Question electronics

[classic87]

Why do some electronic devices from the 60s, 70s and 80s still survive today? Is it just a matter of luck?
They were exposed to decades of oxygen

Is there a big difference between corrosion by oxygen and humidity and corrosion by sea spray? And I don't live near the beach.

 

[stepleton]

A question for the veterans: just how often have you encountered tin whiskers or problems that seem attributable to tin whiskers? Particularly in pre-RoHS electronics?

Online sources agree with what's here: leaded solders are far less likely to whisker...

But talking of a "fault that appears absurd or inexplicable or counter intuitive based on a circuit's design", the first hurdle we encountered in reviving our PERQ 2 (1985) was a mysterious short on two adjacent address lines for the fast SRAMs that make up the computer's 20-bit microcode-addressable registers. There are five such chips, and believing that the fault was internal to one of them, we started gently clipping the leg for address line 8 on one SRAM IC after another until we reached and cleared the short. Having isolated the fault, we solder-bridged each clipped leg so that we could turn the machine back on and confirm that we'd caused no new damage.

To our surprise, instead of seeing the original fault, the machine merrily charged through its CPU tests and tried to boot! The short did not come back, in other words. We don't know how we fixed it, but we joked that there must have been a whisker between two of the legs on the broken IC that we missed during visual inspection. By this account, the act of soldering the leg back together melted the whisker. I'm still not convinced of this story, but we have no real explanation, and the CPU registers have been rock solid ever since.

A similar phantom problem has frustrated my attempts to fix a problem with the PERQ's disk controller: the problem greatly attenuated itself (and these days basically seems gone) after I soldered flying leads to an IC that I was suspicious of so that I could probe it. We have no idea what's going on unless some of these things like dopant migration can heal themselves with use somehow. (And no, it's not electrolytic caps reforming themselves this time ) So we joke that the tin whiskers have struck again.

The PERQ is a complicated machine with loads of solder joints and hundreds of 74Sxx ICs pushed right up to their frequency limits. Ours lived in a dank cellar for years to boot. But two episodes of "tin whiskers" still feels a bit rich to me. I don't think ICL was using unleaded solder in 1985, either. So, to bolster or erode my scepticism: how often have folks here encountered whisker-y problems on old electronics?

 

[Hugo Holden]

The PERQ is a complicated machine with loads of solder joints and hundreds of 74Sxx ICs pushed right up to their frequency limits. Ours lived in a dank cellar for years to boot. But two episodes of "tin whiskers" still feels a bit rich to me. I don't think ICL was using unleaded solder in 1985, either. So, to bolster or erode my scepticism: how often have folks here encountered whisker-y problems on old electronics?

You can get plenty of tin whiskers without solder being involved at all and nothing to do with lead free solders. All you need is Tin plating which can be on the legs of IC's, transformer brackets, many tin plated surfaces. Tin was/is very common to protect thin steel used in electronics. I have a great photo somewhere of a transformer with a Tin plated bracket, that came out of storage and it was covered in millions of long Tin whiskers, so many the photo of it looked out of focus.
The whiskers that affected my 2465B scope came from Tin plated pcb tracks devoid of conformal coating. The ones that killed many Germanium transistors of the AF11x type made by Mullard & Philips (That NASA did the research paper on) it was derived from the tin plated housings that encapsulated the transistor crystal.They can grow between Tin plated pins where there is no solder about.

 

[whartung]

Is there a big difference between corrosion by oxygen and humidity and corrosion by sea spray? And I don't live near the beach.

If you live near the beach, everything, and I mean EVERYTHING, rusts, or corrodes. The salt in the air eats everything.

Maintenance and such can keep it at bay, but there's a reason why everything near the sea is rusty, while everything near the lake it just faded. There's a reason they repaint the Golden Gate Bridge every year (it's always being painted).

Keeping things at room temperature, out of the sun, low humidity, and things will last -- but that doesn't mean entropy will not creep in and do it's thing anyway.

 

[classic87]

i not live near the beach corrode is same i electronic devices storage?

 

[johnx993]

Living on the Gulf coast, it gets humid.
Years ago, I invested in dehumidifiers for each room with VC's/electronics.
The air conditioning alone gets the humidity down to 53%.
I run the dehumidifier machines to get it down to 45%.

And I turn it off in the electronics lab a day before I do some bench work.
This to let the humidity rise to reduce static issues. Just to be safe.

-J

 

[classic87]

Does turning on electronic devices once a year help prevent failures?

Do all components fail equally in long-term storage?

 

[johnx993]

I turn mine on once a quarter. I devote an entire weekend to startups, documentation upkeep, logging any issues that need work before the next cycle, etc.

 

[johnx993]

There are good engineering statistics out there on failure rates and modes.
But generally for computers, it's mechanical, corrosion, capacitors.

 

[johnx993]

And understand corrosion that can affect performance may not even be visible with the unaided eye.
A very thin coating can block a connection.

 

[classic87]

I don't live near the beach, even though I use a dehumidifier in my room, the humidity varies from 59-70%, if the dehumidifier dies it remains at 70%

What is the maximum interval for energizing electronic devices and remove humidity and preventing oxidation and corrosion?

 

[johnx993]

Well, I've heard I don't do it often enough (every 3 months). But I have a large collection and can't devote lots of time to it.
Realistically, I'd say once a month should be fine.

 

[johnx993]

As far as humidity - I think somewhere in the 40's should do.
I kept it in the 50's for a while then noticed no serial output on one of the Altairs during quarterly start up.
Visual inspection of the molex connectors showed no visible corrosion.
However, after a single wipe with deoxit dabbed swab, it worked.
My guess is a thin layer of corrosion - not enough to be visible, but enough to interrupt the connection.

 

[johnx993]

OTOH, there is the philosophy that old computers are like old cars.
Expect some 'tinkering-time' every now and then to keep them running.
I guess that applies to all things really.

 

[cjs]

...the problem greatly attenuated itself (and these days basically seems gone) after I soldered flying leads to an IC that I was suspicious of so that I could probe it.

Did you desolder the leads after? Adding random wires that do nothing in order to fix things is a time-honoured tradition. :-)

BilHerd OP • 8y ago​

​

Third response: There is a whole story behind the wire on the top of the C128 PCB. Its redundant to the trace (A10) that’s there except that due to the layout there would be a reflection when the Z80 was driving the address bus. 30% of the units wouldn’t complete booting CPM, they would hang.​

​

The back story is that I and my boss were fighting, and he literally took me off the project for the week and gave the problem to someone else to fix. I spent the week sleeping, taking showers and having sex. On Friday he walked in and basically told me to fix it or I was fired. I lucked out and found the problem in under two hours, I had found that if I stared at the scope trace for a minute and then looked quickly up at the white wall I could see a reverse of the signal due to retinal burn and there I could see a glitch right at a critical time (It was a standing wave actually). I remember Frank Palia giggling when I showed him this trick as he admitted he couldn’t see the glitch on the scope but he could staring at the wall (Astronomers do this also as its easier to see dark on light for humans). Since no one really could believe that a redundant wire fixed this, let alone that I found it so quickly given the hundreds of signals on the board, we did a test run of 10,000 units over the weekend which would be a production run quantity for any place except someone as big as CBM All of the 10k units worked, we didn’t have a single CPM boot failure which is crazy that such a simple fix really was needed and really worked. We didn’t relay out the board as we might have just moved the problem somewhere else, so this was a matter of the devil that you know, so we added the wire to the board assembly instructions.​

 

[alank2]

Things used to also be built better. They were expected to last a long time so they were built so. They were also much more repairable.

 

[cjs]

Things used to also be built better. They were expected to last a long time so they were built so.

Really? I do look forward to your explanation of how Sinclair and C64 built better than, say, modern laptops.

 

[JDW]

Why do some electronic devices that were working stop working if they remain unpowered for long periods?

I'm not surprised by the broad sweeping replies subsequent to the opening post in light of the fact "long periods" was not defined. Even so, I was pleased to see you defined it on page 2 of this thread, as follows:

Why do some electronic devices from the 60s, 70s and 80s still survive today?

You then went on to other questions like these...

Does turning on electronic devices once a year help prevent failures?
Do all components fail equally in long-term storage?

What is the maximum interval for energizing electronic devices and remove humidity and preventing oxidation and corrosion?

I think the questions themselves illustrate an excess amount of concern akin to asking, "If I walk as if I'm on rice paper, will all be well?" And the answer to even that question is no.

As others have stated, fluid-filled capacitors fail over time, but you will most often get more than 20 years of life out of them, except for hardware made during the "capacitor plague" years.

You cannot easily answer a broad sweeping question like why SOME devices from the 60s through the 80s "survive" today. You probably mean "still work today" but even that is not easy to answer and involves a lot of speculation.

Turning on electronic devices once a year won't really hurt or help.

No, there is never a case when "ALL" (the word you used) components "FAIL EQUALLY" in long-term storage. But it goes without saying "nothing lasts forever." Everything slowly degrades over time. That is why the keyword of this thread should be "MAINTENANCE." If you see capacitors leaking, change them. If you power on an old device, try to find schematics, get an oscilloscope and power supply and start troubleshooting in cooperation with others on specific points you suspect to be a problem.

Just as your car will last a long time with regular maintenance (oil changes, etc.), so is true of your old electronics. Just take them one machine at a time. And when you eventually get stuck, ask in a very specific way in forums like this and others. That's what I do.

 

[stepleton]

Did you desolder the leads after? Adding random wires that do nothing in order to fix things is a time-honoured tradition.

I did. It took about a month to work up the courage. The machine still worked.

This is the same computer that I was asking about in the tin whiskers thread: it seems to have weird problems now and then.

The disk was playing up a few days ago, but that was probably due to the hot weather. When things cooled down, it was back to normal.

 

[johnx993]

Really? I do look forward to your explanation of how Sinclair and C64 built better than, say, modern laptops.

That 'time' thing with quality was valid for a short time window. What it really means is that business took a critical look at producing cost-effective (cheaper) products starting in the late 70's and then in full force starting in the 80's. That in turn came under scrutiny for some products, like frequently used laptops that needed to take everyday knocks. Businesses make cheap items they can reasonably sell with acceptable losses to poor quality, customer support, etc. Other 'mission critical' products get high quality. However, most consumer goods do not fall into this category. It is the ALQ - Acceptable Level of Quality.