If cycling power doesn't work...
Try dropping it to reseat the connectors. Never mind I'm off...
The Kepler space telescope is back in action after a tense couple of weeks for NASA engineers. On April 8 the space agency reported that Kepler was stuck in emergency mode, which not only made stellar observations impossible but also burns through the telescope's limited fuel reserves at a prodigious rate. In a couple of days …
As opposed to the Acorn Atom where the rule was take the back off and push all the chips back in - Atom was single board with keyboard on one side and all chips hanging from sockets underneath and could slowly work loose as you typed. The Atom also came as a "Build-yourself" kit and one of the computer mags at the time printed a set of helpful tips for people doing this which included "if you are finding it hard to push the chips into sockets try loosening the socket first with a pin" .... sadly some people rather overdid this - had no problems putting chips in sockets but when they finished and turned the board over they'd all fall out!
Try dropping it to reseat the connectors
Regularly used that same trick to unpark the heads on an external 20(?)MB HDD hooked to the store Tandy 2000 POS machine at Radio Shack back in the day. Worked a treat every time. Word subsequently got out to the other stores, and I eventually received a call from corporate to "stop recommending that," along with a firmware patch to fix it.
"Regularly used that same trick to unpark the heads on an external 20(?)MB HDD "
Picking it up and a jerky twist on the platter axle axis usually achieved the same result without risking unseating the heads from their carriers (which makes for an interesting sound as they bounce off the spinning platters on their wire tethers)
The real solution to drives which had sticktion problems (other than replacing them) was to never turn them off.
<quote>As opposed to the Acorn Atom where the rule was take the back off and push all the chips back in </quote>
That sounds like the EXACT same problem my TRS-80 Model I suffered from. RAM chips slowly rising from the sockets. (TBH, I had also seen that same quirk in some broadcast telemetry equipment. The mfgrs 'solution' was to ty-wrap the chip in place.) I ended up solving that problem at the same time that 64K DRAM chips became available, yank out the old sockets, and solder the new chips in. Good bye trouble prone sockets.
"Atom was single board with keyboard on one side and all chips hanging from sockets underneath and could slowly work loose as you typed."
The second generation of their colour encoder that turned the signals from the 6847 into PAL relied on you removing the 6847 from the mother board, plugging it into the colour encoder, and then plugging the colour encoder board (with the weight of the PCB itself, a 40pin dip package, various 16pin packages, all socketed, and discrete components) so it was hanging upside down and held in by the friction of a 40 pin dip socket - not sure how that was ever supposed to not fall out!
Probably the biggest advance Acorn made between the Atom and the BBC was moving all the chips to the top of the PCB!
There were two main types of socket for DIL packages, one was considered the quality option - turned pin and they tended to hang on to their component. There there was another with flat pins and a sprung (be being bent) contact, these ones could work a component out by flexing of the PCB regardless of which way up the PCB was, even against gravity.
I was recruited by a US company to work on PC-clone products in the early 80s. Its the first time I came across 'designing from the price list'. Early computers had ridiculous markups but that didn't stop the people running the companies squeezing every last cent out of the design. Needless to say, natural selection favored those who made sure the product had a reasonable life expectancy.
Incidentally, hot glue works well to hold stuff like ICs in place.
> Try dropping it to reseat the connectors. That was actually a common solution with the Commodore 64, drop it flat on the table from say 25 cm up.
And 30 years on, that's a a well-known technique for bringing an iPad back from the dead.
We have a large service running on about 20 Wintel boxes (a mix of MS databases and .NET apps). I think they are still on Server 2003 (planned to be refreshed at some point in the next year or two (but then they've been saying that for a while now!)).
They all get rebooted automatically every Sunday afternoon (Service is only used Mon to Sat).
If they don't get a reboot, it gets slower and slower, and eventually hangs.
It was deemed by management to be more cost effective to just set up the scheduled re-boot, than to do the root analysis to find out what the real issue is (the assumption is a memory leak somewhere). :-/
"The pointing tables and science targets – instructions that tell the spacecraft where to look and at what – were reloaded and confirmed, onboard logs and counters were reset, and a new command sequence was created, tested and uploaded to account for the late start of the campaign."
That is not a power cycle it's more like pushing policy in checkpoint to unstuck things.
"Kepler's rebirth is yet another example of why NASA has some of the best engineering hackers"
I hate to think what the Fleet Street write-up would be if you tell them hacking involved. An industry that thinks guessing the default PIN still used on voicemail counts as "hacking phones"? That effectively says 'hackers' always equals 'bad' but argues vehemently that the journos paying corrupt officials were just a few bad apples?
Ungraceful shutdown (or reboot) may be useful on some rare occasions. If a process writes incorrect data to the nonvolatile storage during a graceful shutdown, it makes sense to skip that part. Also, boot sequences may force additional checks when they detect traces of an unclean shutdown.
/speculation.jpg/
Sorry. Hoover Constellation. Best vac I ever used. Not the best at cleaning, but the most fun.
More likely some chip turned out not to be hardened enough against cosmic radiation - a high energy particle (which do not happen on Earth, thanks to magnetic field) may easily flip memory state or do similar harm. Electronics sent out there are always shielded against those, but as we have learnt, some particles have very large energy indeed, even if they only hit anything very, very rarely.
Yeah watched a program last night where some NASA guy was talking about working out the best way to get thrust from the Titan Mare explorer. He said how normal propellers would make the whole space thing a bit easier. Like firing a one ton submarine designed to sail in a sea of methane, into space and hitting a ocean on a moon several planets away ain't no big thing.
Anyway..
1 Seriously clever guys.
2. Who says science and engineering is boring?