Braiins has pushed a new firmware cycle focused less on headline hashrate and more on what keeps an ASIC fleet alive through restarts, curtailments, cold starts and field troubleshooting. Braiins OS 26.09 arrived September 16 with new power-supply temperature visibility, searchable logs and startup controls, followed by version 26.09.1 on September 24 to correct an intermittent temperature-sensor fault that could temporarily halt mining.

According to the release notes, supported Antminer S19 and S21 machines can now log temperatures from APW11, APW12 and APW17 power supplies. Operators also gain a configurable startup cooldown limit, GUI log search and safer preheat behavior designed to keep chip frequencies within safe values while power settles toward its target.

PSU temperature becomes a first-class mining metric

A miner can have healthy hashboards while its power supply is moving toward a thermal failure. Exposing PSU temperatures in logs gives operators another signal to compare with hashrate, chip temperature and wall power. That matters as S21-generation machines replace older S19 fleets and power density rises.

The change also complements the economics behind joules-per-terahash and daily electricity cost. A firmware setting that keeps a miner online at the intended power target can matter operationally even when the release itself does not change the ASIC’s silicon efficiency.

One field test shared publicly by Red Panda Mining provides useful context without establishing a universal result. On an Antminer S21 XP, the tester reported roughly 268–270 TH/s at about 3,900 W on stock firmware, then 270+ TH/s at roughly 3,740 W after Braiins OS autotuning. Hardware condition, ambient temperature, power supply, firmware settings and measurement method can all change those results from machine to machine.

Hydro miners get a specific cold-start fix

Braiins OS 26.09 also addresses an I2C: no ACK power-supply communication error after cold boots or reboots on hydro miners using Zynq control boards. It fixes cold-start power settling below the requested target and improves preheat behavior.

Those details are increasingly relevant as industrial sites deploy dense hydro-cooled mining containers and manufacturers push machines such as BITMAIN’s S23 XP Hydro. Liquid cooling removes the high-speed miner fans found in conventional air-cooled machines, but it does not eliminate PSU, control-board, sensor or startup-sequence failures.

Version 26.09.1 shows why fleet updates need staging

Eight days after 26.09, Braiins released 26.09.1. The company documented a T5 temperature-sensor issue affecting some upgraded miners that could freeze mining for 60 seconds. The patch targets S19 and S21 series machines.

For a farm operator, that sequence is a reminder to stage firmware deployments. A small test group can reveal control-board, PSU, sensor or tuning behavior before a release is pushed across hundreds or thousands of machines. The same operational discipline applies to competing firmware ecosystems, including the recent LuxOS S21++ firmware update.

Curtailment makes restart behavior more important

Modern mining sites increasingly treat ASICs as flexible electrical loads. When miners curtail for grid conditions, power prices or site thermal limits, thousands of machines may stop and restart together. A configurable cooldown limit and better startup diagnostics can therefore affect how quickly a fleet returns to stable hashing after an event.

That operational layer becomes more valuable when operators are trying to extract more compute from fixed electrical infrastructure, as seen in PowerCompute’s recent same-power miner refresh. Firmware cannot manufacture new megawatts, but it can influence how effectively installed ASICs use the power already available.

What miners should measure after updating

The useful post-update numbers are sustained accepted hashrate, wall power, J/TH, PSU temperature, chip temperature, rejected shares, restart time, tuning duration and error frequency. Operators should compare those measurements against a pre-update baseline rather than assuming a firmware release automatically increases profitability.

That is the important part of Braiins OS 26.09: it adds observability around the components and startup states that can quietly cost uptime. The release is not a new ASIC. It is an attempt to make the existing machine easier to diagnose and more predictable when real-world operating conditions interrupt ideal steady-state hashing.

Sources

• Braiins Academy: Braiins OS 26.09 and 26.09.1 release notes
• ASIC.tools: independent technical analysis of the release
• Red Panda Mining: S21 XP wall-power and hashrate test

BitcoinVersus.Tech Editor’s Note

BitcoinVersus.Tech reports on Bitcoin, mining hardware, data centers, energy, semiconductors and emerging technology. Support independent reporting with Bitcoin: 3C9o19EH5HSiwEPyCTmEKzxhNCbo2X6TTb.

Financial disclaimer: This article is for informational and educational purposes only and is not financial, investment, legal or tax advice. Bitcoin, cryptocurrency mining and publicly traded securities can be volatile. Readers should conduct their own research before making financial decisions.