💾 Archived View for envs.net › ~kkernig › gemlog › 2020-12-17-electronic_repairs.gmi captured on 2021-12-05 at 23:47:19. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2021-12-04)
-=-=-=-=-=-=-
Repairing things is one of the best ways to reduce your consumption and environmental impact. You also learn about the technology you use, which has practical advantages. Additionally, it might give you a sense of trust and understanding of your environment, because you stop seeing devices as magic black boxes with buttons.
Since my early teen years I have been tinkering with electronics. Over time, I have accumulated skills, practical knowledge and many spare parts, and moved to more advanced repairs. However, I think that I'm still far away from being able to do serious repair work on modern (micro)electronic equipment. This frustrates me quite a bit, since I've learned to loathe throwing away electronics.
But there is no way around realizing that nearly all modern electronic devices are of frightening complexity, and integrated to the hilt. Tiny SMD components, mulilayer PCBs and digital communication between components all make repair work nearly impossible, unless you have specialized equipment and deep knowledge of the system. Just imagine a Raspberry Pi, where a randomly chosen component is faulty. How would you locate it? Do you have a replacement part? Can you do the delicate SMD soldering? Could you think of a reason for the failure and estimate the durability of your repair?
Fortunately, Raspberry Pis are known to rarely break in normal use, which is representative of most other modern electronics. Along with complexity, durability has also gone up. This and the reduced production costs are the main reasons that electronic repairs are seen as a niche service today, worth the effort only for obvious faults on rather expensive devices. How good of a long term solution is this?
For me, the gold standard of production quality and repairability is laboratory test equipment made roughly between the mid 1960s and mid 80s. It is usually very well built in terms of mechanical design and layout, and for many devices there are service manuals online. These typically include a description and explanation of the circuitry, PCB layout(s), part lists and the entire schematic diagram(s). Did I mention the detailed description of calibration routines and troubleshooting information to locate faults? It is a pleasure to repair these devices, and probably the largest obstacle for a sucessful repair is the availability of replacement parts. Some ICs and specialized transistors/thyristors/etc. are simply "Unobtainium" (CuriousMarc quote), but fortunately modern replacements exist for the majority of components.
Current example: I got hold of a laboratory power supply. Made in Germany in late 1977. Internally it's very straight-forward and solid: a big transformer, bridge rectifier, big capacitors and a linear regulator, with a OP-amp based feedback loop for precision and stability. It's fairly compact for its size and power rating, and very solid mechanically. All wires are beautifully laid out, though disassembly takes some time due to the slight over-engineered design (there are probably more than 50 screws and nuts holding everything in place). During somewhat careless testing, a potentiometer malfunction caused the power transistor in the linear regulator to die shorted, which took the rectifier diodes with it. Also, the power switch had contact issues. Luckily, the faults were in the "traditional" part of the circuit, so I managed to locate them even without documentation (a glitch somewhere in the feedback/ regulation circuit would have been trickier to find). I then replaced the broken parts with similar ones: a 2N3055 replacing the obscure PD1515 NPN TO-3 transistor, two 1N4004 for each unmarked diode, and a nice chunky power switch from my spare part box replacing the old one.
Now the thing lives again. How long? I don't know. Maybe some of the capacitors will die at some point. The front panel potentiometers still have some contact issues. But I do know that even then, there is a good chance of repairing these issues.
The question that naturally arises for me is whether the qualities of old lab electronics can be somehow applied to modern devices. Ideally, such a device would combine the benefits of both - durability, low power consumption and powerful hardware one one hand, excellent documentation, durable mechanical construction and easy replacement of components on the other hand. Would it be worth the effort? Would it become any more than a niche market? How many people would be able to perform in-depth repairs even if the device is specifically built to be easy to work on?
In the physics lab where I was doing my B.Sc and M.Sc., I noticed that manuals for newer equipment contain less and less information, up to the point where nothing beyond a plain user manual is available. And very few people besides me cared.
While I believe that such super-durable and repair friendly devices could be made (technically), I highly doubt that they will sell in anyhow significant quantities, because the majority of people simply doesn't care. And even if people would care, there need to be enough with the knowledge and skills for such repairs. In my experience, people who have these skills are usually of the older generation, often retired engineers or ham radio enthusiasts. In my generation (1990s-born), *very* few perform electronic repairs. I think one of the main reasons is the complexity of the devices in daily use. If you are surrounded by devices which even experts struggle to understand, how could you find a starting point? Why should you even begin to think of the possibility of making repairs or modifications yourself? The spiral of technical advance further consolidates the mentality of "throw away and replace".
Maybe my feelings here are unjustified, but I feel a great unease knowing that the vast majority of the people have absolutely no sense of understanding for the basics of electronics. Was it better 50 years ago? Or did it just matter less back then? We spend more and more of our time with computers, and rely on them for anything from EC/debit/credit cards to aircraft and nuclear reactor control. There may be quite fundamental psychological effects at work here, regarding our trust in these man-made, yet intangible machines.