Timeline for Design for long term reliability?
Current License: CC BY-SA 3.0
24 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 22, 2015 at 14:54 | vote | accept | Olin Lathrop | ||
| Feb 6, 2015 at 4:13 | answer | added | Guill | timeline score: 1 | |
| Feb 5, 2015 at 2:54 | comment | added | HKOB | 1. Solid-state ICs are typically much more reliable because they have few exposed parts (see 4.). 2. Whatever the component - you need testing to be sure. 3. Do what you can to reduce temperature&power. 4. Cover the design in goo, making it unserviceable and offer replacement boards. | |
| Feb 4, 2015 at 20:22 | comment | added | iggy | Let us continue this discussion in chat. | |
| Feb 4, 2015 at 20:11 | comment | added | iggy | What about connectors? | |
| Feb 4, 2015 at 0:02 | history | tweeted | twitter.com/#!/StackElectronix/status/562763106167029762 | ||
| Feb 3, 2015 at 23:09 | comment | added | Olin Lathrop | @iggy: The board will need a few caps. From what I've investigated so far, I'll use 0805 ceramics with decent voltage rating margin. That seems to be what I'm hearing is best for reliability. If you know something else, please tell us. | |
| Feb 3, 2015 at 22:03 | comment | added | iggy | Will you be using capacitors? | |
| Feb 3, 2015 at 21:42 | answer | added | iggy | timeline score: 5 | |
| Feb 3, 2015 at 21:01 | comment | added | Olin Lathrop | @iggy: No rubber or clay that I know of. This will be a regular FR-4 board in a chassis. | |
| Feb 3, 2015 at 20:59 | comment | added | iggy | Definetly use lead soldering. There are quite a lot of papers detailing tin whiskers @NASA :) Any chance of rubber (sulfur) or clay being present somewhere nearby? | |
| Feb 3, 2015 at 19:23 | history | edited | Olin Lathrop | CC BY-SA 3.0 |
added 706 characters in body
|
| Feb 3, 2015 at 18:53 | comment | added | Nick Alexeev | @Olin Can you tell about the mechanical environment in which your circuit will be operating? Will there be thermal cycling? Will there be shock and vibration? Ingress or condensation? | |
| Feb 3, 2015 at 18:06 | comment | added | user16324 | "Availability of parts" suggests that repair is a possibility (not undersea or in space) and that changes the requirements. You can reflow the solder joints every 40 years or so, rather than worrying about their eventual failure for example. Or consider the life of PCB pads vs plated holes under re-soldering, when choosing DIP vs SMD. Some interesting background reading in "Voices across the Sea" about transatlantic phone cables (by A.C.Clarke - yes, that one!) - they chose vacuum tubes over transistors, because they had seen tubes operating for 20 years non-stop, for example. | |
| Feb 3, 2015 at 18:04 | comment | added | Oleg Mazurov | Why "being able to get replacements in the future is important"? If your customer is planning to regularly inspect the circuit then I can understand why do they want TH. | |
| Feb 3, 2015 at 17:24 | history | edited | JYelton | CC BY-SA 3.0 |
Minor typos/spelling
|
| Feb 3, 2015 at 17:16 | history | edited | Olin Lathrop | CC BY-SA 3.0 |
added 608 characters in body
|
| Feb 3, 2015 at 17:09 | comment | added | George Herold | Can you say a bit more about the environment? Thermal cycling, Temperature extremes, vibration, exposure to elements, (Rain, Sun, salt water.) @PhilFrost, I only skimmed your link, it looks like all computer modeling, did they do any "real" testing? I heard a rumor that through hole might be better in thermal cycling, 'cause the leads would take up some of the strain.. (but this was for transistors.) | |
| Feb 3, 2015 at 17:03 | answer | added | Oleg Mazurov | timeline score: 11 | |
| Feb 3, 2015 at 16:17 | comment | added | Some Hardware Guy | You looking for predictive studies/methods like Belcore/Telcordia SR-332 and MIL-HDBK-217? Or are you looking more for actual research studies conducted over 10-20 years. I guess this goes beyond the usual calculate MTBF, and do HALT testing approach. | |
| Feb 3, 2015 at 16:05 | comment | added | PlasmaHH | This is a really intresting question, but it seems to be really broad. There are so many things that you might want to think about, e.g. tin whiskers (esp. with rohs). A lot will be pure speculation (e.g. we might completely run off spintronics in 50 years). I would really like to see some answers though, but probably this will be a big list, and many answers giving just some ideas and hints. Maybe one community wiki answer that people edit their points, with optionally links to papers would be a good format? | |
| Feb 3, 2015 at 16:05 | comment | added | Edward | I can't answer your full question, but there was a flurry of research in the late 1990s and early 2000s that looked at the impact of moving to lead-free solder and the conclusion was, surprisingly, that lead-free solder was actually more reliable for typical parts on typical boards. | |
| Feb 3, 2015 at 16:04 | comment | added | Phil Frost | on the point of thru-hole vs SMT: wpi.edu/Pubs/E-project/Available/E-project-042513-011426/… (executive summary: SMT is more robust against thermal cycling, vibration, etc due to smaller size) | |
| Feb 3, 2015 at 15:52 | history | asked | Olin Lathrop | CC BY-SA 3.0 |