You finish a photo or video shoot, insert a CFexpress card into the reader, connect it through your USB-C hub—and something is wrong.

Sometimes the card never appears.

Other times it mounts normally, but a transfer that should feel fast crawls along at a few hundred megabytes per second or less.

For photographers and video creators moving RAW photos, 4K/8K footage, or large project folders, those are two very different problems:

Not detected means the data path is not being established correctly.

Detected but slow means the path works, but one part of it is limiting performance.

The most useful way to troubleshoot both is to stop treating the CFexpress card reader as one device.

Instead, look at the complete ingest chain:

CFexpress Card → Card Reader → Cable → USB-C Hub → Computer → Destination Storage

Your transfer can only perform as well as the slowest compatible part of that chain.

First: Make Sure You Actually Have the Right Reader

Before measuring speed, confirm the memory-card standard.

CFexpress is not the same as CompactFlash (CF).

The names look similar, but the technologies are very different.

The CompactFlash Association's official card guide identifies CFexpress as a PCI Express/NVMe-based format with three form factors:

• CFexpress Type A
• CFexpress Type B
• CFexpress Type C

Legacy CompactFlash uses a different interface and should not be treated as interchangeable with CFexpress.

That means a reader labeled:

CF / CompactFlash Reader

does not automatically support:

CFexpress Type B

Likewise, a CFexpress Type A reader cannot simply read a Type B card.

Before troubleshooting the USB-C hub, check the label or specification of both:

your CFexpress card type
and
your dedicated reader

If those do not match, changing cables or USB ports will not solve the problem.

Build a Baseline Before You Benchmark Anything

The fastest diagnosis comes from four short tests using the same card and reader.

Do not change the card, reader, computer, and cable simultaneously.

Record one result at each stage.

This gives you a usable performance baseline.

For example:

That result tells a very different story from:

“My CFexpress reader is slow.”

The reader itself is capable of higher performance.

The slowdown appears when the hub is carrying additional traffic.

If the CFexpress Reader Is Not Detected

Start with the data connection, not the benchmark.

Check the Hub Port Label

A USB-C connector does not guarantee that every USB-C port performs the same function.

Your hub may include ports labeled:

USB-C Data
5Gbps
10Gbps
USB 3.2

and another USB-C port labeled:

PD
100W PD
Power Delivery
Charging

A PD-only input is generally intended for charging the laptop through the hub, not for connecting a card reader.

If the reader powers on but the CFexpress card never appears, confirm that it is connected to a real USB data port.

USB-IF's official USB Type-C guidance makes an important distinction between connector type and data capability. USB-C products can support very different data rates, including 480Mbps, 5Gbps, 10Gbps, 20Gbps, and higher-performance USB implementations.

So:

USB-C describes the connector. It does not guarantee one transfer speed or function.

Use the Direct Reader Test as Your Reference

Disconnect the hub.

Connect:

CFexpress Reader → Computer

using the reader's normal high-speed cable.

The card still does not appear

Do not blame the USB-C hub yet.

Check:

• the CFexpress card type;
• reader compatibility;
• reader cable;
• computer USB port;
• another CFexpress card;
• another computer;
• Windows Disk Management;
• macOS Disk Utility or System Information.

If the reader also fails directly, the hub is not the first layer to troubleshoot.

The reader works directly

Now return the hub to the path.

If:

Direct → detected

but:

Through Hub → not detected

you have isolated the problem to the intermediate connection.

Now test the hub port, data speed, cable, power, and connected peripherals.

If the Reader Appears but Transfer Speed Is Poor

This is where CFexpress differs from ordinary SD-card troubleshooting.

CFexpress was designed for high-performance imaging workflows.

The CompactFlash Association specifies CFexpress using PCIe and NVMe, and CFexpress 4.0 significantly increases the available media-side performance. Type B in CFexpress 4.0, for example, has a theoretical bus throughput of up to 4 GB/s. (CompactFlash Association)

The USB connection outside the card reader may be much slower.

Consider these theoretical signaling ceilings:

Actual file transfer will be lower because protocol overhead, reader controllers, storage performance, thermals, and workload all matter.

But the table reveals the important point:

A fast CFexpress card cannot bypass a slower USB link.

A reader capable of 10Gbps connected through a 5Gbps USB-C hub is now operating inside a 5Gbps-class path.

A USB4 reader connected through a 10Gbps hub is likewise limited by the 10Gbps intermediate link.

Check the Link Before You Blame the Card

Do not rely only on Finder or File Explorer transfer estimates.

First determine what USB connection was actually negotiated.

Our guide on how to check the actual USB link speed on Windows and macOS explains how to separate three different numbers:

Card/reader advertised capability

vs.

Negotiated USB link speed

vs.

Real file-transfer throughput

Suppose your reader advertises 10Gbps.

Direct connection

Negotiated link: 10Gbps

Through the hub

Negotiated link: 5Gbps

You have already found a major bottleneck.

There is little value in changing camera settings or benchmarking the CFexpress card repeatedly until that USB-path difference is understood.

The Same-Hub Copy Trap

There is another performance problem that matters especially to creators.

Imagine this workflow:

CFexpress Reader → USB-C Hub

and:

External SSD → Same USB-C Hub

Then you copy:

CFexpress Card → External SSD

It looks like a direct transfer between two storage devices.

It is not.

Both devices communicate through the hub's upstream connection to the computer.

So the workflow is really closer to:

CFexpress Reader
↓
Hub
↓
Computer
↓
Hub
↓
External SSD

The reader and SSD may both be individually fast, but they are now competing for the hub's shared connection.

This is why you may see:

Reader → internal laptop SSD = fast

but:

Reader → external SSD on same hub = much slower

If your normal workflow includes an external SSD, temporarily disconnect it and import the CFexpress files to the computer's internal SSD.

Compare the result.

If performance improves dramatically, the card reader itself may be working correctly.

You are seeing shared-bandwidth behavior.

Our guide to why USB-C hubs slow down when fully loaded covers this multi-device bottleneck in more detail.

Use One Large File Before Testing Thousands of Small Files

A folder containing thousands of RAW photos may transfer differently from one large video file.

That is because real file copying is affected by:

• file count;
• file-system metadata;
• destination-drive performance;
• antivirus scanning;
• Finder/File Explorer overhead;
• preview generation;
• thermal behavior.

For a cleaner performance test, start with one or several large video files.

If a single large file transfers quickly but a folder of thousands of photos is slower, the USB connection may not be the only limit.

This matters because:

slow workflow ≠ slow USB link

and:

fast USB link ≠ guaranteed maximum file-copy speed

Watch the Reader Temperature During Long Offloads

CFexpress workflows can involve sustained high-speed reads for many minutes.

That creates heat in:

the card

the reader controller

the hub

and sometimes:

the external SSD receiving the files

Some cards and readers can reduce performance when temperature rises.

So if the transfer begins quickly but gradually slows after several minutes, repeat the test after allowing the reader and card to cool.

Also avoid:

• placing the reader under the laptop;
• covering the hub or reader;
• stacking hot devices together;
• running multiple storage transfers simultaneously during the test.

A thermal slowdown looks different from a fixed USB bottleneck.

Fixed bottleneck

Transfer begins slow and stays relatively similar.

Thermal behavior

Transfer starts fast and gradually drops during sustained use.

That difference is useful evidence.

Give High-Speed Storage the Fastest Hub Port

If your USB-C hub contains ports with different speeds, do not assign them randomly.

Put lower-bandwidth peripherals such as:

keyboard → mouse → USB receiver

on slower ports when possible.

Reserve the highest-speed data port for:

CFexpress reader → external SSD → other high-speed storage

For example, if your dedicated CFexpress reader itself uses a 10Gbps USB-C interface and you need to keep several USB-C data devices connected, a USB-C hub with four USB 3.2 Gen 2 10Gbps data ports and 100W PD gives you clearly specified 10Gbps USB-C data connections rather than forcing a high-speed reader onto a 5Gbps or USB 2.0 port.

That type of setup makes sense when:

Reader capability = 10Gbps

and:

Hub data port = 10Gbps

and:

Host connection supports the required USB speed

But it is important not to overstate what a 10Gbps hub can do.

A 20Gbps or USB4 CFexpress reader will not suddenly run at 20Gbps or 40Gbps through a 10Gbps hub.

The hub simply removes a 5Gbps-class bottleneck for a 10Gbps-class reader.

Don't Let the Destination Drive Become the Hidden Bottleneck

Suppose the CFexpress card, reader, hub, cable, and host all negotiate correctly.

The transfer can still be slow because the destination cannot accept data quickly enough.

Test:

CFexpress → Internal SSD

then compare:

CFexpress → External SSD

If the internal SSD transfer is substantially faster, investigate the destination storage.

Possible limits include:

• external SSD USB link speed;
• SSD enclosure;
• destination cable;
• nearly full SSD;
• thermal throttling;
• slow write cache;
• another shared hub connection.

If the destination SSD is also disappearing or failing to mount through the hub, use our external SSD detection troubleshooting guide before benchmarking the CFexpress workflow.

A benchmark is only useful when both source and destination remain stable.

Read Your Test Results Like a Workflow Engineer

After testing, your result should fit one of these patterns.

Reader fails directly and through the hub

Focus on:

card type → reader → reader cable → OS → host port

The hub is not yet the main suspect.

Reader works directly but is not detected through the hub

Focus on:

hub data port → cable → power → USB enumeration → hub compatibility

Reader is 10Gbps directly but negotiates 5Gbps through the hub

The hub path is limiting the connection.

Reader negotiates 10Gbps through the hub but real transfer is still slow

Check:

card performance → destination drive → workload → thermals → shared traffic

Reader is fast through the hub alone but slow when other devices are attached

Investigate shared bandwidth and simultaneous peripheral activity.

Reader → internal SSD is fast, but Reader → external SSD on the same hub is slow

The same-hub storage workflow is the strongest clue.

Final Takeaway

When a CFexpress card reader is not detected or runs slowly through a USB-C hub, start by separating detection from performance.

For detection, verify:

correct CFexpress type → compatible reader → real USB data port → direct connection

For performance, follow the entire ingest chain:

CFexpress Card → Reader → Cable → Hub → Host → Destination Drive

Then compare the same reader:

Direct

vs.

Through Hub

vs.

Through Hub Alone

vs.

Through Hub With Your Full Workstation

The fastest CFexpress card does not determine the final transfer rate.

Neither does the fastest reader.

The slowest active link in the complete workflow sets the ceiling.

Once you know which link changes the result, you know whether to troubleshoot the reader, USB-C hub, cable, host connection, shared bandwidth, or destination storage—instead of replacing hardware at random.