USB speed depends on the flash media, connector, host port, file size, and the other drive in the copy. A label such as USB 3.x describes an interface capability, not a guaranteed sustained write rate. Small-file copies usually run much slower than large sequential transfers.
Step 1. Back up important files. Plug the drive directly into a known USB 3.x port, without a hub. Note the drive model, file system, free space, and port type.
Step 2. Close large downloads and background copies. Use the same PC, port, and test size for comparisons. Allow a hot USB drive to cool; thermal throttling can change results.
Step 1. Choose a single large test file already stored on a healthy internal SSD. In File Explorer, copy it to the USB drive and note the file size and elapsed time after completion.
Step 2. Calculate approximate write MB/s as file size in MB divided by elapsed seconds. Copy the same file back to a different folder on the internal SSD for a read test. Do not rely only on File Explorer's momentary peak display.
Step 3. Repeat once after safely ejecting and reconnecting. A much slower second write can reflect an exhausted write cache or heat, not necessarily a broken USB port.
For very small files, report the total folder size and file count rather than pretending the result is a pure sequential speed. Windows has to create directory entries and metadata for every file, so a folder of thousands of photos may copy much slower than one video of the same size.
Step 1. Install a reputable benchmark from its publisher. Select the USB drive by letter and capacity. Choose a moderate test size that fits easily in its free space.
Step 2. Run sequential read and write tests, then a random-I/O test if needed. Record units and test size. Benchmarks write temporary data, so avoid them on a drive with missing files or suspected physical failure.
Test another cable or port and compare a second drive on the same port. Check for counterfeit capacity, file-system errors, and repeated disconnects. If the USB drive contains irreplaceable data, copy it first and replace an unreliable device.
If the reported write speed falls only after a long continuous transfer, the drive may have a small fast cache and slower underlying flash. Compare the sustained rate after the cache fills, not just the first few seconds. A higher-rated interface will not improve a drive whose memory is the bottleneck.
Qiling Disk Master's disk-health function is not a USB speed benchmark. Some removable drives expose no meaningful SMART information, so a missing health value should not be interpreted as a pass.