Your stupidest mistakes

[mmagin]

The most embarrassing one I can think of is this rather bulky portable Agilent base station test set (i.e. a spectrum analyzer somewhat specialized for stuff related to mobile phone network stuff). 

Now, I got it cheaply in non-functional condition.  But in trying to disassemble it incorrectly I managed to break the PC board that acts as a feedline from the input connector 

I still have to try fixing that up again.  May just try replacing the broken part with a length of .141" semi-rigid coax and a matching panel bulkhead mount connector and carefully tack the other end to the broken board.  I think the actual problem might be a power supply issue, same as a similar model Shahriar (The Signal Path) fixed in a video on his channel.  (Edit: this is the one, wish I'd watched it before making that mistake!  )

[BradC]

Had a jury rigged protocol analyser comprised of a PIC on breadboard connected by USB to a laptop (an extremely expensive laptop) connected up to a mains power meter while reverse engineering the comms protocol. No problem as everything was floating with the laptop on battery. Noticed laptop battery getting low, so plugged the adaptor in. Laptop no longer floating with the associated smoke and flames.

First kid was only a matter of a few weeks old. I blame sleep deprevation. Frustrating thing was the isolation transformer was in the next room, I was just too lazy to use it.

[mmagin]

I'm still resisting ! I have one analog oscilloscope and since I'm only fiddling with audio amplifiers, I need it for observing sine waves
and the clipping level. Digital ones still can't beat the analog in this field. 

 Many here will disagree with you, but I too have been happy and well served with my vintage Tek 2213. 

You guys are so far behind the times; move into the digital age; my go to oscilloscope is a 2232.  You should always have at least two oscilloscopes so you can use one to repair the other; two of several different oscilloscopes is even better.

This reminds me of how I ended up owning four scopes, two of them 500 MHz scopes.  First I got a HP 5370 counter.  Then I realized it's clearly a bit out of calibration, so I start looking at the service procedures.  Most things I have covered, but I don't have a really fast pulse generator (they specify the HP 8082A) nor do I have anything faster than a 100 MHz scope (they specify one of those HP portables that was hot shit in the late-70s.)  So now I have an HP 8082A.  I also have a Tek 7904 (I already had a 7613, so another 7000 series scope seemed like a great idea.)  But getting the 7904 properly working has been an adventure, so I've gotten a HP 54615B which was cheap and entirely working.  (But I can't sell my 54645A -- it's a lot lower bandwidth, but it has a lot less memory depth.)  At least I gave away the Tek 468 that was truely redundant for me.

(I'm not counting a Tek 323, it's so tiny and lives elsewhere.)

[Powermax]

When constructing my first computer build, (the computer I am currently on )  I was debugging a legacy PCI card and when I plugged it back in something shorted and caused the computer to shut down. (did not know PCI was not hot-pluggable at the time, I learned that the hard way!  )I could not get the computer to boot after that for some reason. Resetting the CMOS worked, but all of my PCI slots died!  Legacy PCI, PCIe 4x, PCI8x, PCI16xv2 (for the graphics card I was going to install in the future) Only 2 small PCIe 1x slots survived.

Since I was having problems with those to begin with, I tried to ship it back as DOA ( PCI were troublesome and something else) but the board came back claiming that the CPU pins were bent. Sure enough, when I looked, it appears as though someone stuck their thumb all over the socket, bending over half the pins! I know that wasn't me, it had to have been ASrock.  So I spent half an hour tediously bending all the pins to as close to where they belonged as possible. I broke one pin but the board does work except for the most of the PCI's.

I still suspect it's a very simple issue, I think I vaporized a power or ground trace somewhere, but I do not know the pinout of PCI slots to verify. Given that it was a legacy PCI, which is not natively supported on Z77 chipset, there is a logic chip responsible for essentially converting PCIe into legacy PCI, and I suspect this chip may be blown. It's a really small package so I don't think I'm going to risk trying to replace it without proper equipment.

[Powermax]

I have a load of stories I can tell about my quadcopter project(s) as well!

I managed to kill not 1... but 2!! microcontroller boards, due to the same mistake that occurred in different ways. The microcontroller board in question is essentially a modified arduino a few sensors on it. Called the Multiwii board.

I had a ratsnest of wires of the quadcopter wired up on my bed and was working on fixing the frame after a crash. The Multiwii board was mounted directly on top of the power distribution board, which connects to a 3S LiPo battery capable of delivering over 1KW. While it was apart, some guests arrived and I left the board alone next to the distribution board while still connected. When I came back and hopped back on the bed it moved just enough to short out and many wires went up in smoke, along with the multiwii.   


Then I had another incident where my OSD display was not functioning properly. Strangely it worked whenever I got my scope out and hooked up the ground. As soon as I disconnected the video out from it back to the transmitter I did not have an OSD overlay from the camera video, or that the picture was really fuzzy, I don't remember anymore. I realized an hour later that while video was being connected, the ground was not connected on the OSD board!  Well I connected it up to 12V as the board requires 12V to go into a small 5V switching regulator, but I had the plug one pin over too far, so the grunty 12V from the distribution board shorted to the ground on the board. It blew the board, as well as the main multiwii board. 

After receiving the replacement boards, I rewired the whole thing so that the minumOSD board no longer required 12V.


more recently, I was rewiring the quadcopter and my hands slipped and well same as before 12V got connected to ground, causing a whole bunch of wires to go up in smoke. At this point I was so frustrated with my clumsiness and having to keep buying replacement parts. I mean, I literally just fried the minimOSD board, the multiwii flight controller, a I2C to UART converter board, and a GPS module, literally everything!!! 

Once I recovered from my mental breakdown, I decided to make my own flight controller board with a old dusty arduino I had lying around. To my amazement, although I killed the ATmega328 chip on the board, a small low dropout regulator (which sacrificed itself) protected all the sensors. I have a box of shame which all dead electronics go to in case I need them later for some reason. Furthermore I discovered that the GPS module that smoked when I tested it by wiring it to a USB FTDI adaptor only smoked because a reverse polarity protection diode which blew and shorted. After desoldering and replacing it, the thing worked fine! And because my arduino mega has more available UARTs, there was no longer a need for a separate board to allow the UART GPS to interface to the multiwii! And not only that, but the mega, due to it's increase processing power, allowed me to enable more features in the later releases of the multiwii code, such as navigation!!! So it turned out to be a win-win! 

So after all that, my quadcopter is now a simple DJI flamewheel frame with a arduino-mega double-sided sticky-taped to it, with a rat's nests of wires going every which way, as well as a plastic cutout mounted on top of it with holes for the standoffs for the old multiwii board, some wires connect the I2C bus on that board directly to the arduino mega which allows sensor data to be collected by the mega. Directly on top of it I have the GPS module mounted with more wires going to the mega, and I had then cut out a small perfboard and mounted four servo headers for the 4 ESC's (brushless motor controllers). It's a ridiculous mess but it WORKS!!! (well, most of the time, anyway.) 

[kaz911]

My first (as far as I remember) and stupid'est mistake - was when I was 7-8 years old. I wanted to make the door to my room automatic. So I rigged a small motor and a counter weight up. The motor was 220v and plugged into a socket next to the door with the old euro "exposed" 2 pin plugs 220v. I used wire for the motor and counterweight - and while "securing" the counterweight with the metal wire by twisting it around - the other end of the stiff'ish metal wire landed on the 220v plug and of course shorted on the two pins.

I got a huge boost of 220v through me - but fortunately the wire burned over and then popped the breaker.  It took a day or two before I finished the project and my parents never found out - apart from they thought it was strange their clock-radio in their bedroom showed the wrong time......

I could have won a Darwin Award for that.... But I got my automatic door

[Scrts]

In early days, I've tried to check the resistance of the 220v in the socket... The smoke from the analog multimeter burnt my eyebrows

[ZeTeX]

In early days, I've tried to check the resistance of the 220v in the socket... The smoke from the analog multimeter burnt my eyebrows

definitely not a fluke multimeter

[Inflex]

Working in Africa, late at night, trying to woo a lady, replacing EU 220V plugs on IEC leads with the South African monster plug,  grabbed the wrong end of a lead, which happened to be plugged in to the wall, the blade vaporised leaving a 1/4" hole in it,  leaving me blinded, shocked and racing down the hall way to check that the AS/400 mainframe was still running, as the lead was plugged in to the UPS outlets.

... did I mention I was trying to woo a lady?

[Pinkus]

My worst mistake ever...

throwing away all my old databooks (Motorola, Texas, National, Linear, Harris, you name it), because with CDs and the Interwebs they'll no longer be needed.

Boy, was I mistaken. I'm still crying.

Ouch - this remembers me that I sold all my vinyl records for small money because CDs were so much sophisticated..... 

But back to the topic: the worst thing I can think of was producing 500 PCBs for a customer where the planned D2PAK voltage regulator was not available on stock. So I purchased a similar replacement regulator .... not recognizing that the main tab was not GND but Vin at this regulator. I did not found out before the 500 units were produced and shipped to me and were not working. I can tell you: manual desoldering of 500 D2PAKs is a nightmare. Lesson learned.....

[Inflex]

But back to the topic: the worst thing I can think of was producing 500 PCBs for a customer where the planned D2PAK voltage regulator was not available on stock. So I purchased a similar replacement regulator .... not recognizing that the main tab was not GND but Vin at this regulator. I did not found out before the 500 units were produced and shipped to me and were not working. I can tell you: manual desoldering of 500 D2PAKs is a nightmare. Lesson learned.....

Been there, experienced that   At least it was only 100 in my case.

[SingedFingers]

Biggest stupid mistake was "Don't take it apart, turn it on!". This resulted in clouds of black smoke pouring out of the back of an old Heathkit scope that had been rewired for 110v as the previous owner could only get a 240/110v site transformer for isolation. Took a week to get rid of the smell. Scope was not recoverable.

[eblc1388]

Directly powering a small electronic device which use a 9V cell from a DC supply.

As I have a spare 9V battery clip at hand, I thought it would make a much better connection than using just the crocodile clips alone. I connected the red lead of the spare battery clip to positive and black lead to negative of DC supply output. Finally, the spare clip was mated with the battery clip in the device.

I'm 100% sure the connection was correct, but....

[Inflex]

As I have a spare 9V battery clip at hand, I thought it would make a much better connection than using just the crocodile clips alone. I connected the red lead of the spare battery clip to positive and black lead to negative of DC supply output. Finally, the spare clip was mated with the battery clip in the device.

Whooops, yes, that can be easy to happen with those 9V connectors.

[nctnico]

I just let the magic smoke escape from a relay. No, not from the coil! 
Edit: took it apart and a picture:

[joeqsmith]

Working in Africa, late at night, trying to woo a lady, replacing EU 220V plugs on IEC leads with the South African monster plug,  grabbed the wrong end of a lead, which happened to be plugged in to the wall, the blade vaporised leaving a 1/4" hole in it,  leaving me blinded, shocked and racing down the hall way to check that the AS/400 mainframe was still running, as the lead was plugged in to the UPS outlets.

... did I mention I was trying to woo a lady?

    And let's here the rest of the story. 


My mistakes are far too many to post.  Even the mistakes in my posts are too many to post.

A long time ago, I was working on a project that had a microphone.  I went to wire it in and there was a red, black and some other color.  Being rather cocky, I assumed I knew what I was doing and did not read the data sheet.  Red is not always positive and black is not always ground.  This was a $70 mistake but a lesson learned.  Read the data sheet....
 

[Inflex]

    And let's here the rest of the story. 

Please to report that was the only disaster for the night, everything else was a greater success

18 years ago now, still have the pocket knife as a reminder.

[drussell]

I have a load of stories I can tell about my quadcopter project(s) as well!
...
more recently, I was rewiring the quadcopter and my hands slipped and well same as before 12V got connected to ground, causing a whole bunch of wires to go up in smoke.

You do realize they make these things called fuses, right?   

[Powermax]

I have a load of stories I can tell about my quadcopter project(s) as well!
...
more recently, I was rewiring the quadcopter and my hands slipped and well same as before 12V got connected to ground, causing a whole bunch of wires to go up in smoke.

You do realize they make these things called fuses, right?   

What my quadcopter really needs is a complete re-wiring, and making a PCB with all the parts on one board. Maybe a ruggeduino too. 

[peteb2]

I look down upon my 57yr old knuckles at the scars from RF burns that happened a good 30+ yrs ago. I was just out of school and i had been lucky enough to earn a position as a Tech-trainee with the nation's State Broadcast Radio...

It was my 1st night at the transmitter site to be shown what transmitter maintenance was all about. I was there with my supervisor and i was to learn-by-doing how to shutdown each of the 5KW halves to what was an aging valved (tube) AM Radio Transmitter, change the output coupling over and bring up a spare while we put the main into a dummy load and run the tests.

It was well past midnight. The supervisor was thinking about lots of other things and i had never in my life seen such gear. He yelled over the noise of ducted air; "Stand on that old wooden chair, up by that big frame full of heater coils (the load i was told) at the back of the cabinet and when i shut off the plate volts you'll hear the breakers bang out so spanner undo that huge brass nut on the insulator and swing the brass link onto the Load and tighten it up".

Upon the sound of his voice of "GO" I did exactly as he said, (only he switched off the plate volts to a totally different cabinet and did not notice)! I heard the bang as he said would happen so i simply UNCOUPLED THE LIVE 10KW transmitter and joined it to the load. Yes i was impressed by the little bits of lightning that came off and out of my fingers as it did it and how the load turned into a white hot stack for a microsecond, dimming to a red glow and then just back to a warm heat... as my ears rang to the sounds of what had been BANG, BAAANG and BOOM...

There was an alarm of some sort going off so i came down off the wooden chair emerged from the rear of the cabinet and saw my super looking confused; "WHAT, What ... what"...

Then we looked at the finals through the cabinet window (airblast glass tubes) which both had holes right through their plates.... He quickly turned the entire TXr off, we opened to doors to see loads of damage... bits of exploded big ceramic capacitors and what he called the final drivers (tubes) with melted glass.

That's when i mentioned my hands smelt like Kentucky Fried Chicken and were kind of sore. The super wasn't really that interested and was completely furious.

We stayed there all night, we'd enough spares and had the poor TX running again on half power for another crew who specialized in such intensive repair work to take on the job of rebuilding the mess another day.

My hands took weeks to come right with all the little burn blisters deep inside. I was told was how lucky i was it because it could have been way worse....!

These days i guess OSH would have been over everything!

[tautech]

Great story peteb2. 
I'll forever be watching for a guy my own age with scarred knuckles and then I'll know it is you.


Going away for a couple of weeks so thought it was best to power off the PC rather than leaving it in sleep mode....you know thinking green. 
Turned it off at the wall in fact, and thinking little about it as one would, flicked the switch on return only to be greeted with a loud pop. 
Dead as a Dodo bird she was with that unwelcome smell of fried electronics.
Hell, it's not that old for a PC, what's wrong ? 
It was winter and some mangy little bug had crawled into the PC PSU to keep warm and unwittingly used its body to create a bridge between an HV point where a SMD resistor had lessened the gap between the PCB and the ATX metal case. Now that wasn't the problem per se, rather that I'd turned off the power that kept the mangy bug in a nice dried/fried state and without it the little bugger was free to absorb a couple of weeks of moisture to then become conductive. 
Now the stupid bit begins:
Bugger it's my main PC and not worth shagging around with so off to the PC doctor for a new PSU....built a few, I know what I'm doing, besides I've got all this EEVblogging to catch up on. 
Hurry, hurry, wack the new PSU in and with X'ed fingers turned it on. No go, bugger. 
Oh shite, the mobo's fried  so off again for another mobo from the doctor....can't get the same processor socket anymore. 
So hand deep in pocket fishes out for new mobo and processor and shoots home to throw it all together. 
Gather tools, remove covers only to find some dumbass hadn't connected the 4 pin ATX 12V supply lead.       

Got a spare mobo and processor though. 

There's one born every minute....there's one born every minute.....there's...................................

[Inflex]

Gather tools, remove covers only to find some dumbass hadn't connected the 4 pin ATX 12V supply lead.       

Got a spare mobo though. 

There's one born every minute....there's one born every minute.....there's...................................

That catches me out about as often as forgetting to put a stick of RAM in to the laptop mainboards after I've been doing rework on them "Cursed thing still wont POST! (wastes another hour)"

[matseng]

I needed one more cable connected to 5 volts while debugging an old arcade game board. So I hooked the wire up to the sense input on my PSU without checking that the "shortening bar" between the +out and +sense on the PSU was securely tightened.

It wasn't. :-(  The sense got some random half-floating voltage from the PCB and the PSU promptly jacked up the 5 volts in order to compensate....

Two dozens of unsockeded TTLs and the CPU blew. By some reason the uv-eproms managed to stay alive.

[tautech]

Another some 35 years earlier while working in a small engine business, happened after spilling some petrol down my leg......thought little of it at the time. 
We wore calf high lace up steel cap work boots, mine had a fluffy band at the top that in dirty environs would prevent any loose matter from getting down your boots.  Except liquids. Oh well, wet petrol foot would smell better than they usually do. 

Rushed off our feet we were, flat out serving customers, one having something that needed grinding....I'll do it and out the back I went and you can well guess what happened next. 

Hot footed it out of the workshop. which luckily had a tap just outside to douse the flames that threatened to cauterise the family jewels. 
I don't normally mind the smell of cooking meat but when it's your own...bugger that.
I ended up with a raw ring completely around my calf that when healed looked for all the world like the original Aussie birthmark. 

I swear I have no Aussie breeding in my ancestry....I swear I don't. 

[LHelge]

I also have a multicopter related incident. 5-6 years ago I was developing my own quadcopter controller. I brought it to our countryside cottage over Easter to finalize tuning of the PID controllers.

To control the motors I used standard RC motor controllers. These take a PWM signal in where a 1 ms pulse means 0% throttle and 2 ms pulse means 100% throttle. If you power up these motor controllers with 100% throttle request and then change to 0% throttle you can calibrate the endpoints on the throttle curve to your timings (if you aren't spot on 1 ms and 2 ms). I created a calibration firmware which looked like this pseudocode:

Code: [Select]

main() {
  setThrottle(100%);
  wait(4 s);
  setThrottle(0%);
  while(1);
}
It worked great to perform calibration.

When I was working on the PID-tuning I had some problem which I cannot remember, but I suspected it had something to do with throttle calibration. I started flashing the calibration firmware but forgot to disconnect the battery leads first. This resulted in the (very overpowered) quad running at full throttle for 4 seconds cutting up my left hand with its razor sharp carbon propellers. My SO was not happy about spending the Easter driving me the 1 hour drive to the ER and back, and waiting for them to stitch me up.

I still have some scars on my left hand to remind me to never work on my multicopters with the props on.