EEVblog

[@trombonista92]

You should make an episode all about components that can be built out of pcb races like pcb capacitors, fuses, inductors, antennas, transformers, there are so many options!

[@MrFirstcause]

Agreed!! That is an awesome idea for an episode!!

[@rich1051414]

Transformers? Really? Wouldn't that be more like coupled inductors? Or do you have to add a ferrite bead?

[@SetMyLife]

I would be all ears for a full fledged input protection tutorial!

[@tom7601]

Back in the 60s, I had an RCA color TV come in that kept blowing a fuse. I discovered two parallel traces that had arced across and the PCB had become carbonized, kind of a low-Ohm resistor. I ended grinding out the carbonized area and painting it with a high-voltage coating that is normally used around the second anode connection on the CRT. That tamed it... :-))

[@QoraxAudio]

Should've used peanut butter on that fuse. TVs don't like blowing fuses when there's peanut butter on it.

[@sickbow8167]

No idea why this appeared on my recommended, but I'm glad I clicked it. I learned a little bit, and this was oddly relaxing to watch. Thanks.

[@kendelion]

i dont make boards, im not knowledgeable about this thing and does not have any background about electronics. But i find my self watching a lot of your videos!

[@TomaszWota]

Kitsune Ken Welcome to the club. ;P

[@petevannuys5642]

😭 you doing this nothingness !

[@AlTheEngineer]

Dave, thank you for listening to your viewers and for making videos about the things they ask you about. You're awesome!

[@joshcryer]

Absolutely brilliant overview, very layman level. Sometimes, I have to admit, you post things that fly past me, but this video is really comprehensive and explanatory.

[@thingyee1118]

Awesome explination for those who were unsure. Basically if its going to fair somewhere then make a weak point thats safe from IC / chips. Spark gaps are great little protection features of PCBs so make sure to add one if needed.

[@linksmith1057]

Beautiful class 3 soldering as well on that PCB. I am class 2 certified, and while I cannot remember the exact numbers, I think they are roughly these for through hole components.. Class 1: it simply needs to work. 50% full in barrels, 50% wetting on the source side. Class 2: it has to work for a long time. 80% barrel fill, 75% wetting in the source side, 50% on the destination side. Class 3: medical grade devices. 100% barrel fill, 100% wetting on source side, 80% wetting on the destination side. Class 4: aerospace and military applications. Perfectly wetted and filled barrels.

[@Rev22-21]

Being in the field for some 30 years, it's refreshing to learn something new. Great video.

[@azyfloof]

During the 1000 volt test a tiny speck of PCB material moved downward away from the upper C shaped terminal. So no spark, but still evidence of the HV.

[@sc0tte1-416]

I noticed that too

[@EEVblog]

azayles Yeah, several have point that out, very interesting. Didn't notice that at the time.

[@Nevir202]

azayles 12:25 In the valley in between the two lower left points you can see a spec of the crud shift in the 1kv test on this PCB as well.

[@azyfloof]

Nevir202 Oh yeah! You have a good eye :-D

[@Nevir202]

azayles Thanks, don't know if I'd have been looking for it if you hadn't mentioned it before I saw that part of the video.

[@zhitailiu3876]

I just realized now people love to watch someone plays video games for them, pretty much the same as we watch how Dave does all these interesting stuff for us! Great work! Thx, Dave!

[@onjofilms]

I use software for protecting inputs: int main (){ if (voltage > crappycomponent II voltage > pcb_resitance){ gap = 0; if (smoke < cigar) { repeat = 1; while (repeat){ trace = releasesmoke; repeat--; } printf("shit hit fan"); return 0; }

[@bloomtom]

Seems legit. 10/10 commit.

[@TorutheRedFox]

What language is that? I'm genuinely curious if that would work.

[@ПетяТабуреткин-в7т]

repeat = 1: while (repeat){ trace = releasesmoke: } That's an infinite loop, by the way.

[@joshfixesall4056]

@@TorutheRedFox looks like C for an Arduino or STM or something

[@Asyss_Complex]

@@joshfixesall4056 That could be C or C++ Arduino doesn't have a main() function.

[@marty-fh1rw]

I like the way Dave speaks on here.Calmer.Interesting videos.Thanks.

[@JohnGotts]

I'm glad that electrical engineers care about ESD nowadays. Back in the 80's it was a simple matter to destroy one of the joystick ports on a Commodore 64 simply by using it. In those days the ports were used a lot. It was uncommon to plug in a joystick and leave it plugged in for any length of time like it is today with consoles. In particular I remember using one port or the other, depending upon the software.

[@xjet]

I suspect that PCB spark gaps are used a lot less these days of lead-free solder. Tin whiskers anyone?

[@EEVblog]

AIM54A I don't recall that, so would like to see it!

[@mausball]

The safety spark gaps I've designed have always been gold plated, both to prevent oxidation, and to mitigate the whisker issue somewhat.

[@waswestkan]

Sarcasm not indicate so here goes... Not I'm any sort of expert, but wouldn't depending on tin whiskers would be as bad on depending on a spark, or even worse? You can be sure the spark gap exists because it was purposely designed and manufactured. depending on whiskers means you're gambling they have been create & create in a manner that they had create in a manner they could provide the level of protection desired. You seemed to miss some important details in the video. Sharp point aren't desirable, while I'm no expert I understand why thee aren't. Also the spark gap should be as close to the power supply input as possible.

[@xjet]

waswestkan I wasn't suggesting that tin-whiskers be used -- I was simply pointing out that using lead-free solder without soldermask and with sharp points in the presence of a reasonable electrostatic field would be a perfect environment for the growth of tin whiskers that would (eventually) short out or reduce the gap, thus creating problems.

[@waswestkan]

mausball I could be wrong, but I thought xjet was implying one doesn't have design spark gaps, because could gamble on the creation of tin whiskers in a manner that would provide the desire protection.

[@doctordover7966]

+ EEVblog Pretty Good Video Dave. You say most IC inputs are pretty tolerant of ESD these days. You're right; here's a story of .the early 1980's when they weren't ... We built an experimental 2-D phased array which was digitally steered. Each array element had a 1K deep 8 bit CMOS FiFO on the input. We had just finished integrating & testing the digital part & were ready to start the RF tests. On that cold, dry morning, one fellow came into the area behind the antenna & peeled off his wool sweater. He was right next to the big digital backplane but did NOT touch it; didn't matter - ALL tests which had passed 2 minutes before, now failed each & every FIFO was fried.

[@alexanderwind4129]

This remembers me about electrostatic experiments in school when it was summer (exact opposite). They just did not work. But still today there are many modules in RF that have a very low ESD Tolerance of just 500V to 2 KV.. bad for the bom cost

[@DarkInsanePyro]

I'm really liking the inclusion of an example in a CAD package. Thanks!

[@kurdim4628]

Here is another viewer from the Kurdistan part of Iraq. Yes we love you Dave

[@JSaysHey]

Kurdim same from Baghdad :)

[@spiros1994]

As RimStarOrg mentions in his videos, it's easier for current to travel through air if ther is a pointing thing like that you show us on the pcbs. The round think you showed as in the cad is not so good as the pointing one.

[@calvinthedestroyer]

Loved that you went back and tried the lower setting after the spark gap had burnt.

[@barthchris1]

+calvinthedestroyer Me too. I wanted to see if it would arc from the carbon track it made. Thought for sure it would.

[@fluxoff]

Redesigning a serial terminal interface card, we added pcb spark gaps because built ones cost a fortune! Spark gaps+ Rc networks worked a treat solving our ESD problems!

[@Bbaass_TMH]

We have an insulation resistance tester at work that measures up to 200MΩ at 1000 volts. If you hold down the test button, it'll continually cause arcing if you hold the clips close enough together, and you see the measurement result jumping around. If we do that when it's attached to a faulty cable (that doesn't have a full short circuit) the resistance slowly goes up.

[@RoelofRichter]

The sharp edges arc easier than flat or round edges. The sharp edge has less surface area so the current density is higher and will thus arc easier than if the current is spread over blunt edge.

[@big0bad0brad]

You're so close to the answer, so close! But it's not current density, it's the density of the electric field. The electric field is concentrated at the tips of points and it takes less voltage to make the air there begin to conduct.

[@105d11]

Dave's pointless (pardon the pun) spark gap is actually a pretty bad idea. Or rather, he's making its arc-over reliability far worse by not concentrating the electric field on a point to encourage breakdown. The idea is not to sustain continuous arcing, but to handle short faults (under a few ms). Same reason lightning rods are pointed.

[@Empolitano]

First time I see it, funny that the circuit passed medical standards! Even though I would never rely only on that.. Breakdown voltage depends also on humidity of atmosphere around for example Cheers!

[@anikaneda]

Hi Dave, nice review! This spark gap is not for protecting the input of the device, it's for leading the voltage flashover when (for example) u have a surge of higher voltage between the primary and secondary side or as u said ESD. You can not compare it with GDT. These gaps are also often being used in parallel to each winding of a common mode choke. If u have any question in detail to this PS I could ask my college at work, that's his baby ;) Greetings from Germany // Kaneda

[@lassehoei]

In combination with a resistor it would protect the IC from ESD, wouldnt you say?

[@crispusb]

Hi Dave, I appreciate very much your work. I miss fundamental Friday :)

[@AureliusR]

InXLsisDeo He's already done LTSpice, and Afrotechmods has already done the switching power suppply basics. mikeselectricstuff has covered supercaps too.

[@robertcalkjr.8325]

Thanks Dave. I don't remember ever seeing such a crude spark-gap before on anything I have fixed.

[@1Andypro]

Great video Dave! Showing us how to design spark gaps in your PCB software was fantastic. I would love to see you incorporate more of that when you discuss layout related topics, because it's great for visual learners.

[@biturboism]

Awesome, Dave! I was wondering about this myself and searched for more info online, but came up with nothing. Thanks!

[@power-max]

Neon lamps also come in handy, there are cheap as beans, are a little more controlled, and obviously trigger at a lower voltage, and also they have hysteresis, so after igniting, the voltage can fall down 60V or so and the NE2 lamp stays on in a low resistance state.

[@0MoTheG]

Power Max The voltage will fall in any discharge or spark.

[@power-max]

0MoTheG Not necessarily. Most real world supplies w/ series resistance will experience a voltage drop, however if you incorporate a feedback loop to maintain voltage at some value and hold it constant, the change in voltage over time will be negligible, and probably not even measurable (at least very difficult). That feedback would not be hard to do with a large resistor divider and some op amps/comparators.

[@dbackscott]

I remember a line of credit card terminals I had to support. The metal casing was not grounded, and there were two traces with exposed metal mere millimeters from a support for the case. One trace went to a RAM chip, the other to a flash ROM chip. One ESD from the clerk touching the case, and it either lost all transactions for the day or bricked, depending on which gap the ESD jumped.

[@johnbeer4963]

Thankyou for this video. It was entertaining to see the spark gaps in action, and most informative to see them designed on a PCB maker.

[@88njtrigg88]

The most interesting & biggest spark gap's you'll see (inline before the gase discharge) are in commercial AM/FM transmission lines, usually set 180°from a right angle of transmission line.

[@curtisbeef]

I'd love to see more videos where you use Altium.... maybe a tips and tricks video?

[@redtails]

curtisbeef I back this up too. I love seeing experienced users tell me what they're doing with software like this

[@EEVblog]

curtisbeef I won't do a tutorial video for such an expensive piece of proprietary software. If you can afford $8K+ for Altium, then you don't need a free video from me showing you how to use it.

[@curtisbeef]

EEVblog I don't necessarily agree... I mean you do say your vblog is for "electronics engineers" etc..etc.. And its not like paying 8k is the only possible way for someone to be using altium. Students... Employees... Pirates. It is your prefered software package, no? Maybe even just some more videos on general PCB layout best practices and tricks. I enjoyed your other ones. Anyway, keep up the great work I watch all your videos.

[@EEVblog]

curtisbeef Students usually have a lab and training. Employees have employers who can (and should) pay for the Altium training course. And pirates? serious, you are going to argue that?So give me another good reason why I should do an Altium video please...

[@curtisbeef]

EEVblog I did give you another reason. More videos on general PCB layout best practices and tricks. Just a suggestion.

[@arbitrarysubpixel]

This is the greatest person on youtube, bar none.

[@vhm14u2c]

I never seen intentional spark gaps on PCBs, thanks for sharing!

[@ArduinoTronic]

With medical devices the main concern is that when a nurse rips the sheet off a bed she can generate over 300kV depending on atmospheric moisture.

[@fjs1111]

Dave: I need to correct something in this otherwise awesome video. PCB spark gaps SHOULD have a sharp point on the conductors between. The reason is sharp conductors always initiate ionization/arc BEFORE non-pointed conductors. This is the same principle as why sharp metal (Fork) creates a discharge in a microwave whereas spoons do not, or why lightning rods always come to a sharp point. It has to do with the concentration of electrical charge in point conductors. en.wikipedia.org/wiki/Lightning_rod#Rounded_versus_pointed_ends en.wikipedia.org/wiki/Electrical_breakdown#:~:text=Conductors%20that%20have%20sharp%20points,concentrated%20fields%20that%20precipitate%20breakdown.

[@electronicsNmore]

Great video.

[@dizzolve]

At 12:15 you can see carbon build up on the right jaw where it obviously had sparked before Dave started. Interesting how the smallest gap sparked until it burned further than it's adjacent gap so the nearer points took over till they burned further than the first. Yet the third never engaged - perhaps because it was already toast from previous event.

[@Qui-9]

They should have a substrate or local underlayer that is not carbon based for this...

[@LarryAllenTonar]

Fascinating. In the early 1970’s, while working on a PCB for operating room EEG equipment, I was told that due to the presence of pure oxygen and flammable anesthetic gases, free air sparks were verboten. Fancy air-seal-before-contact power plugs, etc. were required. Are free-air spark gaps up to code (allowed) on medical equipment? Does your formula represent maximum or minimum breakdown voltage estimate? Is the formula for a rounded trace end or a pointy trace end? Does air humidity have an effect (from high school Van de Graaff experiments, humidity is critical: high humidity seemed to cause discharge without a spark and a long-haired person’s hair wouldn’t flair out on a high humidity day) on pre-spark discharge? (But another comment stated that 50-100% humidity required 40/39th the voltage to cause a spark at a given distance than low humidity.) Also, the more pointed the charged surface, the more concentrated the electric field which causes a spark, thus a lower voltage is required to cause breakdown, giving better protection, unless it overlaps mains peak voltage. Another use-case for this protection is shorting-out back-EMF pulses from collapsing magnetic fields, e.g. in motors, connected to the same mains. In your design discussion, you want to short out to “earth-ground”, not digital-ground or analog-ground, right? Or, at worst, the mains “neutral”.

[@Pugglewuggle]

I'd really like to see a video or series of videos on input protection just as you suggested. It'd be great to see this for both analog and digital (high speed and low speed, busses, etc). Maybe some tidbits on protecting against impulse/ESD vs slow ramping overvoltage vs long periods of overvoltage.

[@krawutzimon]

Thanks for the video, Dave! if you need some ideas for some follow-ups: I would very much love to see you talking about general input esd protection. i know there's a lot to consider, line impedance, capacitance, inductance, response time of involved components (tvs vs zeners), so maybe some kind of hands-on walkthrough of how you personally would approach something like out-of-case interconnection between two boards involving some analog, some digital (serial, maybe rs232 or 485, perhaps i2c or spi) and power lines. I know there are a lot of specialized parts from different manufacturers on the market, but how to choose between them (and are they really needed or can you do similar things with jelly bean components?)

[@Borednesss]

I've always wanted to see you play with high voltage.. I guess this is as close as it'll get =P

Ummm... better than nothing. I remember a saying about "better than nothing" :p But that wasn't what I wanted to comment about.. You reminded me about when I was a little kid and I was marveling at the spark gaps wondering what the hell are hose. They looked like they would make great spark gaps :))

[@AlBarathur]

Wow, my mind is blown. I always thought that the lower the air pressure, the lower conductivity would be, requiring thus greater voltage to produce arcing. Damn you now I must research XD

[@alext9067]

Think Crooke's tube.

[@88njtrigg88]

Just like filtering dust or particulate material from water or air (cyclone air filter OZ invention) make it turn for ejection.

[@Qui-9]

Gasses are complete insulators, and compressing more insulating material in a space also compresses the breakdown gradient. Think how oil has higher density and such a high breakdown voltage per mm. It's a similar situation.

[@jonathanseyfert8256]

To a point. A vacuum is a pretty good insulator, much better than air. But a high pressure insulating gas or liquid is better yet.

[@badatcad]

I always learn a lot on this channel, thanks as always for youtubing

[@justrandomtech521]

Love your channel, every video is interesting and very informal.

[@Arek_R.]

Thanks Dave! I saw this thing(spark gaps) many times but didnt thinking its for protection :D I see this very often on smps(eg. smartphone 5V charger/power supply).

[@Kenneth_H_Olsen]

Beautiful . The carbon that is formed after a discharge in a solid, is conductive. Carbon residue is a good conductor. I agree that it would be better practice to remove the solid material from between the spark gap terminals . That is to have an "air gap" .

[@greenpogo]

As always, great video - and thanks!

[@CoolAsFreya]

You know you can light a candle with high voltage discharge I took a transformer out of a plasma globe and hooked it up to usb changer 5V 3A and got sparks up to 5 mm away, and if the neutral wire (ground) is touched against the end of the wick snd the placed to spark anywhere else along the wick (yes apparently paraffin is conductive) and with some luck it'll light

[@toolhog10]

I recall from physics many years ago the discharge was 3 million volts/meter in air... so that little formula seems good, not sure what the + 1350 is about though, maybe a humidity factor?

[@EEVblog]

***** It was likely a number that just popped out of the experimental data. I don't know the history behind that formula, if anyone does I'd love to see it.

[@LPFthings]

I think there's just so many variables it'd be difficult to get a totally reliable formula for it, you generally just have to go with the worst case scenario. For example humidity is probably just as, if not more important in that calculation than pressure, yet it's not included. Sparks will also travel a little more easily over the surface of a material than they would in free air, but of course that's going to vary with the material also.

[@power-max]

EEVblog I actually learned from when I was building my singing arc project and tesla coils that a spark is 1000V/mm. Very simple but not at _all_ accurate. Seems to be grossly underestimating too! That _must_ mean my singing arc is actually producing 30KV instead of only 10KV!!! You can actually make small arcs with as little as 18V at 1A, they are hard to see, but nonetheless it is possible!

[@EEVblog]

LPFthings I have never seen a formula with humidity added. I agree that it's very important, maybe even the most important. But the fact remains that no such formula exists (to my knowledge), and it's not like I can just make something up.

[@EEVblog]

Power Max You can't get an arc-over spark from purely 18V, it's just not possible. When you get sparks from high current stuff, other things are at play than the supply voltage.

[@kyoudaiken]

"...argue until the cows come home". I laughed so hard. I'm German and didn't know about this phrase. Nice one! :D PS: I think it was a phrase of farmers back in the day, where a long time was described to be until the evening where the cows came home.

[@dumbo800]

en.wiktionary.org/wiki/until_the_cows_come_home

[@bigjohnsona2]

Anyone else just like seeing some High Voltage stuff? A part of me just wanted to see something explode. nothing better then venting some frustration with HV and some unsuspecting caps :-) love it. Id love to see a random video of component failure points too. :-)

[@MilanKarakas]

At your video, after 2:54, there are few red marked spark gaps. I found that red marked spark gaps (for old telephone company line protection) had radioactive paint on the inner wals of the gap. The one marked black has no such radioactive thing inside, but only red one. Check it if you have chance (GM counter, gamma spectroscopy probe). It is such low activity, that one should be broken to show onto GM counter.

[@shaunclarke3]

Is the change in breakdown voltage with the change in altitude the reason some equipment is specified to only operate within certain altitudes? Or are there other reasons as well?

[@ashrasmun1]

It's funny, because I don't really care about content of your videos, google recommended your channel quite a while ago to me and guess what. I tend to come here, listen to you talking for few minutes while you are elaborating something smart about electronics and I just go away after that time. You are strangely appealing though LOL

[@sprybug]

I believe the audio issues you were having at the end had to do with the software you were using. Were you recording the audio on the same computer at the same time you were video recording? I noticed that every time you would move or rotate the image on the screen, there would be a popping sound.

[@lightaces]

What I would like to see now is a Fundamentals Friday on better ways to handle input protection!

[@marvelknight-dx7sp]

Thanks for doing this video, had a SMPS but they put those on the common mode choke.

[@ArcadeGames]

I have seen some ceramic capacitors with spark gaps on them, it kind of looks like a chunk has been cut out of them with a tiny gap. Does this function the same way so the capacitor doesn't blow?

[@ServerFarmGuy]

Thanks, I was just looking into spark gaps for a design and this video showed up on my feed :D

[@MrSparker95]

It would be interesting to see how flux left after soldering affects the spark voltage.

[@DjVortex-w]

I know that it's just a back-of-the-envelope equation, but the lack of proper units still bothers me, so I'm trying to figure out what the units must be on the right-hand side of the equation to equal Volts. A Volt is defined in SI base units as (kg * m^2) / (A * s^3). Thus the right-hand side must produce that. The 1350 must be in Volts because it's a summation. Atmospheric pressure is measured in Pascals, which is defined as kg / (m * s^2). This already has many of the same units, so it's somewhat promising, but there's still quite a way to go. Atmospheric pressure times distance gives us kg/s^2, which means that the unit for the 3000 factor must be m^2 / (A * s). Now I need to figure out what kind of unit this is, if it's anything. m^2 is obviously the unit of area, and A*s is the unit of electric charge, ie. Coulomb. I can't find anything that corresponds to this. The closest thing I can find is the unit of electric displacement field, which is the inverse, ie. C/m^2. If this is indeed the accurate formula, the unit for the 3000 is either strange, or a combination of units I can't figure out. Anyway, the most accurate I can come up with is: V = 3000(m^2/C) * Pressure(Pa) * distance(mm) + 1350(V).

[@WeAreGRID]

Ah, but your error lies not in your research, but in your assumption that the 3000 NEEDS a unit at all, for in fact, it is an arbitrary number! You need no unit, because it is only there to show reduced voltage at less than one atmosphere, , just as the 1350 is arbitrary as well, it is the V on the left that is the actual unit for the final number, not the numbers on the right, which are all distance and pressure with arbitrary values for the math, because as he showed, 0.2 gives you about 2000 volts, and it took over 2k volts to spark the spark gap that was presumably close to 0.2.

[@DjVortex-w]

WeAreGRID It needs a unit because both air pressure and distance have already units, and the units on the left hand side of the equation must match the ones on the right hand side. The only way to make them match is to add the necessary units to the 3000 factor. (This in itself would be very useful _if_ the equation is correct, because it would tell us what the 3000 represents, which would be both interesting and useful information.)

[@pesshau6508]

WarpRulez The "3000" figure represents the breakdown voltage for air at 1atm (30kV / cm). He has lowered it by one order of magnitude, since he is working with distances in millimeters.

[@DjVortex-w]

pesshau I'm not so much interested in its magnitude as in its units.

[@mrlazda]

You are over complicating, unit for dielectric strength is V/m (usually it is used as MV/m) and for gases dielectric strength varies with pressure so they probably just put pressure in constant so it is something like this V/(m*Pa). If you are interested in how exactly it works google Paschen's law, but in quick formula for it is Vb=(a*p*d)/(ln(p*d)+b) where Vb is breakdown voltage in Volts, p is the pressure in Atmospheres, d is the gap distance in meters, a and b are constants based on composition of the gas and unit for a is V/(atm·m) and b have no units.

[@sparkyprojects]

They used to use neon lamps as spark gaps/surge arresters many years ago Also, points are best for spark gaps, that's why lightning rods are pointed, the capacitance builds up on the surface area mass, and is concentrated on the point, in fact the GDT's symbol has points, so essentially a neon with pointed electrodes i think. The pcb spark gap was probably for short static arcs, rather than long duration, so wouldn't have built up so much carbon in a short time

[@justincomisk]

Great video. Around 10 min I can see your reflection in that excellent solder joint. Ha!

[@ForViewingOnly]

Interesting video Dave. It would be good to find out more about the TVS and GDT devices if you fancied doing a video on them. I was looking into TVS devices recently because I wanted to do an Arduino project in a classic car and read that car 12v power can vary wildly when the ignition is turned off (load dumping). I was also interested in what happens in a mains surge protector like my Belkin Surgemaster 8 way extension lead. There used to be a huge flash of arcing whenever I plugged in my laptop PSU (maybe inrush current sent to earth?). Eventually, after about a year, there was a massive pop when I plugged the PSU into it and the mains tripped. There was a smell of burning from the Surgemaster but no damage to the laptop PSU, and from that day on there was no more arcing when I plugged in the laptop PSU. I guess the GDT gave up the ghost?

[@pintokitkat]

Another good video. So now I understand that the spark gap protects inputs/components from over 1500v, but surely they can be killed by less than this. So as protection solutions go, it's like wearing rubber soled shoes to stop you getting struck by lightening. Possibly effective, but no good at protecting yourself from more likely threats.

[@dwDragon88]

Hi Dave, a couple questions. Why does the transient choose the high impedance path through air instead of simply propagating down the trace. I understand the impedance of the gap changes drastically once the spark forms, but it seems that might take some time, during which the copper trace would be more inviting. Second question, are there considerations for the impedance of the trace after the gap, to ensure no reflection of energy? Or perhaps current handling capability, say for when someone shocks the hell out of it with an insulation tester? :) Thanks.

[@mikeguitar9769]

It might be less likely to fail from a carbonization-short if it didn't have any PCB material between the 2 points. This could be done with an unplated hole or a notch in the board edge. Although due to machining limitations, the gap might have to be larger, so the breakdown voltage might end up being slightly higher.

[@dLLund]

lovely demonstration & explanation. thank you for posting. take care & stay safe.

[@anthonyvolkman2338]

When I design a board with these I normally place a small grounded box over the spark gap to prevent noise propagating from it causing issues.

[@etmax1]

The problem with the spark gaps shown is that the tracking resistance of the PCB is lower than in air and changes over time. Proper gaps have a slot/hole punched between the 2 poles of the spark gap. They still suffer from dust that is drawn into the gap via the electrical field.

[@ralfschooneveld3186]

Great info Dave, thanks.

[@blackIce504]

Dave thanks for sharing your rounded spark gap Idea, i really like it and it makes sense to do it that way plus its simple, i am not 100% sure on how this works but Asus "board maker" use an ESD protection on there boards but i believe they use some chip to do it, any idea's of what they could be using as it offers another option.

[@11danzizzle11]

Great description, I was wondering it myself!

[@ajledezma2147]

Didn't he sound like a football announcer during that second spark gap test? 😂

[@davidgrenet]

I know this is an old video but on the medical plugpack board it looks like the spark gap is there to protect against HV from the unpopulated Y cap (C21) that bridges the primary and secondary

[@gelectoelectronics2032]

Dave, can you please do the tamper proof device that you received in a mailbag a while ago.

[@peckelhaze6934]

Another interesting topic.

[@dariodraiman6339]

I guess it would be OK to leave out the solder mask if the spark gap is in the top layer, but I wonder if the lack of solder mask in the gap won't make both contacts to be joined by solder with the wave soldering process. The few spark gaps I saw in commercial boards had the solder mask in the gap between the exposed and covered by solder pointing contacts.

[@rockerdude02]

hey there dave I do love your content, I was just curious, the program you use for your pcb layout looks really good, what is it? I sure would love to get my hands on it for possible future projects!

[@MrRoyzalis]

Very good video- one of your best. When you said you may do a separate video on protection design/consideration and when to use gas discharge tubes vs MOVs vs TSVs etc, Please do. I could never figure out why some designs use one protection scheme vs another. Been a mystery to me.

[@TonyLing]

Wow! sacred geometry man!

[@zynthos9]

What does the carbonization do? Reduce the resistance of the path so it's easier to spark over?

[@Radonatos]

Imo the carbon layer isn't thick enough to shorten the air-gap so it would make any electrical difference - I assume the carbon layer is just mentioned because it is any indicator that the spark burns part of the surrounding pcb material.

[@chenbh4029]

Judging by the fact the the spark moves around when the carbonization gets worse, I assume that it raises the resistance. We can see clearly that after some time on the left peek of the spark gap, when the resistance is fairly equal to the middle one - the spark moves back and forth. Not a professional, maybe somebody can enlighten us.

[@dalebertdurando1496]

Carbon can be quite conductive (see carbon resistors) lowering the resistance path. In this video I believe the points burned off before the carbon layer became significant.

[@keithminchin1817]

An input design tutorial would be awesome 👏

[@QoraxAudio]

If the tinned surface of the spark gap oxidizes because of humid air, does it still work as intended?

[@avikazen4439]

About how much current is that insulation tester outputting during the continuous arc? You should do a teardown on that!

[@azharAD]

Thanks Dave! that was helpful :)

[@JohnDoe-qx3zs]

Wonder if a spark gap in parallel with a MOV (e.g. 250V) would be better than the MOV alone? Or is it better to keep the trace distance high on the mains side and rely on the MOV alone for overvoltage protection?

[@himanshuankleshwaria785]

Really nice understanding given in this video. Thanks Dave. I am little bit worried about aging effect due to environmental effects (Humidity, dust etc) and also due to spark gap, carbonization occur on those exposed patterns and can it affect our product warranty/life cycle? What will be the MTBF analysis for the same?

[@vargasmir]

I would love a video about input protection! Tips and tricks!

[@nineball039]

I fixed regular analog TV's which had PCB spark gaps, typically near the flyback.They appeared to be for short circuit protection safety not ESD protection.

[@jessstuart7495]

The high electric fields near the sharp points of spark-gaps can cause dirt or contaminants to collect in the gap. Unless you are protecting isolation devices, from breakdown for one side to the other, spark-gaps are pretty useless. Most devices would be damaged by the larger number of 100V to 1000V ESD hits that the spark-gap would offer no protection against. I like to use TVS diodes.

[@robinheijblom2929]

I wished they did a little bit of protection in consumer ADSL modems and VoIP Gateways. Almost all phone ports on consumer crap in rural area's dies during a lightning storm. And it's always the phone ports. Ethernet usually survives because of the 1,5kV isolation the HF transformers have. The only place I've seen Ethernet causing problems is with ungrounded electric installations.

[@zacharykrawczyk3942]

Would the spark gap work in a vacuum? I would assume not, or at least it would take significantly more voltage.

[@kkobayashi1]

+Zachary Krawczyk As the empirical formula shown here (5:15) shows, the lower the air pressure, the less voltage it takes to create a spark. But this relationship breaks down at very low air pressure; below a certain pressure, the trend reverses: the closer to perfect vacuum, the higher the voltage it takes to create a spark. Look up "Paschen's Law" for more detail.

[@bigboom378]

Awesome! What software were you using for the pcb design? 3D view looks really handy

[@iborschtsch]

bigboom378 It's Altium Designer.

[@skson5838]

Could this formula be used to estimate the max voltage that a air variable capacitor could handle before arcing over?