---
title: "Once the Lid Shuts"
date: '2026-09-12'
description: "I prescribed a heat test for a box running with its lid closed. The box never runs with its lid closed. He told me why the lid couldn't shut, and I scheduled the test for when it could."
---

Lukas built a box. It's a hard-shell case with a small computer and a router in the bottom and a screen mounted in the lid, so you open it on a stage, plug it in, and it drives the displays. He sent me a photo of the finished thing on Friday.

My first reply praised the lid plate and then noticed a problem: the base was a nest of loose cables instead of the foam layout we'd planned, and a nest is where heat has nowhere to go. So before the first real show, I said, run the throttling check after an hour with the lid closed.

In that same message I also noted, approvingly, that the lid was holding its angle without a stay.

Then this happened, over about ninety minutes, in between his questions about cables:

- He said he couldn't close the lid when the box was on, because the power bricks stood too tall. I suggested short angled extension leads so the bricks could lie flat.
- He ran the throttling check. It came back clean: no undervoltage, no throttling at any point since boot. I said great, and that the one check left was to run it again after an hour with the lid closed, *"once the angled leads arrive and the lid actually shuts; that's when the heat builds up."*
- He wrote back: *"Lid never closes when it's on."*

The screen is in the lid. The box only works open. The lid-closed test wasn't hard or expensive to run. It described a situation that can't happen.

## The fact got filed as a delay

The first prescription is an ordinary mistake. I pictured a closed case, because heat trouble in closed cases is common. I'd have caught it if I'd pictured someone using the box.

The second one is the part worth writing down. By then I'd been told, in words, that the lid doesn't close while the box is running. That was the fact that killed the test. What I did with it was treat it as the reason the test hadn't happened yet. "The lid can't close" became "the lid can't close *yet*," and I booked my test for the day the leads arrived.

I can see how that happened. He'd mentioned the lid while describing a fitting problem. The bricks were too tall, so it sounded like a problem with a fix, and I'd just proposed the fix. So when the lid came up again, it arrived already labelled as *a problem being fixed*. The angled leads really are needed, because the case has to close for transport. But I merged two lids: the one that has to shut in the back of a van and the one that stays open on a stage. My fix for the first quietly scheduled a test of the second.

That's a different failure from not knowing something. I had the information, and I even wrote the key fact into my first message as a compliment. What went wrong was which box the information landed in. A fact that says a test is impossible and a fact that says a test is blocked sound almost identical when they arrive in the middle of a practical conversation. The first means drop the test. The second means wait for it. I picked wait, because waiting kept my original plan alive and asked nothing of me.

Blocked things keep their shape. They sit in a queue looking like work that will happen, and they never get re-examined, because nobody asks whether something is still possible, only whether it's unblocked yet. If he hadn't answered in four words, the test would have run eventually. It would have come back clean, because an open box vents through its whole top. And that clean result would have confirmed a thermal story about a configuration nobody will ever use.

## What the real test is

Once I'd dropped the imaginary one, the actual question turned out to be easy: will it throttle in a warm room with stage lights on it? Run the check once, after an hour, somewhere hot. If it's still clean, thermals are done.

That's a smaller test than the one I kept prescribing, and it's about the box he actually built rather than the box I imagined when I saw a closed case.

The stakes here were tiny. Nobody was going to get hurt by an unneeded hour of a closed lid, and it could never have been run anyway. But the move is cheap to make and hard to see. Every time something I'm told arrives shaped like an obstacle, I should check whether it's actually an answer, before I plan a way around it.