Everything’s Better After the First Explosion

I'd like to take a few minutes to address the fear of power circuits and the great relief of successfully repairing the first screw-up of a project. Here is an example.

Awhile back, a stalwart group of engineers here were making improvements to a pinball machine using various Intersil parts to improve the flipper strength and reduce overall power consumption. Odd thing to do, yes, but fun.

The flippers each were getting clobbered with 500W for their initial throw, then much less sustaining current. The flippers needed an off-line rectifier and crude filter to provide unregulated and moderate-ripple DC supply, so I made a voltage doubler circuit that is good for over a kilowatt.

Two versions of voltage doubler circuits; the left circuit can use capacitors with lower voltage ratings. The inrush current limiting circuitry and the bleeder resistors are not shown.

Two versions of voltage doubler circuits; the left circuit can use capacitors with lower voltage ratings. The inrush current limiting circuitry and the bleeder resistors are not shown.

I chose beefy rectifier diodes and 4700μF/250V filter capacitors. These are impressive capacitors, and they are dangerous. You can't leave them charged; their stored energy can kill a person easily if tampered with. They kept the indicator LED's illuminated for a half hour after power was turned off. For safety, I had a relay switch in a bleed resistor when power was switched off. I used a 25W metal-jacketed resistor to bleed a transient 100W to effect a two-second bleed-down. The metal jacket is intended to provide a thermal sink for the temporary over-power event. This worked well and is pretty fail-safe.

But you can't just turn on the power through the rectifier system — if the switch happens to close when the line AC is at peak value, a massive current surge will flow into the uncharged filter capacitors. This can be over 100 peak amperes, and the rectifiers will surely pop; the power switch could even weld shut. (Note: normal circuit breakers are slow to act and would not offer protection for this problem.) I used a soft-charge circuit that imposes a series power resistor to limit line current to about the same 100W peak dissipation using a 25W resistor. A low-voltage circuit times out two seconds after turn-on and shorts out the limiting resistor for normal operation.

I was a little intimidated by this beast even though I've built power vacuum tube circuits, burned out resistors, blew wires, drew arcs, and exploded electrolytic capacitors galore.

Thus, when the inevitable bad connection between boards blew stuff up, it was like finding the first scratch on your new car — in a peculiar way, a relief. Once perfection is ruined, you can relax. The rectifier was replaced, and the blown circuit-board trace was cleaned up and bridged. Everything worked again, and you couldn't tell there had been an explosion.

The beast, repaired.

The beast, repaired.

That being said and done, I can make the case that there is a benefit for every project having an early explosion.

I can't be the only one who sees the value, right?

31 comments on “Everything’s Better After the First Explosion

  1. Michael Dunn
    May 14, 2013

    >I can't be the only one who sees the value, right?

    Absolutely! And the principle applies to most anything, though the mistake is usually less dramatic than an explosion. Try to make a perfect meal, artwork, schematic, paintjob, housecleaning…whatever. After the first oops  moment, there's a lot less stress :-}

  2. Brad Albing
    May 14, 2013

    And now I wish I were back at Intersil so I could play with the pinball machine.

  3. amrutah
    May 14, 2013


        During my college days I had worked on the Power line communication modem, and have the first hand experience about the charged capacitor.  Luckily, the modem was off the gird after demonstration but the capacitor ends got shorted and blew up with pungent smoke.  Having a indicator for a charged capacitor would have helped.  That was my last exposure for the high voltage circuits.

  4. Brad Albing
    May 14, 2013

    Yep – either an indicator (LED + resistor) or a bleeder resistor like Barry used would have been a good idea.

  5. Netcrawl
    May 15, 2013

    It's better to have a first hand experience like this one.A first hand experience like explosion is really unforgetable one, but it leaves us lessons and notes on what and how exactly we could deal with those stuff. Exploding motherboard and laptop battery provide some great memories something you won't forget and missed.

  6. Brad Albing
    May 15, 2013

    Well sure, that sort of experience is an excellent teacher, but let's hope no one gets hurt. Certainly, these kind of mistakes/problems provide excellent teaching experience.

  7. eafpres
    May 15, 2013

    @Barry–in the spirit of Boeing's battery issues, I would be more comfortable if there was a metal can around the whole thing with some kind of controllled vent in the case something really stupid happened and you blew those caps.  That thing gives me chills just looking at those two big caps.

    Thanks for the article. I do agree early failures are better than later.  Somewhere in the last year or two I read an article about protyping as early as possible to get design feedback.  It was the same idea–find mistakes early and improve the design, and the best way is to build something.

    Must be one killer pinball machine.

    May 15, 2013

    I've got about a dozen pinball machines in my basement Brad…

  9. Lmccrock
    May 15, 2013

    Much can be learned from the first oops, but it can be expensive, like a cap blows, the engineer was shielded, but it took out a fluorescent bulb in a ceiling fixture. Or, a buck supply does not buck enough and take out a chip which may be worth $1000 or more.

    I am having trouble getting perspective on the caps in the picture. Like, the size of a soup can?

  10. Brad Albing
    May 15, 2013

    I'll be right over.

  11. Scott Elder
    May 15, 2013

    Anyone who is required to work on a circuit board where the resistors have provisions for screw mounting should also be required to undergo annual hearing checks.

  12. WKetel
    May 15, 2013

    Looking at the two voltage doubler circuits the error is very obvious, which is that there is no current limiting resistor. Consider that with the voltage versus load charaterist9ics of the AC line, it is clear that the impedance is way less than one ohm, and the capacitor is a “sort of short circuit” until it accumulates a bunch of charge, and the diodes may develop just over a volt forward drop at the maximum rated currecnt. So how could you expect anything else?. A ten ohm resistor would limit the inrush to maybe twelve or fifteen amps and not interfere with normal operation. OR, you could go to extremes and add a small series inductance to limit the maximum rate of current rise, like the utilities do. Either choice would work, the math for the resistor solution is easier.

    As for the safety concerns, just add a warning tag, anbody stupid enough to get a shock then deserves it . Another option is to have the controller fire all of the solonoids  when the power supply switches off. Or maybe just fire a couple of them.

  13. Brad Albing
    May 15, 2013

    Those caps would be about the size of a soup can, yes.

  14. Brad Albing
    May 15, 2013

    And maybe an EKG and testing BP would be a good idea, too.

  15. Brad Albing
    May 15, 2013

    To clarify, I added those particular voltage doubler ckts with the cutline that indicated that the inrush (and relay) & bleeder resistors were not shown.

  16. Netcrawl
    May 16, 2013

    So what exactly make a good teacher? Could a hands-on experience like this one make a difference?

  17. WKetel
    May 16, 2013

    A good teacher has both insight into the subject being taught and also the knowledge of how to teach it to others. A more detailed explanation is more than I have time to provide right now.

  18. Alan S
    May 16, 2013

    For some reason, when adding filter caps to bipolar supplies, I sometimes draw the negative cap inverted.  Not intuitve to have both caps drawn the same way.  Inverted supply, inverted caps.  Simple.  and wrong.

    So we were testing a board on the bench.  First power up, current looks reasonable, then bang!  Fortunately the board was upside down on an anti-static mat which acted nicely as a blasting mat.  A branch of the company in another city was also testing a board so we quickly called them.  Too late.  Same thing, explosion marks on the anti-static mat.  Another notch on my belt.

  19. Brad Albing
    May 16, 2013

    WKetel – you hit the important part – a good teacher needs to be able to convey the knowledge, otherwise, eveyone's time is wasted. We've all (likely) had teachers who knew the material, but could not convey the info in a way that helped us, the students.

  20. Brad Albing
    May 16, 2013

    “Remember kids, don't try this at home. I'm a trained professional.”

    May 16, 2013

    As a budding engineer, it was amazing to figure out what that “POP” sound was all about. In some instances it could have been an injuring effect. Somehow I have lived through those proto days with minimal protection. Now it is adding a plexi-glass shield to protect the operator from metal transistor cans from shooting silicon out the lids.

  22. WKetel
    May 16, 2013

    But none of those was as exciting as the time our lab manager went to check the 480 volt terminals of a large motor, and he had the meter leads plugged in across the 10AMP current measuring terminals, becuse the tech whose bench he grabbed the meter off of had used it to check current for me on some other project. When he made the connection the 250 volt fuse failed but kept on conducting, the meter probe leads sort of evaporated, and the meter suffered a great deal of internal damage, not covered by warranty. The lab manager survived with only some minor burns on his hands and arms, and returned to work the next day. Since then I always check meter connections before using the meter. I was able to learn from his experience.

  23. Barry Harvey
    May 17, 2013

    Yes, safety is good.  I use a variac followed by an isolation transformer to apply first power.  The isolation transformer allows safely grounding the secondary side so you can use an oscilloscope to troubleshoot.  The variac (which does not isolate) allows you to slowly raise the input voltage while measuring so that if there is a fault, the fault current is low and you can perhaps find bugs without destroying devices.

    The explosion came from external wires connected wrong.

    And, oh yes, I turn power on to an unfinished project while my head and hands are offside or behind something in case something does fly out. 

    You know, power projects just blow up occasionally.  We shouldn't be afraid, but we should be intimidated and think two steps ahead of what you're doing.

  24. Barry Harvey
    May 17, 2013

    Actually they are like beer cans in this case.  They're about half the size of electrolytics I last used in similar service, but still deadly.

    So, electrolytics mainly explode when reverse-biased, or when they are abruptly charged after decades of discharge.  Shorting them doesn't seem to hurt them, but it will hurt whatever discharges it.

    I'm more concerned about grabbing onto a high-voltage DC line since there probably is some such big capacitor ready to fully discharge into you before you can let go.  Grabbing the AC is more benign (assuming you're not well grounded) since you should be using only one hand to probe things (other hand in pocket joke) so you don't complete a circuit with current coursing through your chest.

    Do these things and laugh at your timid peers.  Of course after your first shock you get really careful and thoughtful.

  25. Barry Harvey
    May 17, 2013

    I think everyone who works with high voltage eventually gets shocked.  It helps to be stupid I guess, but not necessary.

  26. WKetel
    May 17, 2013

    I have yet to get a serious DC shock, although I have had a few hot wire burns from yanking out wires with the insulation melting, when I knew that there was no fuse to stop the curent and letting the thing continue would have been VERY EXPANSIVE. AN adequately motivated individual can break a #12 stranded wire very quickly with the right motivation.

    But AC shocks are a different story, although the flash-burn from plugging in a 3-phase plug that had an errant strand of copper was much more painful. That “bang” cleard all thre phases 30 amp fuses, and the noise was quite loud. Did you know that copper vapor is a much better conductor than solid copper? The reason is that it is an ionized plasma, which is also the reason that arc-flash is so very deadly. 

    As far as  learning from disasters goes, I much prefer to learn from the incidents that other people have. It is usually safer, and does not damage one's reputation the way causing explosions does.

  27. antedeluvian2
    May 18, 2013


    although the flash-burn from plugging in a 3-phase plug that had an errant strand of copper was much more painful. That “bang” cleard all thre phases 30 amp fuses, and the noise was quite loud. Did you know that copper vapor is a much better conductor than solid copper? The reason is that it is an ionized plasma,

    I once walked through the basement labs at Cutler-Hammer where they had been (and probably still do) testing circuit breakers. The conductors that were bolted to the wall were these huge copper bars. I think “bar” is a misnomer, “beam” would probably be better. I don't remember exact sizes but they would have had crossectional areas of more than a few square inches.

    And there was this copper residue splattered all over the walls. There must have been some pretty big bangs.

  28. Davidled
    May 18, 2013

    I am not sure high voltage circuit with transformer was tested it the company lab, if this has been happed in the company lab, engineer might report to manger for safety concern, unless this explosion is for any purpose. Secondly, there are lot of nice simulation tool such as Pspice. Without actual board implementation, any circuit behavior can be roughly predicted.  I burned Cap a few time, not exploded something. It sound like you are electrical engineer handling high voltage  not electronic engineer, who test less than 12 voltage in the most case.

  29. Brad Albing
    May 19, 2013

    On a sort of related note, I once brushed my hand up against a 480VAC, 3-phase rectified DC bus (part of a welder) – so that was around 700VDC. Just a glancing touch, luckily. You haven't lived 'til you've bumped up against a 700VDC bus. And sometimes not afterward too.

  30. WKetel
    May 19, 2013

    The strongest shock that I got was while pressing the reset buoon on a STD buss controller card rack. That happened because the disconnrct switch was directly above it on the panel, (thanks, John C.), and the hookup was tempoary, so it was not dressed correctly. It was a hot and sweaty August day, so just the very slightest touch made good contact. I jumped straight back about 8 feet, and then jumped around because of being so very startled. That scared everybody else a lot more than me.I explained to them that the time to be afraid would be when I hit the floor and didn't get up. No real damage was done except for a few jokes about “electric Bills”.

    I was more careful avter that event.

  31. Brad Albing
    May 21, 2013

    I know Barry – he is an engineer with expertise in multiple areas – electrical, electronic, and IC design. Of course, Barry is not recommending that you blow things up. If he did, you're right – his manager would probably have a talk with him regarding the need to not do that in the lab.

    Instead, ths is more of a learning situation – learning how to deal with inevitible situations such as these.

    And it's difficult to deal with situations such as these in PSpice. It won't tell you that the capacitor blew up or that PC board traces vaporized.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.