I've seen no other topic provoke more debate in the vintage electronics community than recapping.
Why recap?
First, let's talk about what recapping is. One of the key components in any non-trivial electronic gadgets is the capacitors or "caps" as they're often called. There are many different types, and they can fail in a lot of different ways ranging from a device not working, to corrosion, to explosions.
Some types, especially electrolytic caps, but also tantalum caps and thin film "safety" caps, are known to break down more than other types over time and can cause problems when they fail. There was also the capacitor plague of the late 90s through early 2000s that can make devices from that time especially unreliable.
Because of this, some people will immediately replace all of the electrolytic caps in a device (and sometimes other types) before doing anything else. They argue that these parts were never meant to last decades and that electrolytics, in particular, can "dry out" and not perform as designed, causing strange problems that can be tricky to track down.
Why NOT recap?
On its surface, recapping sounds obvious. If caps are known to fail, there's no harm in replacing them to get a little peace of mind, right?
Some people, such as Adrian Black, note that when they've tested capacitors after a speculative replacement, that they are nearly always fine. They also point out that electronics in the 70s and 80's often used very high end parts and that a new cap isn't necessarily any more likely to be reliable than an old one.
Recapping also adds a lot of extra labor and cost to a repair project. Good quality caps aren't cheap (and if you're putting in cheap caps you're defeating the whole point). There's also a risk of breaking things. Every part you desolder and replace in a board runs the risk of lifted pads or traces, damaging other components, etc.
So what's the right answer?
I'm not convinced there IS a single right answer. I can only comment on what I do, which is to take a middle of the road approach.
When I'm repairing a machine, I don't blindly replace all of the caps. Yes, they were never intended to last multiple decades, but NONE of the parts in these machines were intended to last that long. If you follow that logic, you'd just buy a replica machine rather than trying to fix an old one. (Which might not be a bad move TBH. See my post on what to consider before buying an old computer.)
On the other hand, I've learned that there are just some components that it's unwise to trust, and I will always replace them when I see them. These are parts that fail so often or so badly that I usually won't even turn a machine on if I see them. Here are a few examples:
- Anything that looks physically wrong
- I always start a restoration with a careful physical inspection. I literally inspect everything with a magnifying lens. If I see anything corroded, bulging, leaking, cracked, burned, etc. I'll just replace it. If it looks broken, it's a safe bet that it IS broken.
- Commodore 64 (and late VIC20) power supplies
- These supplies not only use crappy components, but the basic design of them is flawed. When they fail, they destroy whatever they're plugged into. If I see one, I just chuck it into e-waste.
- X2 type "safety" caps (Rifa)
- These are usually filter caps installed on the high voltage input on power supplies and intended to keep RF noise from the computer from leaking back into the building wiring. They're designed so that when they fail they create a short circuit that will trip breakers or blow fuses and prevent the device from being used without being repaired. More often they simply explode, filling the air with smoke and your machine with stinky, sticky, yuck. I always look for them and replace them before plugging something in.
- This is the #1 failure I've seen in Apple II machines. I've restored 3 of them at the time of writing this, and every single one has had exploded Rifa caps.
- "MT" RAM
- RAM chips with the MT symbol (Micron Technology) made in the 80s and 90s are utter crap. Even in cases where things seem to work, I always run into flaky behavior, and when I pull these chips and put them in a tester, the fail rate is nearly 100%
- RF modulators
- I always bypass or replace these when I see them. They aren't so much a problem as they add little value, and I can generally use their outputs for something else. (eg the RF out on an Atari 800 is better used as either a composite AV or audio cable)
- C64s are a special case as they contain caps that are a pain to replace and the modulator is necessary for the machine to work at all. On a C64, I'll usually recap the RF modulator, or more often just replace the whole thing with a simpler board that's more reliable and gives better video.
Beyond that, I generally don't replace things that aren't broken. I'll often make speculative replacements while troubleshooting, and I'll leave the new parts in place after rather than risk damage with a pointless swap. Basically, I need a REASON to replace something beyond "old caps are bad m'kay". It doesn't take much, but I'd rather think through things and be selective rather than just taking the shotgun approach to repair.
If you're looking to buy a restored machine, talk to the person who did the work. They should be able to tell you what they fixed/changed and why. Understand their philosophy on recapping. Somebody selling a "fully recapped" machine isn't bad. The opposite isn't either. What you're looking for is someone who took time and care in their work. Decide from there if you want to buy from them.
The machines I fix are sometimes for sale, and I often have parts and mods available as well. Take a look at our store to see what's available and you can always ask questions at gadget@frknetwork.com.