|
General Vintage Technology Discussions For general discussions about vintage radio and other vintage electronics etc. |
|
Thread Tools |
1st Sep 2022, 9:26 am | #41 |
Dekatron
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
|
Re: How would we design electronics to last 100 years?
I'd like to know that too!
I have designed equipment that's worked 30 years without service, so approaching a third of the target time. I'd attribute this to: * Using components of established reliability - plastic film capacitors, conservatively-rated electrolytics, metal-film resistors (I routinely use Welwyn MFR4, MFR5 types, and vitreous wirewound for higher power). * Avoiding open-type switches eg the make-it-yourself wafer switch assemblies and open relays * Keeping temperature rises low (noting that adding a fan helps massively but itself adds a life-limited item) * Laying out circuit boards avoiding closely spaced tracks, and narrow conductor widths * Vacuum impregnating coils and transformers * Building using leaded solder * Wiring with high-quality wire (not PVC!), usually silicone rubber or Raychem/Tyco Type 44 Etc! (NB I have to declare an interest in the Welwyn resistors as I work for a company which, like Welwyn, is now part if the TT group. Other brands are available...) |
1st Sep 2022, 9:30 am | #42 |
Dekatron
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
|
Re: How would we design electronics to last 100 years?
In fact the founder Dan D'Agostino of the high end audio electronics manufacturer Krell took the company name from Forbidden Planet.
Craig
__________________
Doomed for a certain term to walk the night |
1st Sep 2022, 10:08 am | #43 | |
Dekatron
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
|
Re: How would we design electronics to last 100 years?
Quote:
The starships need to be repaired on a not infrequent basis. So even at this speculated technology level, they still need to be fixed. I only mention that because Reynolds has a Physics PhD, and worked for ESA in Nordweik in space systems design. After his first novel was a success, he quit ESA and became a full-time author. Of course approaching light speed is a strange and wonderful thing. If you can excuse the robotic female voice, and the somewhat grainy quality, this is mesmeric https://apod.nasa.gov/apod/ap111018.html Craig
__________________
Doomed for a certain term to walk the night Last edited by Craig Sawyers; 1st Sep 2022 at 10:18 am. |
|
1st Sep 2022, 10:43 am | #44 |
Hexode
Join Date: Jul 2004
Location: Paris, France
Posts: 346
|
Re: How would we design electronics to last 100 years?
Hi to all,
Franco-German cultural channel ARTE aired two documentaries on the subject : what would it take to go to Alpha-Centauri and what technologies would be required to do this? Every technology aspect was replied to by scientists actually working in each specialized field. The solution was described as thus: Entirely automated mission, no humans on board Vehicle (a long metal girder structure), 2km long, assembled in orbit Initial impulse by standard chemical rocket propulsion, then Electro-ionic propulsion, accelerating for 25 years, reaching 15% of light speed Gas cylinders for ionic propulsion, cast off when depleted Electric power from portable Fusion drive reactors, cast off after use (vehicle gets lighter) Central AI control, multiple redundancies, automated diagnostic of failed electronics & auto-replacement. 25 years of deceleration, insertion into orbit around Alpha-Centauri planet Shuttle vehicle drops to ground, fleet of intelligent drones including under water ones. Drones feed back data to craft in orbit, relay to earth Total mission duration 50 years + 1 to explore + 5 for data to travel back to earth. Mission planned as possible circa 2100 to 2156 considering technology advances necessary to make it possible. ARTE has a phone/tablet/computer app for free viewing. The 2 programmes are not currently available but are re-aired usually at yearly intervals. May be available on DVD, like many of their productions. Best Regards jhalphen Paris/France Last edited by jhalphen; 1st Sep 2022 at 10:51 am. |
1st Sep 2022, 11:13 am | #45 | |
Dekatron
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
|
Re: How would we design electronics to last 100 years?
With the greatest respect, even though the answers have been interesting, the OP asked:
Quote:
Self-repairing assemblies, and time-dilation by speedy travel, are 'cheating' as to the spirit of the title, so Off Topic, surely? I attempted to answer in Post #41 |
|
1st Sep 2022, 11:32 am | #46 |
Nonode
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,288
|
Re: How would we design electronics to last 100 years?
I am quite happy that the thread has expanded to cover space travel, even using "cheats" like time dilation you have to accelerate and decelerate so 100 years is still a reasonable design goal.
Post 41 certainly sets out some excellent guide lines, but is it enough? Peter |
1st Sep 2022, 11:50 am | #47 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
|
Re: How would we design electronics to last 100 years?
Thinking about the non-repairability of integrated circuits, I am sure I read about someone in the early 1970s who had designed a mechanism to include some extra gates in an ULA or FPGA which could be brought into use if there were failures of the main gates, to essentially 'patch round ' the failure.
Might even have been Clive Sinclair who came up with the idea.
__________________
I'm the Operator of my Pocket Calculator. -Kraftwerk. |
1st Sep 2022, 11:51 am | #48 |
Moderator
Join Date: Jun 2003
Location: Oxford, UK
Posts: 27,787
|
Re: How would we design electronics to last 100 years?
Interstellar missions are effectively impossible with current technologies and understandings of physics, even if unmanned. The timescales are so great and the returns so small and speculative that the economics simply don't work.
Such missions require some sort of FTL technology, which may well never exist (that's certainly what Einstein thought). |
1st Sep 2022, 12:00 pm | #49 |
Banned
Join Date: Jun 2015
Location: Mareeba, North Queensland, Australia
Posts: 2,704
|
Re: How would we design electronics to last 100 years?
TRy a book called " Rendezvous with Rama " Arther C Clark, 1973. Organic/mechanical spaceship that comes and visits then leaves. Completely rebuilds itself by "digesting " broken bits to make new bits.
Took my fancy when I was much younger. It still fascinates my imagination. Clark was one of the masters. Joe |
1st Sep 2022, 12:38 pm | #50 | |
Dekatron
Join Date: Oct 2003
Location: Southport Lancashire, UK.
Posts: 3,221
|
Re: How would we design electronics to last 100 years?
Quote:
The item itself may last for a long time but technology would probably overtake it and you can't build in anything that can outlast that.
__________________
Paul |
|
1st Sep 2022, 12:45 pm | #51 | ||
Dekatron
Join Date: Aug 2006
Location: Madrid, Spain / Wirral, UK
Posts: 7,484
|
Re: How would we design electronics to last 100 years?
Quote:
At the same time, gear left standing unused can also suffer, though that seems more true of electromechanical items. I am often surprised when plugging in TVs in storage for 10+ years and they work first time.
__________________
Regards, Ben. |
||
1st Sep 2022, 5:58 pm | #52 | |
Nonode
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,288
|
Re: How would we design electronics to last 100 years?
Quote:
Peter |
|
1st Sep 2022, 6:00 pm | #53 |
Dekatron
Join Date: Aug 2016
Location: Colchester, Essex, UK.
Posts: 4,081
|
Re: How would we design electronics to last 100 years?
I like the idea of voting logic, as a progression of the double redundancy/triple redundancy etc. safeguards...the gremlins still, on occasion, seem to find a way through all the trapdoors put in their way....i guess it's a function of a world where a huge number of processes are occurring every day. As Arthur said, coincidences must happen, because of the laws of chance...not despite them. (or something along these lines)
Dave |
1st Sep 2022, 6:12 pm | #54 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
|
Re: How would we design electronics to last 100 years?
The "voting logic" or having multiple sources of the same data to guard against the failure of one source is all very well, but if all the sources-of-data are made by the same supplier and develop the same failure, you have 'issues'.
This very problem - with the Rubidium frequency-standards made by a Swiss company - really screwed-up the 'Galileo' satellite-navigation project half a decade back. Each satellite had several time-standards, bunches of which failed, rendering several satellites 'unreliable'. The same failure affected Indian satellites. https://spacenews.com/rash-of-galile...ming-launches/ https://www.firstpost.com/tech/news-...n-3697087.html Source the components of your 'redundant' system from different suppliers!
__________________
I'm the Operator of my Pocket Calculator. -Kraftwerk. |
1st Sep 2022, 6:49 pm | #55 |
Nonode
Join Date: Jan 2016
Location: Stafford, Staffs. UK.
Posts: 2,529
|
Re: How would we design electronics to last 100 years?
Part of the problem is, that as you fix one point of failure you simply find some other failure mode that you had never thought of. Even if you think of a cause of failure, and test against it, you might not test in a particular scenario.
Example, I know of a couple of issues with circuits. The failures were seen once during testing (actually testing of updates on another part of the whole system), but were impossible to replicate (until I got involved). Both were triggered by a brownout on the supply. The second was only seen during multiple attempts to replicate the first. The first was easier to investigate, as once the failure occurred, the device was essentially dead even after a power off reset. So the supplier was able to investigate and determined the system parameters stored in an area of flash was corrupt. Reflashing the device restored operation. The second was much harder as a power off reset restored operation. In both cases the timing of the brownout was critical. The devices were designed to have essentially two modes, running and sleep. The brownout testing had been carried out in both modes, and always passed. Experience suggested to me that the problem was if the brownout occurred during the transition between states, as the device might well be performing housekeeping during that time. The first was easier to reproduce, and once I'd narrowed down the timing, I could reliably 'kill' it within a couple of hors of repeated testing. The second was trickier and wasn't houskeeping as such, but a similar issue with the combined power supply managment and bus communication controller locking up if hit with a brownout at the critical time during mode transition. But I still found a way to repeat it, and send the supplier of that device a script to use with test equiment so they could reproduce and fix it. Unfortunately both problems had made it into the field but failures were rare and extremely rare, but were soon resolved. So with potential failures like that, how could you be sure your system will last 100 years? |
1st Sep 2022, 7:11 pm | #56 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
|
Re: How would we design electronics to last 100 years?
There's also the issue that the world as you designed your device for moves on over that 100 years.
I'm reminded of the UK's introduction of radio; initially a broadcast-receiver only really needed to receive one station - because that was all there was. So simple was elegant/cheap. Then the BBC came along and Mr. Eckersley implemented his 'Regional plan' which meant that loads of people then found their early, simple/cheap receivers had to have selectivity to separate-out the regional and national stations. A single signal-frequency tuned-circuit was no longer good-enough. [This issue was rather more pressing in the urban US, where their more-liberal approach led to a proliferation of stations and selectivity/resistance-to-swamping really pushed the move from TRF to superhet radios]. Designing for a 100-year lifespan, all very well, but your 100-year-old design can easily be overtaken by reality. Anyone want to buy a Horse?
__________________
I'm the Operator of my Pocket Calculator. -Kraftwerk. |
1st Sep 2022, 8:32 pm | #57 | |
Dekatron
Join Date: Jan 2012
Location: Brentwood, Essex, UK.
Posts: 5,316
|
Re: How would we design electronics to last 100 years?
Quote:
By the time the application reached me, it had been successively assigned to Ternland and then to Anarmartic ( possibly it was the other way round). I remember it well because the necessary copies of the assignment documents had not been filed and I had to ask the agent for copies of them. At that time, IBM and others were proposing matrix memory arrays incorporating redundant rows and/or columns that could be substituted for defective ones. I have no idea if any of these proposals made it into production. |
|
1st Sep 2022, 11:31 pm | #58 | ||
Dekatron
Join Date: Dec 2007
Location: Haarlem, Netherlands
Posts: 4,185
|
Re: How would we design electronics to last 100 years?
Quote:
It's clear that valves (having heaters) can notably fail (burn out) upon switching on, and can also work themselves loose by thermal cycling. So I very much agree that's not a myth and there would have been very good reasons (likely more than only those) to leave those computers powered on permanently. If only one thing, maintenance will become much more predictable (for example swapping all valves every 20.000 hours instead of one every 5 minutes, which would render the computer completely useless). The myth in my eyes, is that tubes (or rather modern computers, or really any piece of equipment) hardly wear when they're powered on continuously. I don't think you even claimed that about valves, but it's claimed way too often for my comfort by small time systems administrators, etc. who once turned off some ancient piece of equipment that wouldn't come on after that anymore. So yes, I meant that the mindless propagation of a true story turned it into somewhat of a myth. Last edited by Maarten; 1st Sep 2022 at 11:54 pm. |
||
2nd Sep 2022, 10:05 pm | #59 |
Dekatron
Join Date: Jan 2012
Location: Brentwood, Essex, UK.
Posts: 5,316
|
Re: How would we design electronics to last 100 years?
According to this advertisment from the October 1945 "Wireless World", the standard Mazda SP41 valve was found to be serviceable after 11,000 hours continuous service, hence its use in a subsea repeater for the GPO.
Last edited by emeritus; 2nd Sep 2022 at 10:14 pm. Reason: typos, better quality scan substituted. |
2nd Sep 2022, 10:49 pm | #60 | |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
|
Re: How would we design electronics to last 100 years?
Quote:
David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
|