Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
Following
BTC How the Bitcoin mnemonic generator works
An open source hardware project called "Entropy32 Plus" was recently unveiled. Its core innovation is to use the physical randomness generated by radioactive decay to generate a Bitcoin seed phrase. The device runs completely offline, does not have any wireless connectivity, and does not store seed data, ensuring security at the physical level.
The core concept of the project is based on the unpredictability of the physical world. Traditional software random number generators are inherently deterministic, that is, they must produce the same output under the same input. In contrast, the interval between radioactive decay events is determined by quantum mechanics, is truly random and cannot be predicted in advance. This physical entropy source provides higher security for key generation.
In terms of hardware implementation, the device is connected to a Geiger counter through a 3.5 mm audio interface to detect decay pulses from radioactive sources or background radiation. The LM393 comparator then converts the raw voltage signal into a clear high or low reading, using a bias point fixed at around 0.5V to ensure consistency. Firmware captures the time interval between consecutive pulses and enters this data into the SHA-256 conditional processing stage to eliminate statistical bias and "bleach" the raw entropy data. Finally, the processed entropy is mapped to the standard 2048-word BIP39 English vocabulary to generate mnemonic words.
The entire process is performed by a custom PCB based on the ATmega328P microcontroller, and is operated locally with a small OLED display and button drive interface. Complete seed phrases can be generated or viewed without additional software other than on-board firmware. The developers emphasize that the device is simple, auditable and can be built by anyone with common components.
Hardware and design limitations
Due to the memory capacity of low-power microcontrollers, the project made many engineering trade-offs. The compiled firmware takes up 30,006 bytes of the chip's 30,720 bytes of flash memory, leaving less than 1,000 bytes for future code updates.
The BIP39 vocabulary itself occupies 13,117 bytes of space, accounting for approximately 43% of the total program storage. The SHA-256 compression routine adds 2,278 bytes, while the state machine and menu logic consume approximately 3,000 bytes. Given the huge space occupied by the vocabulary, regular display libraries cannot accommodate it at the same time. Therefore, the project failed to use the widely popular Adafruit_GFX and Adafruit_SSD1306 libraries, but instead switched to the U8x8 library's text mode to communicate with the OLED screen to save memory resources.
This open source repository contains the complete Arduino code, SHA-256 implementation, KiCad schematics, as well as fabrication files and 3D printed shell designs for manufacturing equipment. However, the developers pointed out that the build was experimental in nature and had not been independently audited. The document clearly reminds that users should verify the quality of the entropy source by themselves, otherwise they will have to bear their own risks. As of now, its entropy quality has not been verified by the NIST SP800 - 90B standard method.
