Slow file transfers reward a methodical approach because the bottleneck is rarely what you first suspect. People blame Windows, buy new drives, and reinstall — when the actual problem was a USB 2 cable, or antivirus scanning, or a nearly-full destination disk. This guide is the ordered diagnostic sequence: measure, isolate each layer, fix what you find, and verify with numbers.

The discipline that makes this work is the baseline test: pick one large test file (a 1–2 GB video or ISO), copy it the same way before and after every change, and record the sustained MB/s. Without that, you are guessing whether each fix helped. With it, each step produces a verdict.

A word on expectations: know your hardware’s ceiling. A USB 3.0 flash drive tops out around 100–400 MB/s depending on quality; USB 2 caps at ~35 MB/s; a spinning HDD does 80–150 MB/s sequential. If you are getting 8 MB/s on USB 3 hardware, something is wrong. If you are getting 90 MB/s on a budget flash drive, that may simply be the drive.

Why this happens

  • USB 2 negotiation: Wrong port, bad cable, or hub silently downgrades a USB 3 drive to ~35 MB/s.
  • Antivirus on-write scanning: Real-time protection scans every file as it lands, often halving throughput.
  • Small-file overhead: Thousands of tiny files multiply per-file filesystem operations.
  • Full destination disk: SSDs and HDDs both slow sharply with little free space.
  • Failing drive: Bad sectors cause retries that collapse speed to single digits.
  • Stale chipset/USB drivers: Poor link negotiation and inefficient transfers on outdated drivers.
  • Write caching disabled: Every write waits for physical commit without buffering.
  • Background contention: Indexing, sync, or updates stealing disk bandwidth mid-copy.

How to fix it

Work through these in order — diagnose first, then act.

Fix 1: Establish the baseline measurement

Pick your test file — a single 1–2 GB file, not a folder of small files — and copy it from source to destination while timing with a stopwatch. Note the sustained speed after the initial burst settles (the first seconds reflect cache, not the drive). Record: source, destination, connection type, file size, elapsed time, computed MB/s. This number is the reference for every fix below. If you change the test file or method between fixes, the comparison is meaningless — keep the test identical. The HP-focused version of this sequence is our HP slow transfer guide .

  1. Choose a single 1–2 GB test file.
  2. Copy it and time with a stopwatch; ignore the initial speed burst.
  3. Compute MB/s (file size ÷ seconds) and write it down.
  4. Use this exact test after every fix below.

Fix 2: Rule out the physical layer

Test the cable, port, and hub path before touching software. Move the drive to a confirmed USB 3 port directly on the laptop (no hub), swap in a known-good USB 3 cable, and re-run the baseline. If speed jumps to expected levels, you are done — the fix was physical. Also try a different port on the same machine; individual ports can have damaged pins or controller issues. Document which combination works so you do not rediscover this next month.

  1. Connect directly to a USB 3 port on the laptop (blue insert/SS mark).
  2. Swap to a known-good USB 3 cable.
  3. Re-run the baseline test and compare.
  4. If fixed, label the good port/cable and stop here.

Fix 3: Isolate antivirus scanning

With the physical layer confirmed, test the antivirus contribution: pause real-time protection briefly and re-run the baseline copy. A significant jump confirms scanning overhead. The permanent answer is not disabling protection — it is targeted exclusions for trusted bulk-transfer folders, or scheduling big copies with protection paused only for their duration. Note the with/without numbers; on some systems with many small files the difference is 2–3x, which reframes the whole problem.

  1. Pause real-time protection in Windows Security.
  2. Re-run the identical baseline copy.
  3. Record the speed difference.
  4. Re-enable protection; add exclusions for trusted transfer folders.

Fix 4: Test small files vs. large files separately

If your real-world transfer is many small files, run a second baseline: copy a folder of a few thousand small files and time it. Compare against the single-large-file baseline. A huge gap proves per-file overhead is the bottleneck, and the fix is archival — zip the set before transferring (see next fix) — not hardware. This distinction saves people from buying faster drives to solve what is actually a filesystem-operations problem.

  1. Time a copy of several thousand small files.
  2. Compare against the single-large-file baseline.
  3. A large gap confirms per-file overhead, not hardware.
  4. Proceed to archival (Fix 5) rather than hardware changes.

Fix 5: Archive small files before transfer

Convert the many-files problem into a one-file problem: compress the set into a ZIP or 7z archive, transfer the archive at full speed, and extract at the destination. Time the whole operation (compress + transfer + extract) against the direct copy — the archive route wins by large margins on sets of thousands of files. For recurring transfers (backups, project folders), script it: a scheduled 7-Zip archive plus Robocopy is the professional pattern.

  1. Compress the file set (built-in ZIP or 7-Zip).
  2. Transfer the single archive; note the speed.
  3. Extract at the destination.
  4. Compare total time vs. direct copy.

Fix 6: Fix the destination: free space, health, and caching

Turn attention to where the data lands. Verify 15%+ free space, check SMART health (wmic diskdrive get status), and enable write caching (Better performance policy in Device Manager) for the destination drive. A nearly-full or failing destination produces exactly the “fast start, then collapse” copy curve people misattribute to Windows. If health warnings appear, stop optimizing and back up — a dying destination is a data-loss problem wearing a speed-problem costume. Our C: drive cleanup guide covers freeing space in depth.

  1. Confirm 15%+ free space on the destination.
  2. Check SMART health; back up immediately on warnings.
  3. Enable write caching (Device Manager > Disk drives > Policies > Better performance).
  4. Re-run the baseline and compare.

Fix 7: Update drivers and eliminate contention, then verify

The final layer: update chipset/USB drivers from your manufacturer’s support site (Intel’s detection tool helps identify Intel-side updates), then re-run the baseline with background contention eliminated — OneDrive paused, no indexing storm, no updates downloading. Record the final number against the Fix 1 baseline. Whatever gap remains between your final speed and the hardware’s rated ceiling is the true hardware limit; everything above that gap that you recovered was software and configuration.

  1. Update chipset/USB drivers from the manufacturer’s site.
  2. Pause OneDrive/sync and confirm no background disk storms.
  3. Re-run the identical baseline test.
  4. Compare final vs. initial numbers to quantify the total improvement.

Closing

The step-by-step method turns “file transfers are slow” from a vague complaint into a measured diagnosis with a quantified fix. Physical layer, antivirus, file-size profile, destination health, drivers — each step either finds the bottleneck or eliminates a suspect. Keep your baseline test file around; the next time a copy feels slow, one timed run tells you instantly whether something regressed.

Frequently asked questions

How do I accurately measure my file transfer speed?
Copy a single large file (1 GB or more) and watch the sustained speed in the Windows copy dialog after the first few seconds — ignore the initial burst, which reflects cache. For precision, time it with a stopwatch and divide file size by seconds. Consistently testing the same file before and after each fix is how you prove what worked.
Why is my USB 3 drive only transferring at USB 2 speeds?
The drive is negotiating a USB 2 connection, almost always because of the port (plugged into a USB 2 port), the cable (a damaged or charge-only cable), or a hub in the path. Less commonly, outdated chipset drivers fail to negotiate USB 3. Test with a direct connection, a known-good cable, and a confirmed USB 3 port.
Does the destination drive's free space affect copy speed?
Yes, significantly. SSDs slow down as they fill because the controller has fewer free blocks for fast writes, and HDDs suffer fragmentation and longer seeks. Keep at least 15% free on any drive you copy to regularly. If the destination is nearly full, freeing space is often the fastest fix available.
Should I use Robocopy instead of File Explorer for large copies?
For very large or interrupted transfers, yes. Robocopy handles retries, resume (/Z), and multi-threaded copying (/MT), and it logs errors instead of aborting the whole job on one bad file. For everyday copies File Explorer is fine, but for bulk migrations Robocopy is the technician’s tool of choice.