, not battery alone; no device daisy-chained through the port being updated unless the vendor instructs it; exact-model package only; do not sleep or shut down during the update; restart as the vendor instructs after it finishes. If a firmware update fails, stop, note the exact error and the versions before and after, keep the laptop on power, and get it inspected. Do not run another firmware package on top of a failed one. If no newer compatible NVM firmware is available from the system vendor, the documented position from Plugable is that the only option is to contact the vendor and ask when one will be available. Dock firmware belongs in the dock article; this guide records dock firmware only as a test-device fact, and host firmware and dock firmware must not be updated together as one casual step.

10. Intermittency by trigger: each trigger gets one discriminating test

Intermittent faults are the hardest to diagnose because the fault may be absent at inspection. Recording the trigger pattern matters more than describing the fault as frequent, and no frequency ranking is claimed anywhere in this guide, because no prevalence data was sourced.

Trigger pattern	Discriminating test	How to read the result
Movement: drops when the laptop moves, the cable is touched, or the lid angle changes; works when held still	Hold the laptop and cable completely still for a timed period, then move only the cable at the port end, then only at the device end, then only the laptop, recording which movement reproduces the drop	Movement at the laptop end only points toward the laptop port side; user-report sources describe slight movement interrupting a high-speed signal, but that is forum and user-report material, not a measurement, so it justifies the movement test and nothing stronger; never keep bending or propping a cable as a fix
Sleep and wake, or lid close and open: works after a full restart, fails after sleep or wake	Three comparisons: full shutdown then power on, sleep then wake, lid close then open, recording which comparison fails	Full shutdown always working with sleep and wake failing points to the resume path, meaning driver power management, BIOS sleep mode, or firmware state, not to a physical port; all three failing identically means the resume path is not the discriminator
Heavy load: drops during large file transfers or dual-monitor plus charging at once	Single device direct in the laptop port with no dock power chain, then re-add dock devices one at a time	Direct working with loaded-dock failing points to the dock, power, or load side rather than proven laptop-port fault; direct failing under no extra load keeps the port suspect; no wattage figure for any specific dock is quoted, because none was sourced for this batch
Idle periods: drops after idle, stable during active use	Record how long the idle period was before each drop, and whether drops also occur during active use	Idle-only drops raise the power-management suspicion, while active-use drops raise the link or physical suspicion; that correlation is the diagnostic value, and any setting change belongs to the technician once the model is known
Adding devices: the last device in a chain drops first, or devices drop when another device is added	Break the chain and test each device direct in the laptop port, one at a time, then rebuild the chain one device at a time	A device that fails direct is the suspect; a device that only fails at the end of a chain points at the chain, the intermediate device, or total bandwidth, not automatically at the laptop port; a Thunderbolt 4 hub gives each downstream port its own branch that can disconnect without affecting the others, while a pure daisy-chain does not have that isolation
After an update: worked before a driver, operating-system, BIOS, or firmware update, and fails after it	Record what worked before, which item was updated, and the versions before and after; separate worked before update from never worked	Worked before update points to a software regression to investigate by vendor-guided driver or firmware handling, not by casual firmware rollback; never worked points to a capability or hardware question, and the two directions must not be merged

On temperature, the boundary is explicit: this research found no primary source stating a Thunderbolt controller thermal limit, a temperature threshold, or a thermal-throttle behaviour for laptop Thunderbolt ports. No temperature figure, no overheats-after-a-number-of-minutes claim, and no cooling fix is published as fact. What can be published is the test pattern: note whether drops correlate with heavy load plus heat versus occurring cold at idle, and report that pattern at inspection as a hypothesis to test, not as a diagnosis.

On PCIe link power management, the boundary is the same: this research did not retrieve a primary source for a specific PCIe power-management setting that fixes Thunderbolt disconnects on a named laptop model. No disable-PCIe-power-management step is published as a verified fix. The article may mention, as a question for inspection, whether disconnects correlate with idle periods versus active use. That correlation is the diagnostic value; the setting change belongs to the technician once the model is known.

11. Common misdiagnosis traps, dismantled one by one

1. Any USB-C cable is a Thunderbolt cable. False. Shape is not capability. A charge-only, USB 2.0-only, or USB 3.x-only cable will charge and may run low-speed devices while never establishing a Thunderbolt link. Check the logo, the number, and the rated speed.
2. The dock is faulty when the cable is passive and too long. A 2 m passive cable at 20 Gbps can make a dual-display or high-speed dock setup fail or drop, while the same dock works on a short cable. Swap the cable before condemning the dock.
3. Confusing USB4, Thunderbolt 4, Thunderbolt 3, and USB 3.x fallback modes. A link that negotiates down to USB4 20 Gbps, USB 3.x 10 Gbps, or USB 2.0 is not broken in the sense of a dead port; it is a fallback. The observable question is what speed or mode it actually negotiated, and which element, cable, host, hub, or device, capped it.
4. Windows approval prompts: missed, dismissed, or never-shown security-level prompts; a device previously refused; a corporate policy blocking PCIe devices. The symptom looks like a dead port for PCIe devices only. Check the approval list and Control Center state before any hardware conclusion.
5. macOS accessory approval: Do not Allow clicked once, or a Mac locked for days re-prompting, or a lid-closed test that never shows the prompt. The symptom looks like a dead port, and charging still works. Check Privacy and Security, Accessories.
6. Daisy-chain and power-budget limits: Thunderbolt supports daisy-chaining up to six devices from one port, and the chain is only as good as its weakest link and its power budget. Bus-powered devices drawing from the port, plus displays, plus storage, can exceed what the port or the dock power supply provides. Symptoms are that the last device in the chain drops first, or devices drop when another device is added. Test with a single device direct, then rebuild the chain one device at a time.
7. It worked before the update, so the port died. Driver, operating-system, BIOS, and controller-firmware updates can each change enumeration, approval, or power behaviour. Separate worked before update, a software regression to investigate by vendor-guided rollback handling, from never worked, a capability or hardware question. Do not instruct a reader to roll back firmware casually; vendor guidance comes first.
8. One port works, so the laptop is fine, or no port works, so the whole board is dead. Both over-read the evidence. One working port localises the fault but does not exclude a controller fault affecting only one port path; no working port does not by itself prove board death, because BIOS disablement, driver absence, and a failed firmware state must be excluded first.
9. Assuming the monitor is the fault when the port is the fault, and vice versa. If the same monitor works via HDMI or via another laptop's Thunderbolt port, the monitor is substantially excluded; if the same laptop drives a monitor via HDMI but not via Thunderbolt, the Thunderbolt tunnel, cable, or controller path is implicated, not the panel. Keep this boundary clean against the screen articles: panel, backlight, and internal display cable are out of scope here.
10. Cleaning and reseating traps: do not recommend compressed air blasted into the port at close range, metal picks, or tightening a USB-C port by bending the tongue. Visual inspection only; any internal cleaning or rework is an inspection job.

12. Repair, replace, or upgrade decision framework

This framework is a decision table, not a recommendation for any individual reader. Every cost cell reads subject to a written quote after inspection, except where a confirmed starting figure genuinely applies to that row, and for Thunderbolt rows it does not, so the quote wording is used throughout this section.

Situation after the decision tree	What it points to	Next step to present	What not to promise
Fault follows the cable across ports and devices	Cable	Replace the cable with a certified cable of the correct type and length for the setup, Thunderbolt 4 universal to 2 m, or active if longer or full-speed length is needed	Do not name a brand or price; do not buy a longer cable to fix a length problem; buy a correctly rated cable at the shortest length that reaches
Fault follows the device across ports, cables, and laptops	Device, dock, or monitor	Device-side support or the dock article	Do not diagnose the device in this article
Fault follows the port across cables and devices, controller present, security and firmware ruled out	Port assembly, Power Delivery controller, Thunderbolt controller, or board interconnect	Inspection and board-level diagnosis	No price, no turnaround promise, and no claim from symptoms alone about which component it is
Controller absent or in safe mode after correct driver plus BIOS enabled	Firmware or controller hardware	Inspection; do not re-flash repeatedly	No rollback promise, and no claim that controller work is always possible or always impossible
One Thunderbolt port works and the other does not, same cable and device	That port's assembly or its controller branch	Inspection; the interim workaround is to use the working port if the laptop has one, stated as a workaround, not a repair	Do not claim the working port proves the laptop is fine; do not recommend living with the working port as a permanent fix where the faulty port is also the only charging port, or where the faulty port is physically loose and may worsen
Laptop has only one Thunderbolt port and it has failed at board level	Board-level repair versus board replacement versus replacement laptop versus workaround using surviving USB-C non-Thunderbolt functions or a different connection for the monitor	Written quote first, then the customer decides with the quote in hand, weighing the laptop's age, value, data, and whether the controller is discrete or integrated	Do not steer the customer to replacement without a quote; do not quote a replacement-laptop price
USB-only fallback is acceptable, meaning the reader needs only USB data, a USB monitor via a separate HDMI port, or charging, and does not need Thunderbolt storage, daisy-chaining, or dual high-resolution video	Surviving functions on the affected port, confirmed explicitly: charge, USB 2.0, USB 3.x, DisplayPort Alt Mode, each yes or no, using only those confirmed functions	Label the port mentally, not by sticking anything into it, so a future Thunderbolt device is not misdiagnosed as faulty when plugged into it	This is a live-with-it branch, not a repair, and must not be presented as fixed; a Thunderbolt device will not work on that port
Port is plain USB-C by specification, so no software change can make it Thunderbolt, or the laptop is Thunderbolt 3 and the user wants Thunderbolt 4 guarantees	Capability difference, not a fault	A different laptop, or a different connection method for the monitor or storage, decided after inspection	Do not present a Thunderbolt 3 to Thunderbolt 4 upgrade of the same port as available; do not present any driver, cable, or setting as turning a plain USB-C port into a Thunderbolt port

On price, stated once more to prevent misunderstanding: the only confirmed starting figures are battery from HK$580, screen from HK$1,280, motherboard chip-level from HK$880, and keyboard from HK$480, as listed on the pricing page. Thunderbolt port, Power Delivery controller, Thunderbolt controller, and board-level work are not among those four confirmed items, so every Thunderbolt cost in this guide is subject to a written quote after the diagnostic service, and the chip-level starting figure must not be presented as the Thunderbolt-specific price.

On replacement, the same caution applies: consider replacement only when inspection shows board-level controller failure, when the laptop has multiple concurrent faults, or when the reader's work genuinely requires Thunderbolt capability that no working port on that laptop can provide. Pushing a new laptop as the default answer to a cable or approval fault would be a misdiagnosis with a large cost attached.

On data, one caution applies to every branch: if a Thunderbolt storage device holds the only copy of data, do not run repeated firmware updates, BIOS changes, or board work until the data position is understood. Back up from a working port or machine first where possible. If the storage device itself is not recognised anywhere, that is a data-recovery question, not a port question, and this guide hands off rather than instructing risky recovery steps.

13. Hong Kong context: what can be said, and what must not be claimed

Only the following Hong Kong-context points are safe to publish, and each is worded as practical advice, not as measured local data. No Hong Kong statistic, humidity figure, corrosion rate, commuting damage rate, local parts availability claim, or local repair price beyond the confirmed figures in this guide was sourced, and none is written here.

Humidity and corrosion: Hong Kong humidity is a reason to look for visible corrosion or debris in a port and to keep a laptop out of a damp bag compartment, as inspection advice, not as a claim that humidity caused this reader's fault. If visible discolouration, debris, or a gritty plug feel is present, stop using the port and seek inspection; do not scrape contacts clean.

Commuting and bag or cable strain: a laptop carried daily with a cable left plugged in, or a cable bent sharply at the port in a bag, is a strain pattern worth asking about in the checklist, meaning whether a cable is left plugged in during carrying, and whether the cable is bent at the port. That is a history question, not a diagnosis, and no claim is made that commuting causes a specific failure rate.

Local repair reality, in safe wording: Thunderbolt port and controller work is board-level inspection work, not a same-visit cable swap. The customer should expect the shop to need the laptop, the exact cable, and the exact dock, monitor, or device that fails, or at minimum their exact model numbers, to reproduce the fault, because a port fault that only appears with one device cannot be confirmed without that device. Price is subject to a written quote after inspection, and no turnaround time is promised.

Language note: this guide keeps technical terms in English where the operating system or vendor uses English, including Device Manager, System Information, Thunderbolt Control Center, NVM, and BIOS security level, and explains them in plain language alongside, rather than inventing translated interface names that do not match what the reader sees on screen.

14. Pre-inspection checklist: bring this, in this order

Tell the reader to bring or send this list before or with the laptop. The shop cannot reproduce a port fault without it.

1. Laptop exact model number, from the base label or System Information, not just a brand or a screen size, and the operating-system version, meaning Windows 10 or 11 build, or macOS version.
2. Which port fails: left or right, front or back, and whether the laptop has another Thunderbolt port that works with the same setup.
3. Exact cable: brand and model if printed, length, Thunderbolt 3, Thunderbolt 4, USB4, or plain USB-C marking, and whether it is the cable that came with the dock or monitor.
4. Exact dock, monitor, or device: brand and model, and its firmware version if its utility shows it.
5. Four-way function split for the suspect port, tested separately: charges, USB device works, monitor or video works, Thunderbolt device, meaning storage or dock PCIe function, works. Record each as yes or no, not as sort of works.
6. What changed before it started: Windows or macOS update, BIOS or firmware update, new cable, new dock or monitor, drop, liquid, or nothing changed.
7. Error messages verbatim, photographed where possible: Thunderbolt Control Center prompts, Device Manager error codes, macOS no device connected versus device listed, any safe-mode message.
8. BIOS, driver, and firmware versions already checked: BIOS version, Thunderbolt driver version, NVM firmware version if shown, and the current BIOS Thunderbolt security level, photographed before any change.
9. Trigger pattern for intermittency: on movement, on wake from sleep, under heavy load, after idle, or random, with one concrete example, such as drops when the cable is moved at the laptop end.
10. What has already been tried, in order: other port, other cable, direct connection without dock or chain, other device, driver reinstall, BIOS check, firmware update, and the result of each. This prevents repeating the same test at inspection.

What to bring physically: the laptop, its charger, the exact cable, and, if possible, the exact dock, monitor, or device that fails. If the device cannot be brought, its exact model number and the checklist above are the minimum.

15. What not to do: the stop list

• Do not update the BIOS and controller firmware at the same time; do not change the cable, dock, driver, and BIOS at once.
• Do not bend or prop a loose port, and do not keep wiggling a cable as a fix; a wiggle test is a localisation test done once.
• Do not buy a dock to test a port; borrow the test order in Section 5 instead, meaning a short certified cable and a single device direct.
• Do not use another model's firmware package, do not interrupt a firmware update, and do not stack another firmware package on top of a failed update.
• Do not leave a lowered security level in place as the fix; restore the vendor-recommended setting after the test.
• Do not hot-plug internal connectors, and do not open the laptop to inspect a port or controller yourself; internal inspection is professional-only, powered off with the battery isolated.
• Do not present a USB-only fallback as fixed, and do not present a working second port as proof that the whole laptop is fine.

Shop details, for inspection: 沙田石門達利廣場1209室, telephone +852 2186 8010, WhatsApp +852 6558 6806, hours 12:00–19:00. Thunderbolt port or controller repair price is subject to a written quote after the diagnostic service; other confirmed starting prices are on the pricing page.

Frequently Asked Questions

Q1: The port still charges. Does that mean the port is fine?

No. Charging and data and video are separate negotiations: charging runs on USB Power Delivery negotiation, while data and video run on the high-speed link and its tunnels. Charging success only shows that power negotiation and the power pins are broadly working; it proves nothing about the high-speed pins, the controller, firmware, drivers, or the authorisation layer. Apple states explicitly that on Apple silicon Mac laptops an accessory can still charge the Mac even if the user chooses not to allow data access. So charges but no data and no video is a clue, not proof that the port is healthy. The exclusion steps are in the main text, prices are on the pricing page, inspection is via the diagnostic service, and any Thunderbolt repair price is subject to a written quote after inspection.

Q2: How do I confirm that my laptop's USB-C port is actually a Thunderbolt port?

The connector shape is not enough, and the lightning-bolt mark alone is not enough. The reliable check is the laptop's exact model number, then the vendor's specification page under Ports or Connectivity, which should state Thunderbolt 3, Thunderbolt 4, USB4, or only USB-C. If the specification says only USB-C, the port carries no Thunderbolt guarantee, and no cable or setting will turn it into a Thunderbolt port. Record the model, which port fails, and how the other port behaves with the same setup; all three matter at inspection.

Q3: Why does one port on the same laptop work while the other does not?

This is the strongest port-localising symptom in this guide. If the same cable and the same device work on port A and fail on port B, the cable, device, dock, driver, and controller firmware are substantially excluded, and suspicion concentrates on port B itself: its physical contacts, its retimer path, its BIOS setting, or its power and authorisation state. The one caveat is multi-controller designs, where two ports may sit on different controllers, so check the model specification before concluding. Do not keep buying docks to test this pattern; bring the whole cable and device set for inspection.

Q4: A dock or external drive is completely unrecognised in Windows. Does that mean the controller is dead?

Not yet. Complete non-recognition can come from cable capability, port capability, authorisation, dock power or firmware, a missing driver, a BIOS setting that disables Thunderbolt, or a controller firmware problem, and only after those does controller hardware become the leading suspect. The correct order is: one short certified Thunderbolt cable, one device, direct to the laptop with no dock, hub, or adapter; then check Device Manager for the controller; then check authorisation and BIOS settings. Only when the controller is entirely absent with the correct driver installed and Thunderbolt enabled in BIOS does the firmware or controller-hardware suspicion rise sharply. Price is subject to a written quote after the diagnostic service.

Q5: How can a BIOS security level cause video to work while a drive is not recognised?

Because DisplayPort video and PCIe data travel in different tunnels, and the security level governs the PCIe tunnel. In DisplayPort Only mode the port connects DisplayPort devices automatically but never creates a PCIe tunnel, so the monitor shows a picture while a Thunderbolt drive never appears. In User Authorization mode a new PCIe device needs user approval before its tunnel is created; a missed, dismissed, or never-shown prompt looks exactly like a dead port for PCIe devices while video stays normal. So a half-working setup should send you to the BIOS security level and approval state first, not to a hardware verdict. Photograph the original setting before any test change, and restore the vendor-recommended setting afterwards; a lowered security level is a test, never a permanent fix.

Q6: On macOS I clicked Do not Allow once. Why does charging still work but data does not?

That is Apple's documented behaviour. On an Apple silicon Mac laptop, a new or unknown USB accessory, Thunderbolt accessory, or memory card triggers an alert asking the user to allow it; if the user clicks Do not Allow, the Mac will not recognise the accessory or give it data access, but charging can still continue, and the next connection asks again. The remedy is to unlock the Mac, disconnect and reconnect, watch for the prompt, and check Apple menu, System Settings, Privacy and Security, Accessories. Two further traps apply: if the Mac has been locked for three or more days, a previously allowed accessory may need approval again; and a lid-closed docked test may never show the prompt at all, which must not be misread as a port fault.

Q7: What is the difference between active and passive cables, and why do long cables drop speed?

The difference is high-speed signal loss over distance. A passive Thunderbolt 3 cable only reaches full 40 Gbps over a short length, and sources disagree on the exact short length: one source says about 0.8 m and another says 0.5 m, so this guide states both rather than inventing a single threshold; longer passive cables fall back to about 20 Gbps. Active cables contain signal-processing electronics and can hold 40 Gbps to 2 m and, in one vendor comparison, to 3 m for Thunderbolt 4, while Thunderbolt 4 universal cable certification targets 40 Gbps up to 2 m. A 2 m passive cable that charges perfectly can therefore still starve a dual-display or high-speed storage setup. A control test must use a short certified cable, not another unknown long cable.

Q8: My device keeps disconnecting, but only on movement, on wake, or under heavy load. How do I separate those?

Split by trigger, because the trigger is the diagnostic. Movement only: hold the laptop and cable completely still, then move only the laptop-end plug, then only the device-end plug, then only the laptop, and record which movement reproduces the drop; laptop-end reproduction points toward the laptop port side, but this is a localisation test only, not a reason to keep bending the cable in use. Wake only, with a full restart always working: suspicion points to driver power management, sleep mode, or firmware state, and the documented test is to disable the allow the computer to turn off this device to save power option as a reversible comparison while noting the original setting. Heavy load only: test one device direct with no dock or chain, then rebuild one device at a time to see whether dock power or total bandwidth is the ceiling. On temperature, this research found no vendor-published controller temperature threshold, so this guide gives no temperature figure and no cooling fix as fact; record cold versus heavy-load behaviour as a pattern for inspection instead.

Q9: The controller is completely missing from Device Manager. Must the board be replaced?

No. A missing controller has layers: when the Thunderbolt driver is not installed, Windows shows a Base System Device under Other devices, and installing the vendor's exact-model driver should replace it with the controller entry; on Windows 11 USB4 systems the controller may appear as a Microsoft USB4 Host Router rather than the older Thunderbolt Controller name, and absence of the old name does not prove a dead controller; a BIOS setting that disables Thunderbolt also hides the controller. Only when the driver is correct, Thunderbolt is enabled in BIOS, and the controller is still entirely absent after a cold boot, or a safe-mode message appears, does the firmware or controller-hardware suspicion become leading. At that stage do not re-flash repeatedly and do not open the laptop; inspection decides, and any Thunderbolt repair price is subject to a written quote after inspection.

Q10: A firmware or NVM update failed, or the controller entered safe mode. Can I just flash it again myself?

No. Controller firmware is not the BIOS and not the driver, and updates must come only from the laptop vendor's package for the exact model. The vendor procedure is explicit: do not touch any key or shut the system down before the update finishes, then shut down and restart to activate the new firmware. If the controller enters safe mode, kernel documentation states that the only available functionality is flashing a valid NVM image; if the vendor offers no newer NVM firmware for that model, the documented position is to contact the vendor and ask when one will be available, not to use another model's package. After a failed update, stop, record the exact error and the versions before and after, keep the laptop on mains power, and get it inspected; stacking another firmware package on top of a failed one only increases the risk of leaving the controller unrecoverable.

Q11: Only one function fails, for example video works but USB does not. Does that mean the whole port is dead?

Absolutely not. One Thunderbolt port is a bundle of independent tunnels sharing one connector: charging, USB data, DisplayPort video, and PCIe data each take their own path. Video working with PCIe storage missing points to a healthy DisplayPort tunnel with a blocked PCIe tunnel or authorisation; USB devices working with a Thunderbolt device missing points to a healthy USB path with a failed Thunderbolt link or PCIe tunnel; a monitor failing on a dock while other dock functions work may be a DisplayPort tunnel, monitor input, or bandwidth contention issue. The useful record is a four-way split, written as yes or no for each: charges, USB works, video works, Thunderbolt device works. That four-way table is the single most useful piece of evidence to bring to inspection.

Q12: What should I prepare before inspection so the visit is not wasted?

Bring five things. First, the exact model number and operating system version, not just brand and size. Second, which port fails, its position, and how the other port behaves with the same cable and device. Third, the failing cable and device themselves, or at minimum the cable length and marking and the device's exact model, because without the original chain the fault may not reproduce. Fourth, the four-way function table: charges, USB, video, and Thunderbolt device, each yes or no. Fifth, the steps already tried in order, with results: which port, which cable, direct or via dock, which device, driver reinstall, BIOS check, firmware update, plus BIOS driver and NVM versions, the original BIOS security level, and anything that changed before the fault, including updates, drops, or liquid. Thunderbolt port or controller repair has no confirmed price, so it is subject to a written quote after the diagnostic service; other confirmed starting prices are listed on the pricing page.

Sources

• Intel Thunderbolt 4 announcement — Thunderbolt 4 certification minima: 40 Gbps, PCIe 32 Gbps, two 4K or one 8K, docks with up to four Thunderbolt ports, PC charging on at least one port, VT-d DMA protection, universal cables up to 2 m. https://www.intel.com/content/www/us/en/newsroom/news/thunderbolt-4-universal-cable-connectivity.html
• Intel Thunderbolt 4 announcement summary — same certification minima, corroboration. https://www.funkykit.com/news/intel-announces-thunderbolt-4-universal-cable-connectivity-for-everyone/
• Intel product comparison — Thunderbolt 4 controller description: 40G support, native USB-C, tunnelling including 32G PCIe. https://www.intel.com/content/www/us/en/products/compare.html?productIds=189982
• Linux kernel Thunderbolt documentation — security levels and BIOS names, user authorisation before PCIe tunnels, NVM safe mode. https://www.kernel.org/doc/html/v5.12/admin-guide/thunderbolt.html
• Linux kernel Thunderbolt documentation, earlier version — same security-level family. https://www.kernel.org/doc/html/v5.7/admin-guide/thunderbolt.html
• Dell community, Demystifying Thunderbolt 3 Security Levels — SL0, SL1, SL2 definitions, user prompt options, SL2 challenge and response. https://www.dell.com/community/en/conversations/latitude/demystifying-thunderbolt-3-security-levels/647f8742f4ccf8a8de66b949
• Intel Thunderbolt 3 and Security on Windows 10 — SL0 to SL3, SL1 as default, per-machine access-control list, DisplayPort connects by default. https://thunderbolttechnology-netproxy.boingomedia.com/security/Thunderbolt%203%20and%20Security.pdf
• Thunderbolt interface summary — controller generations Titan Ridge, Maple Ridge, Barlow Ridge. https://en.wikipedia.org/w/index.php?title=Thunderbolt_(interface)&oldid=1367447017
• JAE USB4 overview — tunnelling concept, dynamic bandwidth sharing, host and hub optionality, USB4 20 Gbps and 40 Gbps. https://www.jae.com/en/column/21-brief-Introduction-to-usb4/
• Kingston USB4 explainer — USB4 20 Gbps and 40 Gbps naming, corroboration. https://www.kingston.com/unitedkingdom/en/blog/mobile-lifestyle/usb4-explained
• USB4 specification summary — USB4 based on the Thunderbolt protocol contributed by Intel, tunnels for USB 3.x, DisplayPort and PCIe. https://en.wikipedia.org/wiki/USB4
• Active versus passive Thunderbolt cables — passive Thunderbolt 3 up to 0.8 m at 40 Gbps, longer passive at 20 Gbps, active to 2 m and 3 m in that comparison. https://recables.com/cables/thunderbolt/active-vs-passive-thunderbolt-cables/
• Plugable knowledge base — longest passive Thunderbolt 3 cable rated for 40 Gbps included with a dock is 0.8 m; longer cables need to be active. https://kb.plugable.com/docking-stations-and-video/can-i-use-a-longer-thunderbolt-3-or-usb-c-cable-to-connect-the-dock-to-my-system
• Apple Support, Allow USB and other accessories to connect to your Mac — approval alert, locked-Mac rule, three-day re-prompt, charging still works after Do not Allow, Privacy and Security Accessories settings, lid-closed note. https://support.apple.com/en-us/102282
• Apple silicon USB4 and Thunderbolt controller, secondary — M1 implements Apple's first custom controller for USB4 and Thunderbolt 3. https://appleinsider.com/articles/20/11/18/apples-m1-macs-deliver-worlds-first-support-for-usb4/amp/
• How to check for Thunderbolt 3, 4 and USB4, secondary guide — port mark does not identify generation, model specification is most reliable, Windows and macOS check paths. https://www.purplelec.com/info-detail/thunderbolt-3-products
• MSI Thunderbolt firmware update guide — NVM firmware version location, BIOS Thunderbolt firmware update toggle, do not interrupt the update, shutdown and restart to activate, tool error meanings. https://www.msi.com/files/pdf/Thunderbolt_update_en.pdf
• Windows Thunderbolt troubleshooting guide, secondary — Device Manager Thunderbolt Controller path, warning-icon driver update, power-management uncheck for intermittent disconnects, second-computer test. https://www.pchardwarepro.com/en/How-to-troubleshoot-Thunderbolt-problems-in-Windows/
• Microsoft Q and A, Thunderbolt 3 devices in Device Manager — Base System Device appears when the Thunderbolt driver is not installed; single report of a controller disappearing after a Windows update. https://learn.microsoft.com/en-us/answers/questions/3766644/thunderbolt-3-usb-devices-device-mgr-and-thunderbo
• Microsoft TechCommunity, Windows 11 Thunderbolt 4 and USB4 Host Router discussion — community reports plus a quoted Intel reply; community source only. https://techcommunity.microsoft.com/discussions/windows11/windows-11-thunderbolt-4-bug/3647965
• Lenovo critical Thunderbolt software and firmware updates — ThinkPad symptom list including controller not visible, safe-mode error, BIOS communication error. https://support.lenovo.com/bm/en/solutions/HT508988
• Dell WD19 Thunderbolt user guide — No Security, User Authorization, Secure Connect and DisplayPort Only definitions, BIOS path. https://www.dell.com/support/manuals/en-ca/dell-wd19tb-dock/wd19_tb_userguide/Thunderbolt-Security?guid=guid-537bc8b4-29b8-42b5-ba5f-31183fc69599&lang=en-us
• Dell support, USB-C and Thunderbolt docks not recognised — Group Policy image block, Kernel DMA Protection pre-OS behaviour. https://www.dell.com/support/kbdoc/en-yu/000146415/usb-c-and-thunderbolt-docks-not-recognized-by-the-system?lang=en
• Plugable Thunderbolt 3 support and troubleshooting on Windows PCs — hybrid-graphics note, maker-only NVM position. https://kb.plugable.com/thunderbolt-3-docks/thunderbolt-3-support-and-troubleshooting-on-windows-pcs
• Plugable USB4 versus Thunderbolt 4 — certification process, 40 Gbps signal integrity up to 2 m, PD and E-Marker cross-checks. https://plugable.com/blogs/news/usb4-vs-thunderbolt-4-who-sets-the-rules-what-s-guaranteed-and-how-cables-get-certified
• TechAdvisor Thunderbolt 4 and USB4 cables — cable length and speed table. https://www.techadvisor.com/article/725009/best-thunderbolt-4-and-usb4-cables.html
• OWC Thunderbolt 4 cable guide — Thunderbolt 4 host minimums, cables to 2 m at 40 Gbps, Thunderbolt 3 passive under 1 m for full bandwidth. https://www.owc.com/blog/everything-you-need-to-know-about-thunderbolt-4-usb-type-c-cables
• HP monitor connectivity glossary — USB-C DisplayPort Alt Mode separated from Thunderbolt 4 and 5 rows. https://www.hp.com/us-en/tech-takes/monitors/glossary/monitor-connectivity.html
• Kordz USB demystified — DisplayPort Alt Mode two-lane versus four-lane, USB4 and Thunderbolt tunnelling, dynamic bandwidth sharing. https://www.kordz.com/articles/usb-demystified-what-av-integrators-need-to-know-about-usb-versions-thunderbolt-displayport-alt-mode/
• BenQ DisplayPort Alt Mode explainer — monitor charging without video when DisplayPort Alt Mode is missing, charge-only cables. https://www.benq.eu/en-uk/knowledge-center/knowledge/usb-c-introduction-what-is-dp-alt-mode.html
• Windows Forum, charging but no monitor signal — charging is a poor test of video capability, charge-only and USB 2.0-only cables lack high-speed lanes. https://windowsforum.com/news/usb-c-charging-but-no-monitor-signal-how-to-choose-video-capable-cables-for-windows-laptops.449875/
• StarTech Thunderbolt 3 basics — Thunderbolt 3 supports Thunderbolt, USB, DisplayPort and Power Delivery on USB-C, up to 40 Gbps, daisy chain up to six devices. https://www.startech.com/en-us/blog/thunderbolt-3-the-basics?ref=hackernoon.com