Are retrocomputers best left on or off? | Hacker News Reader

Are retrocomputers best left on or off?

retrocomputing.stackexchange.com

66 pointsazeemba76 comments

My two cents:

If we are talking about 1970s and '80s machines, they tend to run pretty hot, related to the PSU technology. Things like electrolytic capacitors "don't like this" and may contribute to cascading failures. PSU failures are apt to fry some chips, some of which may be hard to come by. So, better keep them cool = off.

If you consider machines from the 2000s retro, continuous run, avoiding the stresses caused by system start, may help keeping them alive. (I've a MacPro running since 2008, with no failures, apart from failing 3rd party ECC DDR3 RAM. Some of this RAM has even failed twice, the RAM originally shipped by Apple is still fine, though. This machine has only been ever off, when I've been on vacation.) – Machines from the 1990s and early 2000s are pretty much the same, but typically suffer from poor capacitors and/or batteries. So…?

It’s almost as if the question is vague on purpose to stir up responses!

It’s almost as if nobody has actually read the question! The OP says in the first sentence that the machine is from 1998, and that they’re going to use it about an hour per day.

Still, the attraction, it got the attention for, may be the more general question.

Something I didn't see mentioned but a consideration: what's the quality of your electricity supply? I live in a location with a lot of summer thunderstorms and even with a surge protector and lighting arrestors I'm nervous about having stuff plugged in unnecessarily during storms, having seen devices destroyed by it.

Power them from a pure sine wave inverter connected to a battery thats being charged. Aka a good UPS

That's not good enough - everything attached to the computer (including network gear, unless it's fiber) must be powered by the same UPS, or you have pathways for lightning current to enter.

Some UPS for consumers even have an ethernet through-port so you can connect your internet cable from the wall to the UPS before it touches your router.

Ethernet has galvanic isolation.

That's functional isolation. Yeah, they're individually tested to 1500 Vrms (or so the factory in china claims), but that insulation is still just the very thin lacquer of the windings and maybe lacquer on the core. There are more expensive (and physically much larger) Ethernet transformers for medical devices and such, these have actual double/reinforced insulation and are tested to much higher voltages.

Typically the Ethernet shield is also only alibi-insulated from the device's ground. Sometimes not at all. (Personally it's always funny to me to see the fat 1kV/1nF capacitor from the shield to the device ground and then the metal of the socket is just bunched up with basically no clearance against the metal case or something like that). This can of course be avoided by simply using unshielded cable.

Doesn't a (plugged-in) UPS have similar isolation?

A online UPS has in line: a rectifier with its transformer and output-smoothing capacitors, a huuuge battery, surge capacitors, power transistors, another transformer and more capacitors. And only then comes the actual device with its PSU.

It takes an awful lot of power to smash through these multiple layers of insulation and bypass all the capacitors, at least if the PSU is properly designed (specifically, clearance between the various power / ground domains).

In contrast, Ethernet transformers are tiny small little things.

Yes, but only the transformers will provide the galvanic isolation. And if the windings are not physically separated then you are basically in the same situation.

Most UPS that you’d plug into a socket aren’t online/double conversion though, they’re line-interactive.

That may be a factor, excellent point. Electric supply is pretty good and there hasn't been a major outage, since I was a child. Also, power lines are buried (under ground), here. (But I have surge protection, just in case, and am running all related equipment from the same protected outlet.) Which also means, I have no idea.

Apple will always have the best RAM. They get to pick from the donut of wafer, and have testing processes rivaling that of the fabs.

I've read that electrolytic caps dry out faster if they aren't holding a charge. Temperature is a concern, but I have a surprising number of devices that worked fine before I left them unplugged for a few years and then didn't work when I powered them back up.

There's also a small amount of wear simply from the drag of electrons running through the fine filament threads of wire they are piped through. The longer this goes on the more electron tunneling will happen, degrading the CPUs performance.

We don't notice them on larger wires as the wear is infinitesimal, but in a CPU where the wire diameters are in the micron scale and below there is likely a maximum lifetime of usage before the whole system becomes inoperable.

Might be 500 years of continual usage, might be 50, I don't know and I don't think anyone has done a long-term study on how long a CPU can be continuously operated before it fails, but the risk is there.

Especially with older disk drives if you don't power them up for a long time (months) the bearings may seize, at which point you may have to use the hair dryer trick to try getting them to work again.

I have a 42 year old apple //e that had been in an uninsulated attic in Boston for 30+ years. It booted fine and all the 5 1/4" floppies worked. I know that C64s need the caps replaced and the org external power supplies can catch fire. I'd be really surprised if 1990's pcs failed. There was a period of transition to non-lead solder where many pcbs from that era suffered shorts from whiskering but everything else should be fine. Electromigration tolerances in the 90's cpu families was even more strict than today.

1990's PC's have Varta bios batteries soldered on that leak and destroy the motherboard. Or they have Dallas real time clocks with dead batteries.

I'm staring at a PC-AT mobo from 1988 with an AMD-branded 286 chip that also says Copyright Intel (lol), and the battery has leaked a white film.

My understanding is that every electrolytic cap will eventually dry out. The capacitance drifts as this happens, so the quality of the cap and the sensitivity of the circuit to incorrect capacitance will greatly affect how long it takes for the device to fail. The late 90s was also when a large quantity of low-quality electrolytic caps were manufactured.

I recently turned on a PC some ten years old which was in storage for five years. First time it turned on, but had difficulties booting due to the M.2 SSD from Intel being corrupt. Second time, it failed to turn on (but some LEDs on the MB indicated that there was still some power), fans not spinning. Third time, it did start again, but the fans at full speed and fan control not working anymore. It seems the lubricant of the fan's bearing dried up and made them hard to spin to the point that the in-rush current exceeded the capacity of the fan-control chip.

Fortunately, those beige^black boxes hold no sentimental value whatsoever and will be replaced with something better at a convenient time. Before turning on something more rare, I'd inspect the MB visually for corrosion, bloated caps and test the fans and disk drives externally.

It’s kind of funny to me how the different manufacturing and supply chain factors have caused these reliability “sweet spots” and “death spots”.

Like you mention, the leaded to lead-free solder transition was one of them. I agree about 90s PCs quite likely being reliable. But then… 1999 to 2007 we had the capacitor plague: https://en.m.wikipedia.org/wiki/Capacitor_plague

And around 2002 there was the Great Fujitsu Hard Drive Recall: https://www.theregister.com/2002/10/31/fujitsu_faces_lawsuit... If I’m remembering that one, they changed out a flame retardant additive in their chip packaging? Something like that? Those drives were approaching a 100% field failure rate and you could mail them back to Fujitsu and get a cheque for $100.

I switch most of mine (70-80s) on a few times a year and of them I use a lot (weekend game dev); I have to replace capacitors quite often, but outside that they all work. Unlike more recent (2000+) pc stuff which is all dead. All Sun and SGI stuff still works well though.

All my old Sun hardware has failed nvram chips. I replaced and restored the one in my 3/80 but can't be bothered with the rest. I haven't turned on the SGI Indigo in many years, so who knows.

I've got an S-100 bus AlphaMicro with bad ram that would probably take a ton of work to get going again.

if you never turn the computer on, why does it matter if it's hypothetically still functional? boot it occasionally, the only thing that keeps a machine alive is someone caring for it

"I intend to use it for about 1 hour per day."

>why does it matter

Because you aren't a cat whose owner is called Schroedinger.

Am I, or am I not? Who could tell?

That's not the question though, it's whether it's better to leave it on or turn it off between uses.

I'm not an expert, but I'm convinced the answer is "off". The reason is "on" accelerates growth of whiskers.

https://en.wikipedia.org/wiki/Whisker_%28metallurgy%29

No matter what you do it will all need servicing.

Capacitors are perishable. Electrolytics dry up, solid (MLCC, polymer, tantalum, paper) dont like humidity (some even require baking before soldering). Mechanical stress and temperature cycling leads to cracking of caps and solder joints.

Mechanical parts will have stiction (hdd heads), dried grease. Plastic parts expand due to moisture (sprockets/cogs in FDD/CD), rubber bands liquify or dry and stiffen.

Then you have vented helium/lost vacuum.

Early 1990s dec machines: 3100 m series and 4000- 60 90 90a 705

If you have a modern scsi2sd card solution as your drive, off is fine. Just make sure all fans can spin prior to boot.

If you have a classic SCSI or expansion box with a big SCSI, on. Same situation if you have functioning dssi drives. Let them spin.

Don't have to read/write or full boot just get to Chevron prompts and do a show dev to get it all spinning.

What’s the point of a retro machine from ‘98? Any application from that era should run natively on a modern machine. It’s past the point of games locked to a certain MHz or special purpose video drivers. If you want a CRT, just get a CRT. What’s the appeal?

Theoretically, but practically a lot of late 90s (and early 00s) games don’t run on modern windows.

they probably run well in wine though

This seems like the most difficult era as far as emulation of PC games goes. There was a ton of change and different 3D accelerators during this time and hardware advanced away from things like 16-bit color and dropped support. And it is old news, look up "thief crappy colors" so see all the work that had to be done to get Thief: The Dark Project (and games using the same engine) to look normal on modern hardware. And that was 10-15 years ago. And many games didn't have the fan base to update game like Thief did.

Because people enjoy using them

applications yes, but operating systems less so.

i am a fan of NeXTstep. i can see myself wanting to run that on an original machine because all alternatives would not provide the same experience.

next time i visit my mother i have to try and check if the NeXTstation stored there still boots up.

Why does there have to be a point?

Old components are still (and will be) manufactured, since so many old tech depends on it, like refrigerators, air conditioners, washing machines, microwave ovens, ...

Some old components will be, sure, but there are some chips that aren't manufactured anymore. There's no one blanket rule.

That's true, but in many cases, it's not the integrated circuits (ICs) that fail, because they tend to be quite reliable. It's more often the capacitors, resistors, storage devices and power supply components that degrade over time. Capacitors, in particular, can fail due to heat, aging, or voltage surges, which can lead to failure in devices even if the ICs are still in good condition. Power sources can also fail due to wear on these exact components.

> It's more often the capacitors

And the electrolyte can and will start to leak and foul the ICs around them, if not by ruining the traces/pads, then they can also foul the legs and/or find their way into the chip package itself, so the ICs will fail as a secondary issue.

So better to turn it on sometimes, without power cycling so much, to keep the electrolyte fluids wet, to extend the equipment life. Let it rot and dry and it will.

That's not how it works.

Electrolytics which have not had voltage applied in a long time (decades) may need gradual reforming, but if they aren't hermetically sealed they will dry out no matter what --- and heat will accelerate that process.

I've seen enough electronics repair videos that it seems the default for troubleshooting absent other conspicuous damage is to start checking the electrolytic capacitors.

https://en.wikipedia.org/wiki/Capacitor_plague

There was a time when Gigabyte started selling more robust motherboards with quality capacitors that would last much longer. After seeing so many old motherboards dying from blown capacitors I was very happy to get one. Those motherboards so far held the full length of their meaningfull life until the CPU just got too slow to use for anything. Still try to look for that type of capacitors when I build new stuff for friends and family.

The Gigabyte line is called "Ultra Durable" and have solid capacitors instead of the electrolyte ones.

I doubt it would be impossible to replace a Pentium 2 machine in most of the western world for at least the next 20 years.

I worked at a company that decided to do a new design because the chips for the old one had been out of production for decades and a solid percentage of what was still available on eBay was sitting in the warehouse to be scavenged for new production.

Pentium 2s aren't the only retro computers out there. There's many different computers, especially in the 1980s, that have components that are just simply not made anymore -- break one of those components, and you'll have a hard time repairing / replacing them.

yes

Honestly emulators work better than the real thing 99% of the time, so I don't feel the need to play with flaky and dusty 30yr old hardware

Yeah, but they don't emulate the smells, the sound of fans spinning up and the gentle breeze they generate, the beeps and boops, the tick tick tick hard drives make, the long travel of keys on that old keyboard with a slightly sticky shift (gotta be from that coffee I spilled in '97).

Game speed too.

Simh does.

I miss the scraping hard drive noises with a flickering “HDD” light and a mysterious “Turbo” indicator when my 90s PC had to work hard

Turbo was all about slowing down a 486-/Pentium-class machine to work with programs and some ISA cards that required a slower CPU and/or XT bus speeds.

wait... so turbo made it slower?

De-pressing turbo made the computer slower. Pressing it made the computer run at its normal speed.

Or that click of death that plagued zip drives.

My Dell had one in college. Everyone thought it was cool until the click-CLICK click-CLICK. Dell replaced it twice. Then I gave up. When did I ever really need to transfer 100MB anyway?

The click of death was transferrable via a disk having been damaged by a drive with the c-o-d, that's why 'replacing the drive' wasn't really a solution, you'd need to also throw all your zip disks and buy new.

Basically, it was a hardware virus.

Yes, but old computers also do not do that, because they no longer work. (And if they do, it's not with some 30 year old hard drive.) Maybe 80% of the working computers I have put on the shelf do not even boot after a few years. It's a fair question, I think. Maybe electronics just lose their spirit.

FS-UAE emulates even the floppy/HD head sounds and scrapping the media too.

These were all very nice, and reminiscent of the time

Then you remember you need to install Windows 9x or 2000 on it and the CD-ROM reader is kinda flaky, and you barely can download anything with IE 6, so now you would need an TLS bridge or use a usb-drive. Oh yeah and don't forget you need a small partition formatted on FAT32

Call me when you can emulate that 30 year old solder smell, among other things.

I once made some rosin for soldering by dissolving the stuff made from tree sap for stringed instruments in isopropyl alcohol, and I think this is pretty close to the smell you're after. Fairly certain it's still not ideal to inhale the fumes from it, but it must be at least slightly less harmful than additionally vaporizing some heavy metals.

(This works reasonably well as flux for through hole sized components, possibly smaller if you're more skilled than I am, but you do have to clean up any excess afterwards because it's very sticky.)

Lead does not evaporate at the temperatures used for soldering.

Is that true, I have always heard that a tiny bit of lead is in the fumes. Either way I still think that one should avoid breathing soldering fumes as a general rule, no matter how pleasant they may smell.

>Either way I still think that one should avoid breathing soldering fumes as a general rule, no matter how pleasant they may smell.

Yes, but due to the flux. So lead-free solder fumes are not really safer (some say they're worse). The main danger with leaded solder is getting lead on your hands, which could get into your food if you aren't careful.

Real hardware is still necessary to verify functional equivalency and correctness.

Also, there's no way you're going to emulate an interlaced or noninterlaced analog SVGA display, especially the sounds of mode changing, tearing, proper vsync, palette cycling, or setting the border color. Or MIDI music as it came from an AdLib, Roland, SB, or a GUS.

Some machines do not have emulators. Notably, most SGI workstations have custom hardware that's poorly documented and very complex.

As far as I'm aware of, there's also no functional Ultrasparc emulators -not that can emulate a late 90's/early 2000's Ultrasparc. The last time I looked at Qemu it wasn't quite there ("there" meaning able to run an arbitrary Solaris the way the original hardware could).

Reply on news.ycombinator.com