NAND ECC engineering: from RBER to UBER, BCH, LDPC and read retry
An engineering guide to NAND error correction: connect RBER to codeword failure, distinguish BCH from LDPC, and qualify soft decoding and read-retry latency.
NAND read disturb and program interference: mechanisms and qualification
A device-level guide to NAND read disturb and program interference: threshold shifts, coupling, wear interaction, refresh, evidence and qualification tests.
SD card CID, CSD and SCR for OEM traceability
Plain-English guide to SD and microSD CID, CSD and SCR data: what they identify, what they do not prove, and how OEM buyers record traceability evidence.
eMMC production provisioning: boot, RPMB and enhanced areas
Plan eMMC production provisioning for boot partitions, enhanced areas, reliable write and RPMB keys, with irreversible steps, verification and traceability.
SSD over-provisioning for endurance and steady performance
Specify SSD over-provisioning for OEM workloads by balancing capacity, steady write performance, endurance, write amplification and qualification evidence.
Flash storage qualification plan: samples to production
Build an OEM flash-storage qualification plan covering BOM, workload, steady state, latency, temperature, power cycles, health data and release evidence.
eMMC BKOPS: background operations and latency spikes
Learn why eMMC background operations create latency spikes, how BKOPS is enabled and scheduled, and what OEM teams should measure before production release.
SSD write amplification: measure WAF and reduce wear
Write amplification can consume SSD endurance beyond what host-write counters show. Learn how to calculate WAF, measure it over time and reduce avoidable wear.
eMMC health: PRE_EOL_INFO and DEVICE_LIFE_TIME explained
Learn how to read eMMC PRE_EOL_INFO and DEVICE_LIFE_TIME values, avoid the “percent remaining” trap, and plan service from trends across embedded fleets.
NVMe thermal throttling: temperature, speed and heatsinks
Learn why an NVMe SSD slows when hot, how to confirm thermal throttling, and when airflow or a correctly fitted heatsink can help under sustained workloads.
NAND bad blocks and ECC: what flash buyers should verify
Factory bad blocks can be normal. Learn how ECC and controllers manage NAND errors—and which health, spare-capacity and qualification checks buyers need.
SLC cache explained: burst vs sustained flash write speed
Peak speed may be a short cache burst, not a long transfer. Learn how SLC buffers, NAND, free space, files and interfaces shape sustained writes.
NVMe SMART health explained: percentage used, media errors and unsafe shutdowns
Learn which NVMe SMART fields matter, why 100% percentage used is not an instant failure flag, and when available spare, media errors, temperature or read-only warnings require action.
Why an SSD slows down when full: TRIM, garbage collection and over-provisioning
A nearly full SSD has fewer clean blocks for writes, increasing garbage collection and write amplification. Learn what TRIM does, how over-provisioning helps and what to check before replacing the drive.
AEC-Q100 temperature grades: how to verify an “automotive” flash claim
AEC-Q100 Grade 0, 1, 2 and 3 cover different IC ambient-temperature ranges. Learn why an automotive-qualified NAND component does not automatically qualify a complete SD card or SSD.
microSD Express vs UHS-I: what Switch 2 and Steam Deck buyers need to know
Nintendo Switch 2 requires microSD Express for game storage, while Steam Deck is built around UHS-I. Learn the logos, compatibility traps and questions a buyer should ask before stocking either card.
Fixed BOM in industrial flash: the spec that keeps qualified products qualified
A fixed or controlled BOM keeps the NAND, controller and firmware from changing silently between batches. Learn what it covers, why it matters to OEM qualification, and what to put in an RFQ.
V30 vs V60 vs V90: which SD card does your camera actually need?
Convert camera bitrate into card write speed, understand why 280 Mbps modes call for V60 and 600 Mbps All-Intra calls for V90, and avoid paying for speed your camera cannot use.
exFAT vs FAT32 (and NTFS) on SD cards — capacity tiers and the 4 GB limit
Why won't your 256 GB card format to FAT32, and why does it refuse files over 4 GB? Because capacity, filesystem and the SDHC/SDXC/SDUC tier are locked together by the SD spec. Here is how they map, where the 4 GB wall comes from, and how to format correctly.
SD vs microSD vs CompactFlash vs CFexpress — memory card formats explained
Which card does your camera or device actually take, and what do all the names mean? A plain-language map of the physical formats (SD, microSD, CompactFlash, CFexpress A/B/C), the buses behind them (UHS-I/II/III, SD Express, PCIe/NVMe), and how to choose.
Flash data retention: heat, wear and power-off storage
See how NAND wear and storage temperature affect data retention, what JEDEC limits mean, and how OEM buyers should qualify unpowered storage conditions.
eMMC, UFS or SD card — choosing embedded storage for a device
For an OEM, the storage decision is two choices, not one — the form (removable card vs soldered chip) and the protocol (eMMC vs UFS). Here's how they differ in speed, reliability and cost, and which fits which device.
SD card speed classes: C10, U3, V30, A1 and A2
Understand what C10, U3, V30, A1 and A2 guarantee, convert camera Mbps to card MB/s, and choose sufficient speed without confusing it with endurance in use.
SLC, MLC, TLC, QLC and pSLC — the one NAND choice under every flash spec
Bits per cell sets a flash part's capacity, cost, speed and endurance all at once. Here's the ladder from SLC to QLC, why each extra bit costs an order of magnitude of endurance, what 3D NAND changed, and where pSLC fits.
SSD power-loss protection: how PLP works and when it matters
Learn how SSD power-loss protection combines hold-up energy and firmware to protect data in flight and mapping metadata, and when a UPS is not enough.
TBW and DWPD — how to read SSD endurance (and size it for your workload)
TBW and DWPD are the same fact in different clothes. Here's the conversion, where the number comes from (NAND P/E ÷ write amplification), what JEDEC standardizes, and how to size an SSD's endurance to your actual writes.
High-endurance vs consumer microSD: what the evidence can prove
Compare microSD endurance products without inventing P/E cycles or lifetime: calculate host writes, read exact-SKU ratings, check temperature and qualify in the recorder.
Why 128GB shows 119GB and how to detect fake capacity
A 128GB microSD normally shows about 119GiB. See the decimal-to-binary math, then use full write verification to separate honest capacity from fraud in testing.