Your biggest electronics stuff-up?

[VK3DRB]

We all have make mistakes with electronics or electrical gear.

Here is my first biggest one. When I was 19, my Ford Cortina car had one dim headlight. I removed it, smashed it with a rock to punish it, then bought a new headlight. To my horror, the new light was dim too. The cause was a corroded ground wire. 

Another one was I soldered a tantalum SMD cap reversed on a PCB soon after they first appeared on the market. The result was a white hot fireball burning a hole in the PCB. I learn two lessons that day, one being the standard of having a negative strip to indicate the positive terminal on a tantalum capacitor is idiotic. 

Anyone else would like to leave their pride at the door and share what silly thing they have done in electronics or electrical work?

[Rerouter]

More recently while working in a spectrum analyser i received in parts i could not get the display to do anything, now the coil wires run off to a nice big 6 pin plug with a mating one on the deflection board to the right, now being the geniuses the manufacturer was put the exact same connector on the CRT power supply board and being how there was no schematics for these boards and i only later found out there purpose went right on ahead and hooked it up to the power supply socket

low and behold this shorted the 80V unreg to the -25V rail on the vertical coil and +50 regulated to ground on the horizontal leading to my now open circuit and damn near irreplaceable CRT being dead,

now the fun is there is a buffered X,Y and Z blanking output, but i will first have to trick the thing into powering up without a CRT,

[dr.diesel]

Not totally electronics related, but I tripped off a 500MW generator replacing a power supply once.  The power supply was actually on a different unit, somewhat complicated and elusive dick-up.

Over $100k between lost generation, equipment damage, startup costs etc.

[nowlan]

Failed my electronics final project, because the Motorola ADC I was using had pins backwards.
Data 87654321 were on pins 12345678 or to that effect. So everytime LSB went up, the output would swing +/- 127 points.
Being a clueless student, it looked totally random to me.

[nctnico]

When I was 16 or so I spend a couple of hours figuring out why the 5 bit counter in my first GAL16V8 experiment counted to 30 instead of 31. It turned out I left the output enable floating and getting at 30 provided enough leakage to disable the outputs.

[TerraHertz]

Urrgh.... thanks for making me think of.... those things.

Hard to pick the worst one. Maybe this:

I'd always wanted a Tektronix 7104 1GHz analog scope. Pinnacle of the beautiful engineering Mt Everest that the Tek 7000 scopes were. All my working life I'd coveted them, but never could afford any. Then discovered ebay around 2000, and finally around 2006 bought a 7104 from the USA. It arrived undamaged, and working fine. I had also bought a service manual.

As advertised, the scope was quite dirty and had slight surface corrosion in a few places. Nothing unrepairable. So I set about restoring it.
Now as we ALL know, the very first thing you do when disassembling anything containing a CRT, is to discharge the CRT anode to a suitable ground, typically the CRT outside grounding braids or case.
In the 7104, the CRT anode lead has an in-line HV plug, which is easier to get to than the clip on the side of the CRT.

But to reach the plug, you have to reach into a narrow space between the deflection amp board and the CRT.
I reached in, both hands, and gripped the two plug halves. Pulled them apart.
Then accidentally dropped the male end one (from the CRT.) It fell down past the amp board, and made that 'SNAP!' sound of a HV discharge. To the 1GHz amplifier board, covered with irreplaceable weird Tek custom ICs.

Needless to say, the 7104 is completely dead. Power supply doesn't start for one thing, and that too contains custom Tek ICs. Half-hearted depressed attempt to get the power supply going with dummy loads failed. After that the whole thing got put away as too horrible.

Sob.

[DomesticHacks]

Scope ground connected to negative voltage of a old Commodore (assuming it was gnd) and wondered why the Pcb trace got black. The scope seems fine...I hope

[VK3DRB]

Scope ground connected to negative voltage of a old Commodore (assuming it was gnd) and wondered why the Pcb trace got black. The scope seems fine...I hope

Hey, reminds me when I connected the ground lead of my oscilloscope to the live chassis of a Sony Trinitron colour TV around 1983.

BANG!

I had connected to the TV a so-called "isolation transformer" that was not isolated - it was an autotransformer. 

[poorchava]

Shorted a 500VA 12V mains transformer through scope ground.

Reversed 5V and 3V3 supplies on a 40€ DDS chip

Welded an ATMega32 into th breadboard plastic when trying to unlock wrongly set crystal

Made a 1.5mm hole in scissors by accidentally shorting few milfarads worth of caps.

Managed to connect USB plug to mae DB25 header (significant ethanol intoxication may have played some role)

DUI (Design Under Influence) resulting in reversing GND and VCC on a big power board (well, a few transistors have survived)

Killed a damn expensive 50W LED with ESD after stroking my cat (o.O)

[daqq]

I've had several minor ones - the normal - reversed polarity, 90 deg rotated TQFPs... but my most major screwup didn't happen, but might have happened.

One of our products has an LED display, 200 or so individual LEDs. I've hand assembled the prototype, worked as specified. So we sent the final version to the fab house, some 100 pcs of the boards, with the assembly drawing. But they were still waiting for one part (switched power supply).

So I waited, so, I was looking over the board, when I noticed that the package looked fishy. After closer inspection I noticed that the cathode was switched with the anode. I nearly crapped my pants - the LEDs were on one side of the board only, what if the assembled the one side already? I quickly called to fixed the drawing, and in the nick of time indeed - they got the parts later that day.

[Psi]

Probably the time i had the oscilloscope on the isolating transformer and 230V connected to the oscilloscope ground clip.

It worked fine until the scope case touched a nearby instrument and all sorts of odd patterns appeared on the scope until it completely died a few seconds later. 
Luckily i found the service manual and managed to fix it with 6 replacement transistors for the horizontal driver.

[G7PSK]

I dropped a socket extension bar onto a 1200 amp hour 48 volt lead acid battery pack when working in a confined space. Of course it had to land across the terminals, there was a flash but not much noise and the bar was fused to the terminals and glowing orange, I tried to knock it free with the ratchet handle but it would not move, by the time I got a pry bar there was clouds of gas escaping the batteries so I decided that the best option was to remove myself.
Shortly after the terminals melted there was a flash and a loud bang and half the cells were split.

[johnh]

Not personally my stuffup. But, I had to fix the result of it.

I  had this work experience job for 6 months in 1977. The company I was working for had built a electronic scoreboard for a sports ground. The scoreboard was been modified to so that it could be used at night.  They used ordinary filament light globes to make up the displays.   Individual controller per globe.

An electrician wired up a the step down transformer back to front.    You can guess the result 

Every controller that was on blew up.   i had to make all these controllers.  They used a LED and LDR  potted in some araldilte as opto-isolator.   

[Abstr7ct]

Unfocused, I connected a 220V to 12V step down transformer in the reversed direction (the secondary wires to mains and the primary wires to a breadboard). You know the rest of the story.

[cosmos]

First ASIC I worked on ... made stupid error in the logic schematic and missed it in simulation too ... almost everything works except it occasionally locks up ... "only" 50K USD (~1990) to do a re spin.

Not mine, but close to me, compiled memory module for ASIC with 5 bit address bus.
40 of 64 entries in actual module, 64 entries in simulation module, address bus signals reversed...  200k USD re spin.

Very fresh out of school, got to do tests with  a very high performance (2GHz clock) ECL DDS/DAC, we had 4 and needed 3 and these where unobtainums and ~2KUSD a pop.
One just stopped dead while I was working on it for no apparent reason ... might very well have been an ESD event as I think the protection might have been minimal compared to commercial circuits.

Made a 12 layer PCB with footprints for the circuit we had samples of (very fast AC logic in mil spec ceramic version) turns out our samples where the only 3 left world wide ...

Enough of these and one can see how being obsessed with details and checks is a good thing.

In our lab we had an appropriate sign (as I remember it):
"Experience is proportional to the equipment ruined"

[schwarz-brot]

Just happened a few minutes ago.

Hacked a laminating machine to work with toner transfer method. Therefore I had ripped out all electronics and rewired them as the device was total crap before. I managed to get the temperature braking switch wrong or it didn't work at all.

While printing my layouts I turned it on to heat up - what it did - to excess. When it started popping and warping I ripped out the mains and brought it outside where it stands now in the rain and cools down.

That's what happens if you forget to double check 

Now I have no idea how to get the boards done over the weekend 

[IanJ]

Two come to mind from the 90's.

Power supply, rectangular pcb with transformer mounted in  the middle. Pair of terminals at either end, 240vac in and 12vdc out.
BANG!!!!!!......yup, shouldv'e marked INPUT and OUTPUT a but clearer!

Cleaning out the large pcb etching tank, found an old metal gallon tin to store the old fluids, half full.......left it on a shelf for a couple days. All looked fine till I went to lift the tin........right at the level mark the metal had been etched right through and the only thing holding the tin together in one piece was the paint on the outside of the tin which hadn't been etched away!.........what a mess!

PS. The amount of tantalums we had go bang were quite high......the assemblers we used barely knew how to solder.

Ian.

[bingo600]

Getting 5 x 15v/3a psu's , the ones that looks like laptop psu's  , for my new FEI Rubidiums
Connecting one via testclips to a Rubi. Innercore to plus & braid to minus.
Wondering why the lock led took so long to come on , the Unit was just removed
From my lab psu.........

Geezz i hate the designer that made plus go to the braid.
And my self for assuming .....
Offcause it was my most stable Rubi that went to silicon-heaven.

/Bingo

[AlfBaz]

Not my stuff up but...

We were doing a big install project at a water treatment plant that had many fiberglass tanks. On the top of one of these tanks a boilermaker had to weld some stuff for a bracket. They had been getting away with just running out the welding lead having the earth attached to the main part of the metal structure. On top of this fiberglass tank, the only earth was the earth wire from a 2C+E 1.5mm^2 cable which ended up with this burnt black spiral all the way along it

Needless to say I've seen this happen many times since

[peter.mitchell]

One project I had was a "suitcase" generator. A generator that was about the size of a cordless drill carry case.

Code: [Select]

2 Stroke motor -> centrifugal clutch -> gearbox -> induction motor -> rectification -> small smoothing caps -> inverter -> output
                                                                                                                                 -> starter charge circuit

Had some throttle feedback with endpoint limits, so if overloaded it would just stall. Some controls for electronic starter, Prototype worked excellent up to around 2kw continuous, assembled a few test units for people to use and report back on.

After about 3 days they were all back, and the starters were all fried, so were the LiFePO4 cells used on the starters, the induction motor looked severely over heated, as did the inverter.

Some poking around revealed that after a overload or two, the people would just tape down the momentary starter switch so it wouldn't stall, and if it did it would start back up again...

I guess you can't underestimate people...

[bookaboo]

Working through until 10pm to get a system ready in the workshop for installation on site 4hrs away. Double checked all the parts were packed, double checked all the tools and equipment.
Up at 6am, got to site and realised I'd left the 3-pin header for one of the boards at the workshop. Had to make up a cock and bull story about software to cover my tracks.

8 hour drive for a £0.20 component.
I check all my headers are in place now.

[Zbig]

Getting 5 x 15v/3a psu's , the ones that looks like laptop psu's  , for my new FEI Rubidiums
Connecting one via testclips to a Rubi. Innercore to plus & braid to minus.
Wondering why the lock led took so long to come on , the Unit was just removed
From my lab psu.........

Geezz i hate the designer that made plus go to the braid.
And my self for assuming .....
Offcause it was my most stable Rubi that went to silicon-heaven.

Center-positive-ness of DC barrel connector is neither implied nor granted. Pretty much all the devices of Japanese make I encountered were center-negative. And I'm pretty sure that's not a rule, either.

[VK3DRB]

A remarkable incident at IBM about 20 years ago...

An recent new hire had a 240V PC power supply mains switch fail (it switched active and neutral). So he used an external cheap replacement switch connected with about 60cm of thin speaker wire. The new switch had a metal lever and the poor solder joints had two wraps of greasy plastic tape around them to "protect" the factory floor operator who wore a conductive ESD coat. When I saw this, I reported the incident and read the riot act. The same guy imported a 110V-only monitor from the US and plugged it into the 240V mains - BANG! He said he thought Australia had 110V in the mains. He seemed very confused. The last straw was he made up a metal test fixture with mains active 240V bare wire about 1mm from the metal case - the case was not earthed. He was fired not long after. He said he had an engineering degree from a university and he stated during the interview his extensive software and hardware skills, but he was proven to have none.

The lesson learnt... ALWAYS test new job applicants and check their academic and work experience credentials.

FOOTNOTE: He once had a book on Klingon to English translation on his desk. I asked him what use it would be, and he replied he was learning Klingon fluently so he could put the skill his CV because if an interviewer happened to be a Star Trek fan, he could talk Klingon and he will get the job above all other applicants. I realised something wasn't quite right with this bloke.

[DL8RI]

Measurement on a Tube-Amplifier. I was 18 or so...
Wanted to check the Anode-Voltage-Ripple with my scope, forgot to remove the signal-generator... and -of course- chose the wrong side on the HV-Caps for the ground-clip.

Except the ground-clip of the probe everything survived. But the light-show was good.

[Tinkerer]

Not mine, but I was right there when something happened.

So me and another guy(EE, been with that company for years) were looking at a certain high priced unit(as in couple hundred k $$) that was supposed to be shipped out real quick, like in the next day or two. This particular kind they never get right, always something wrong. Well long story short, we are looking at it and he goes to hook power to it directly via two leads. We suddenly hear a large pop that startled the crap out of us. Everything in the area has gone dead.
Turns out he had accidently dropped the two wires on to the metal frame around the connector and blew a small crater into it. haha Which had then tripped the breakers. Luckily, nothing else seemed to be damaged. He then proceded to make sure that would never happen again. The metal connector frame got replaced and thankfully that POS left the next day.(its basically a damned computer and there is always something wrong with those units)