Yes, roughly 14 MB/s can be a perfectly normal sustained speed when copying thousands of small files to a microSD card, even on a fast USB 3.1 Gen 1 reader. The USB connection is almost never the limiting factor here — the card's NAND flash memory and the nature of the workload are. Understanding where the time goes will tell you whether your setup is healthy or genuinely underperforming.
The reader and the card are not the same bottleneck
A USB 3.1 Gen 1 (SuperSpeed) link offers around 5 Gbps of theoretical bandwidth, or roughly 400–450 MB/s in practice. Your reader connecting at SuperSpeed, as confirmed by tools like USB Device Tree Viewer, means the interface itself is not the constraint. Likewise, the card's marketing numbers — 130 MB/s read, U3/V30 speed classes — describe sequential read and minimum sustained sequential write under ideal conditions with large blocks.
Real-world copying to a card is dominated by random and small-file write performance, which manufacturers rarely advertise because it is far lower than the headline figures. A card that sustains 30+ MB/s writing one huge video file can easily drop to 10–20 MB/s when the transfer consists of thousands of small items, which is exactly your scenario: 5,196 items across 172 GB averages out to files of a few tens of megabytes, with plenty of small ones mixed in.
Why small files are so much slower
Three properties of SD card storage work against many-file transfers:
- Erase-before-write overhead. NAND flash must erase entire blocks before they can be rewritten. The card's controller manages this, but on a used or heavily fragmented card, more of each write cycle is spent on housekeeping, cutting throughput.
- Filesystem metadata churn. Every file written also updates the file allocation table and directory entries. With thousands of files, the card shuttles between writing data and updating metadata, and each switch costs time — a bit like a tape drive constantly rewinding to update its index.
- Thermal throttling. Sustained writes heat the card. When it gets too warm, the controller deliberately slows down to protect the flash, so speed often starts high and decays over a long copy session.
Windows File Explorer also reports an instantaneous, fluctuating average rather than a benchmark, so dips during metadata-heavy stretches drag the displayed number down even if bursts hit 80–90 MB/s.
How to verify your setup is healthy
Before assuming a hardware fault, run a controlled test:
- Copy a single large file (a 4–8 GB video or disk image) to the card and watch the speed. If it reaches 60–90 MB/s initially and then settles lower, your reader, port, and card are all working as designed.
- If even large-file writes never exceed ~20 MB/s, try a different reader and a different USB port. Some card readers use cheap bridges that cap speed regardless of the card.
- Check that the card is formatted as
exFAT(default for large cards) and that Windows is not applying some unusual write policy.
A quick way to rule out the reader is to test the same card in another device, or test a different card in your reader. If both combinations are slow only for small files, nothing is broken.
Ways to speed up the transfer
You usually cannot change the card's physics, but you can reduce the overhead:
- Archive first, then copy. Compressing the 5,196 items into a single archive (even a store-only ZIP with no compression) turns thousands of tiny writes into one large sequential write. This is often the single biggest win.
- Let the copy finish without interruption. Restarting a partial copy forces the card to re-verify and re-write, adding overhead.
- Use a fresh or recently formatted card. A wiped card has clean erase blocks and performs better than one that has been filled and deleted from repeatedly.
- Keep the card cool. Avoid enclosing the reader in a pocket or against a warm laptop surface during long transfers.
When an SD card is the wrong tool
If your goal is simply to move 172 GB quickly and repeatedly, a microSD card is a poor choice regardless of brand or rating. A small portable SSD connected over the same USB 3.x port will sustain hundreds of megabytes per second, handles small files far better, and does not throttle the way flash cards do. For one-off transfers to devices that only accept SD cards (cameras, handheld consoles, phones), the slow write is the price of the format — but for general external storage, an SSD is dramatically faster and no larger than a card reader.
Frequently Asked Questions
My card says V30, so shouldn't it write at least 30 MB/s always?
The V30 rating guarantees a minimum sustained sequential write of 30 MB/s under controlled conditions. Small-file and random writes are not covered by the rating and are routinely much slower — often a third or less of the sequential figure.
Could a USB 3.2 or USB-C reader help?
Only marginally, and only for high-end cards. A UHS-I microSD card tops out around 90–104 MB/s regardless of reader bandwidth. To exceed that you need a UHS-II card and a UHS-II-capable reader — and even then, small-file performance remains limited by the card's controller.
Why did the speed start fast and then drop?
That pattern is normal. The card's SLC-style cache absorbs the first gigabytes at high speed, then the controller must write data to slower TLC/QLD flash while erasing used blocks, and heat buildup can trigger throttling. Long transfers settle at the card's true sustained rate.
Is 14 MB/s a sign my card is counterfeit?
Not by itself, given the small-file workload. Counterfeit cards usually fail in more dramatic ways: reported capacities that don't match reality, corruption on writes, or sequential speeds far below even the rated minimum. Verify capacity and sequential speed with a dedicated test tool if you're suspicious.
Would a different file system help?
Not meaningfully. exFAT, FAT32, and NTFS all incur per-file metadata updates; the flash erase overhead dominates either way. The archive-first approach sidesteps the problem far more effectively than switching file systems.