Introducing WeldLab. ⚙️
We believe technology should be for everyone.
Professional engineering knowledge and advanced tools shouldn't be locked behind expensive software, complicated interfaces, or exclusive access.
We started WeldLab with a simple idea: make technology more accessible, practical, and easier to understand and use.
Our goal is to build tools that help people learn, create, solve real problems, and do better work — regardless of the size of their business or the resources they have.
For us, engineering isn't just about complex calculations. It's about making those calculations useful. It's about turning knowledge into tools that people can actually use.
We believe progress comes from sharing knowledge, learning from one another, and growing together.
A better industry isn't built by a few people working alone. It's built when more people have the opportunity to contribute.
That's the spirit we want to bring to welding and engineering.
We are starting with welding simulation, engineering tools, and quality management — but our vision goes beyond any single product.
Technology for Everyone. Engineering Without Barriers.
Welcome to WeldLab. 🔥
Rough day on the sim:
1. 17 steps into the t8/5 module, the environment restarted and the process died. Went through the disk and most of it survived, only two connections were missing. Wired those up myself, self-check passed 6/6, and t8/5 measured 2.672 s.
2. Two self-check items came back FAIL even though they printed exactly what I expected (3.000000 and 2.000000). That one's on me. I'd set the tolerance to 1e-9 s, but the data passes through a float record once, which adds about 2.4e-7 s of error. Loosened it to 1e-5 and left the expected values alone.
3. The preheat job: the source file is 158,833 bytes, but the runner cut off at around 145k characters, couldn't see the rest, and stopped. Checked everything afterward and nothing was corrupted.
4. Stage 5 pre-check failed. I used dt = 0.001, but the stability limit is 0.000342, so I was 2.92x over. There's also no way to set up a 20x20x14 small plate because the spacing is hardcoded at 2 mm. On top of that, the graded long run would produce about 19.7 GB of output, which doesn't fit in 16 GB of RAM 😭
Schooled by my welding sim. Pass 2's arc point spiked to 1948 K, sank to a 485 K plateau, and never crossed the 150 °C interpass line — 120 s, gate stayed shut. Reran it with predictions locked in first: gate opened at 11.5 s, pass 2 lit and blew past the 3000 K limit. We only count numbers we ran