Admin needed server fast, skipped factory config … then bricked it Welcome to another edition of “Who, me?”, The Register’s confessional in which readers explain how they broke things. This week meet “Pete”, who way back in 1986 “was installing a dedicated file server for a small advertising agency.” And not just any server – a 3Com “3Server”, an oddity that tried to improve on early PC …
A pop, and not a bang?
What a damp squib.
Glad to hear all ended well though :)
First time I've heard about an application for the elusive 80186 - had fortunate experiences with the 8086, 80286+80287, 80386+80387 and so on...
My first 80486DX I was in such a rush that I did not notice what I was doing and mounted the CPU the wrong way round - part of it glowed briefly before magic smoke escaped... had to toss the board and everything. Learnt my lesson that day.
"Hats-off to the inventor of the CPU notch or whatever it's called that prevents you from doing just that!"
A workshop colleague once sat a CPU on the socket then dropped the retainer flap over and quickly jammed the lever over and under the latch to the sound of the corner of the CPU snapping off. Yeah, notches are good if you check the notches are in the right place before you apply loads of pressure. His face was a picture!
"My first 80486DX I was in such a rush that I did not notice what I was doing and mounted the CPU the wrong way round"
About 20 years ago I was struggling to get a video card to work on a new PC motherboard. After trying yet another hardware change - I mistakenly plugged the motherboard's two side-by-side power connectors in the wrong sides of their common socket.
After it failed to power up I realised my mistake with sinking heart. Swapped them round - and to my amazement the motherboard still worked. Not sure if it was luck or a piece of good design of the sockets' pins layout. It was a definite improvement when PC motherboards started to come with a single polarised power connector.
That still didn't stop me plugging the power connector into a floppy disk the wrong way round. They were often tricky - being somewhat hidden with apparently no industry agreement on how they were mounted. One dead floppy.
I have seen memory chips filed down in order to make them 'fit', RJ45 connectors cut in order to make them fit into RJ11 sockets, HDMI plugs butchered to make them fit a USB port and on and on ..
We talk about the dawn of Artificial Intelligence, but what we need is a dusk to Genuine Human Stupidity.
"How did you manage to force a female molex connector the wrong way round into a male connector?"
The small floppy female connector is almost symmetrical. The captive male end only had a very flexible tongue as a key - not a profiled shroud. Normally you had to do a blind fiddle inside the case to get the thing inserted against the sometimes considerable resistance of that tongue. So when the manufacturer fitted the captive male end unusually "upside down" - it was still possible to make the connection with no more than the usual amount of fiddling.
"How did you manage to force a female molex connector the wrong way round into a male connector? I tried it, just could not be done without a lot of excessive force..."
If I remember correctly, the connector for a floppy drive is just four bare pins and it's your responsibility to do any lining up and orientation. No hand-holding here.
4 pin on 3.5 was awful but the 2 insides were ground, outsides were 12/5. I had some 8's with non-standard connectors. However I have reversed the 4 pin molex on both hard and floppy drives. Often very cheap plastic with minimal material around the key side, or brittle to the point of splitting with minimal force.
Yes, with enough force, you could get them into the drive the wrong way round, it was damned hard work...
It goes along with many things. My much abused battery in an (old) MBP was misbehaving a little. I asked our admin to get it replaced. Not quite understanding the battery catch (as I said an old Macbook) she decided to use a screwdriver to help it out. She was extremelty competent in most things and certainly worth while bailing out.
The replacement machiune was 'siezed' by de Kops (I was away at the time and I never got an adequate explanation - even a plausible one) as to why our personal machines were searched. Being away, for, again, reasons that remain murkey even new (it was only 10 years ago), no one asked for or demanded a password or even tried to ask me. However they obviously decided that trying to clone or at least look at the disk (encrypted anyway). Some plod did the same thing - assuming the battery was a disk. Hum de hum. Obviously embarassed they paid up for my and five other machines with similar case smtoms without a peep.
Almost unbelievable... And the f***ing things even had screws to dissasemble them then.
I've done it before as well. Fortunately, only here and amazingly the few times I did "succeed" in making the wrong way around connection, I got clicking from the drive. Put me in mind of my Amiga 1000. Worked after reversing to "proper" orientation.
At times, I've been quite amazed at what some machines live through. Design or accidental protection? No idea.
"Design or accidental protection?"
I often wonder if low power consumer kit has a blocking diode - or even a bridge rectifier - in order to survive people applying power the wrong way round.
One of the apparently poor designs are R/C toy controllers. Often they require a 9v PP3 battery pressed into a rectangular slot with two flat terminal contacts. Seems possible to get it the wrong way round - and the polarity markings are just embossed in the plastic. I always insert a piece of yellow electrical tape and mark it to indicate the terminal polarities clearly.
A lot of the sensors I work with are 24VDC, and use the current consumed to indicate the present reading. (4-20 mA systems, if anyone is curious.) A disturbingly high number of these don't much like having the power hooked up backwards. But one particularly bright manufacturer seems to have put a diode bridge onto the power connections - it really doesn't care which way it's connected, to the point that they didn't bother marking a polarity on it at all.
I recently bought a very second-hand wheel balancer, but there was no power brick supplied with it. On investigation (RTFM), it became apparent that it would accept either 10 volts AC or 12 volts DC, but as the input went first to a bridge rectifier, the DC polarity was immaterial. A wonderful piece of kit, and a well designed power input system.
"I often wonder if low power consumer kit has a blocking diode - or even a bridge rectifier - in order to survive people applying power the wrong way round."
That might appear logical at first glance, but if power can be applied incorrectly, then it's probably because someone realised they could save money in the manufacturing process by not using keyed or otherwise "foolproof" connectors. I think it unlikely they would then add "unnecessary" components. See, for example, your own observation of PP3 batteries in R/C toys.
"If I remember correctly, the connector for a floppy drive is just four bare pins and it's your responsibility to do any lining up and orientation. No hand-holding here"
Yes, on 3.5" floppies, especially in the early days where sometimes even the "tongue" was absent. The fact the person questioning the possibility mentiond Molex connectors means he was thinking the FDD referred to was a 5.25" drive which, by the time of the IBM-a-like PC had a standard connector which only fitted one way.
"If I remember correctly, the connector for a floppy drive is just four bare pins and it's your responsibility to do any lining up and orientation. No hand-holding here."
Now I think about it, it definitely does. I was teaching a friend how to build computers and I took my eye off him for a second and he connected the power to the floppy the wrong way round. A little popping and it was back to Jungle (gives you an idea when this was) for another floppy drive.
"If I remember correctly, the connector for a floppy drive is just four bare pins and it's your responsibility to do any lining up and orientation."
It's not just possible to get it upside-down, it's also possible to plug it in offset, so only three pins connect. This will still cause to to let the smoke out.
I worked at a medium sized PC manufacturer about 15 years ago, and I saw pretty much every possible error it was possible to make, even the ones that are "impossible". Yes, even getting a molex upside-down (it clearly required a lot of force). Things like getting a DIMM in the wrong way around were relatively commonplace.
"I mistakenly plugged the motherboard's two side-by-side power connectors in the wrong sides of their common socket.
...
Swapped them round - and to my amazement the motherboard still worked. "
Me too...
I later noticed that all the 0V wires are towards the centre of that connector, so swapping them shorts all the other voltages together, presumably triggering the PSU short circuit protection. I decided not to test whether the motherboard would always survive.
Many moons ago, when I worked in Education, the 186 processor was used in some Research Machines (RM) kit which we had. Thankfully when I first started, my boss told me that he would look after this kit as it was so old/non-standard (and this was the early 1990s), but it was good enough to run for a couple of years after everyone else had stopped using it (even beat out some 386/486 kit in lifespan). From memory, the kit ran some weird version of DOS with some networking engine thrown on top of it, running some "really interesting" DOS programs.
This was the first and last time I ever came across the 186.
I remember them at my college in the early 1990s. I remember coming in to the college and the mouse wasn't working properly. Other students also said their mice were not working. Some moron had cut up all the rubber mouse balls in the computer labs. The only mice working were the RM Nimbus ones which were metal.
My school (Abingdon aka "Royce's") had a RM Nimbus in the D&T workshop. With Windows 1 on it and a rather early AutoCAD.
About 1 year later I (cheers Mum and Dad) had a 80286 based beast with 1MB of RAM. I saved up for a 80287 maths co pro so I could run ACAD on it. To be honest I had more fun reading Peitgen and some other bloke's fractal related shenanigans and attempting to replicate their results in GW Basic.
Then I got hold of a C compiler - cool and quick and bloody complicated. I went back to BASIC to get the results that I wanted at the time, in linear time. Before you smite me with righteous indignation, bear in mind that I just wanted results and not a cool paradigm.
Those Convergent Technology machines were awesome for the 80s. It's a shame they went nowhere.
The operating system (which was an advanced for the time pre-emptive multi-tasking networking operating system) was called CTOS - Convergent Technology Operating System. Burroughs badge engineered them and in a magnificent display of can't-be-arsedness, changed the OS name to BTOS. What did the T stand for? Nobody knew, but it meant that the string length did not need to change...........
I've done the same thing but not cross country..migrated a clients environment to a true datacenter with 208V PDU's and one of their remote "workstations" got plugged in and same pop, immediately followed by crap....
As soon as it popped I realized I needed to move that stupid little switch...luckily it was same thing..swap PS and away we went...
I knew exactly what he had done at that point..
Likewise. Especially more or less since the same thing had happened to us - just got our first shipment of brand new (386?) desktops directly from Taiwan.
We were used to buying stuff from UK vendors and so didn't bother checking the power supply settings. Cue one whole benchful of new machines going pop..
(Supplier eventually agreed to replace those machines - about 30% of the rest of the first batch and the next batch were DOA anyway due to build problems and our account rep was told that the next batch had to be better (and all the DOA machines replaced free) or not only would he get no more business from us but we'd see him in court. Since we were their first UK customer[1] he managed to get it sorted - but only after he spoke to the head of the company directly).
[1] Which we didn't know about until after.. we wouldn't have touched them with someone else's bargepole if we had known.
Yup, me too.
Exactly the same thing happened to a Dell tech a decade later at a place I was working at in the UK. We'd just gone to a new service contract where us admins were to be "hands-off" the hardware. PSU died in a server, Dell shipped a new one from the US, followed by a shiny new tech to install it.
We all stood around watching the new support arrangement in action... and managed not to giggle aloud when he unwrapped the PSU and installed it without the merest glance at the voltage switch position (poor dude was a bit nervous with the entire systems admin team "observing").
After a nice bang! and the magic smoke escaping, we ended up getting a brand new CPU, memory, motherboard AND PSU.
Hah!
Standard equipment on both the HP 200-LX and HP 100-LX.
They were brilliant machines that gave me pocketable work-from-a-cafe capabilities in 1994 that I didn't achieve again until Wi-Fi, smartphones, and netbooks became widespread in around 2010. Being old-school HP devices, they both still function perfectly.
2708 and 2716 eproms literally showed this through the little round window if arse about face on some of the less sophisticated programmers, you would look to the chip and see all the liitle wires light up dull red and go out. (If you didn`t blink.). "programmer declared chip not seen" is an accurate report. I have never done this myself, ever.
That report leaves a promising person who just fkd up having a serious reconsider about concentration on the job rather than out of the door.
Those chips were once very expensive, and they only have a notch on the small face in the ceramic in the middle.
Ah, the value of the chinagraph pencil
2708 and 2716 eproms literally showed this through the little round window if arse about face on some of the less sophisticated programmers
I remember someone doing this with an early 27256, in the days when those chips had far too many zeroes in the price. He was quite proud of it, stuck the chip on the wall beside his desk with its legs in the air "dead bug" style.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
