Field Notes · IIoT · Industrial controls
Dispatches from the intersection of industrial control systems, IIoT architecture, and biolab experiments. These are the notes I wish existed when I was debugging SCADA at 2 am — written close to the metal, after the hardware answers back.
Field Notes from a Bioreactor — read the series in order →
-
The TV That Would Not Stay on the IoT VLAN: A Lesson in Network Enforcement
My Samsung Frame TV kept wandering off the IoT VLAN onto the Default network — a perfect example of why SSID-based isolation is a fragile illusion. Here’s how I diagnosed the firmware-driven roaming behavior and why real network…
-
The NAS Wasn’t Broken, Just Slow: A QNAP Timeout Debug
Three weeks of a cryptic TypeError turned out to be a 5-second timeout fighting an 8-second NAS. Here’s how a library’s hardcoded timeout masqueraded as broken code, and the simple reset that fixed it.
-
Field Notes from a Bioreactor — Entry 12: The Culture Is Dead
Twenty-four consecutive nights of NO_FLASH. Eighteen days at the OD scatter floor. The culture is dead — and the hardest lesson is that we should have known sooner. Entry 12 confronts the null result, falsifies the heat hypothesis,…
-
I Gave My Home AI a Radio and Let It Explore
A cheap USB radio receiver on the house AI server, a deliberately dumb service that lets the agent use it without a shell, one night of sweeps from 24 MHz to 1.7 GHz, the spectrum turned into sound,…
-
Field Notes from a Bioreactor — Entry 11: Ten Days at the OD Floor
Ten days of silence at the detection floor. The optics have never been cleaner, the baseline has never been lower — and the vial has never been emptier. The bulk-flash test has been rehearsed four times and executed…
-
What Old Hardware Is Still Good For: Eleven Real Jobs for Two 2016 GPUs
Two used datacenter GPUs from 2016, with none of the features modern AI assumes, now search a family photo archive by description, label who said what in a recording, split songs into stems, clone a voice, enlarge images…
-
Field Notes from a Bioreactor — Entry 10: The Cumulative Heat
The blob is gone. The instrument has never been cleaner. But the culture isn’t growing — and the reason reveals something deeper about our garage biolab than I expected. The cumulative heat problem, the two gates, and why…
-
The Blob That Wouldn’t Stop — and the Routine Swap That Fixed It
For weeks, a mysterious green intruder crept across the glow-capture camera frame, threatening to contaminate every bioluminescence measurement. The bench trip was scheduled. The protocol was written. Then a routine culture swap on October 2 wiped it clean…
-
Field Notes from a Bioreactor — Entry 9: The Acceleration
The Pyrocystis culture is in its best measured state — OD has accelerated 3.5-fold in 17 days. But what is really growing? A field-notes entry on non-axenic cultures, honest metrics, and the gap between optical density and microscopy.
-
The Floor Is a Cliff — Calibrating the Glow Detector
The glow detector’s sensitivity cliff is exactly one 8-bit DAC step wide. At level 7, the LED pulse is invisible. At level 8, the sensor saturates. There is nothing in between. Sixteen nights of NO_FLASH are redefined: not…
New posts are infrequent — when something worth writing happens in the lab or the rack room.