Passing It On

For years I watched retro repair videos mostly as entertainment. It would be a long time before I actually tried to fix anything. When I finally did, I realized that the mechanics of it were all new, but the knowledge about troubleshooting and figuring out how something worked stayed with me. Things clicked in a surprising way. The reason is thanks to creators who produce extremely high quality content like Adrian’s Digital Basement, Action Retro, This Does Not Compute, TronicsFix, and Nanofix. YouTube repair videos, crafted by competent hands, can be equal parts entertaining and educational.

If you have recapped late-80s electronics, you know the problem. Surface mount electrolytics sit on pads that lift if you look at them wrong, and I had watched people with far more bench time than me fight them with hot air and desoldering tweezers, prying at them, tearing pads off as they went.

Adrian’s method is to press down on the cap and twist. Do not pull up. It shears off, and on a corroded one the legs give way first and leave stubs you can rock off with light finger pressure. Fresh solder and a bit of flux on the stubs, wipe with the iron, clean the pads, done. He says he has never lifted a pad since he started doing it this way, even on boards in terrible shape. My experience is the same, the only pad I ever lifted was when I tried other techniques.

First time I tried it the cap came off, the board was fine, and I sat there thinking, oh, that is how you do it. That was the first time at the bench I felt like I knew what I was doing instead of hoping.

Here is my favorite part. It is not his trick. He credits Mr Carlson’s Lab. So the chain was already three deep before I got near it. Somebody worked it out, Adrian passed it on, I picked it up from Adrian, and now I am telling you. That is the beauty of this community of people, all investing their time and experience to save the electronics that, for many of us, defined our childhood.

Knowledge is power

I have been building up a knowledge base for almost a year now. It contains pinouts, schematic details, copies of manuals and pictures I’ve collected from the internet. How I collect the bulk of it is a post for another time. Stored as plain text, indexed and carefully organized, an AI agent can read this text and use it to help me fix things. It isn’t training - it is agent memory, loaded on demand into the context.

I have watched most of the videos on Adrian’s basement. I wish I had discovered the channel earlier, but I have been an ardent viewer for many years now. The thing about creators like Adrian is, his knowledge is buried in hundreds of hours of footage and I can never find it again when I actually need it. Whether it is me, or an agent, it takes time to go search a video by keywords, find the right video, and then find the part where he talks about the subject at hand. So I pulled the auto-generated captions off every video with yt-dlp, 899 of them, and had an AI read them and write down the technique, with the video and the timecode attached to every entry. Nothing but the caption text, no video downloaded, no reuploads, nothing published. It is a private index on my own machine. Now I, and my AI agent, can know these facts too! More importantly, every entry carries a link and a timecode, so when a technique comes up I can point people at the source.

What have I learned?

I’ve learned a lot of basic stuff from Adrian - how to use a scope to observe a clock, how to use a thermal camera to find a short from the heat. Most of these techniques are all over and if I hadn’t learned them from Adrian, I would have learned them from someone else. I like to think that Adrian’s style is particularly well suited for learning though, he always explains his thought process, why he does what he does, and when possible where he learned it. I model my own documents at Console Artisan the same way.

Here are a few of the more interesting observations I was not going to stumble on:

A leaky cap can pass an LCR meter. A cap bridging a high voltage bias rail measured fine in circuit and was breaking down badly under real operating voltage. LCR meters test at a couple of volts, so a part rated for hundreds can fail in a way that measurement cannot reach. You want a leakage test at rated voltage. Watch it.

Counterfeit chips, thirty seconds, no gear. Cotton swab, 99% isopropyl, rub the markings. On rebadged parts the printing lifts and the swab comes back black. Plenty of genuine period chips were ink marked too, so it isn’t perfect, but it gives you a strong signal. Watch it.

Both were sitting in videos I had already seen. That is Adrian handing them to you now, same as he handed me the cap trick.

Thank you

Adrian, thank you. I have learned an enormous amount from you, and you are a big part of why I went from watching this stuff to doing it. The community around your channel is one of the good ones and I am glad to be in it.

I support him on Patreon and you should too.