I had lighting strike the radio antenna of a cellular remote oil well monitor that I had built a long time ago. It skipped right over the lightning arrestor on the antenna cable and blew the ground track off the entire circuit, through the cellular modem, through the serial cable connecting the modem to the PLC, through the PLC circuit board, through the ground wire and into the electrical circuit on the engine skid that was supplying power to the automation panel. Was a hell of a thing. The ENTIRE circuit from top to bottom vaporized. There was nothing left of the tracks at all.
@@greatscottlab Well, it didn't happen while I was standing next to it. I got a call saying the install wasn't powering on. Remote trouble shooting didn't turn up anything so I had them pull the unit and send it back to me. I took it apart and found dust and vaporized copper.
What brand of PLC? That's not something that you see happen to a PLC very often. The problem with voltage protection is that it's like walking under the hoover dam with your umbrella open. You should be good, right?
I repair control circuits for some industrial equipment, and they regularly use both a TVS diode and GDT in parallel on each line. The TVS diode handles the lower voltage transients, while the GDT protects against higher incidents like lightning surges. It works very well, and it's easy to test and cheap to replace if it does blow.
Hearing the way each device worked that was my immediate thought "can you use them in parallel to protect against a wider range of voltages?". Good to hear I was on the right track.
@@lostsauce0 So, electricity always takes the path of least resistance (more or less). Think of each of these devices as a conditional short, it's normally not shorted to ground, but under specific sets of voltages it triggers a short, at which point IT becomes the path of least resistance, and so the voltage spike will take that path rather than flowing through the higher resistance (sensitive) circuit. Each device has a limited range of voltages and response times it's good for, past a certain point it will fry (or under a certain point it won't trigger), stop conducting, and then you're back to the circuit you're trying to protect being the path of least resistance which you don't want. By putting a bunch of them in parallel you're covering a wider range of potential voltages, and even if multiple of them get fried, it's cheaper and easier to replace these components than the rest of the circuit.
a TVS failing would mean that it does not conduct to ground, so a parallell device that also conducts to ground during a high voltage transient, would take over. Or what did you mean by "take the TVS line"?
Varistors are way slower than TVS diodes and aren't fit for fast transients such as ESD & EFT. Another thing to remember is Kirchoff's law, any current pulse will return to its source. Therefore instead of trying to suppress a pulse it is often more practical to divert it from the sensitive components using spark gaps etc
Seen a park gap do it's job in practice once (that is, after the fact not during it), on my TV's power supply. The comment I talk about it in case you wanna read: ruclips.net/video/MzxBBXpgwrE/видео.html&lc=UgydNWuE3TvVqRYRkm94AaABAg The board does have an MOV there too I assume to protect against lower voltage surges instead of just lighting strikes or such high voltage stuff. Those two in combination sure did their job.
Spark gaps still leave residual overvoltage which needs to be handled. They are also not suitable for a low transient range. Neither are repeatable or reliable. Mov and tvs coupled with esd suppressor is the way to go.
Here's a video idea: Signal integrity When and how to use shielded wires, coaxial wires, differential twisted pairs, ect and avoiding issues like ground loops (like connecting the shield only on one end) ect
@@iamdarkyoshi although I think Scott's area of expertise is high voltages and mains stuff? so he might not be the best person for this as much as I love the way he explains things.
Based on his stair light PCB layout, running small-signal traces parallel with the PWM-regulated unfiltered LED power traces with the two LED power traces routed about as far apart from each other as possible for maximum EMI, I doubt Scott knows anything about signal integrity. He'd certainly gain a lot from at least looking into the basics.
When designing automotive type electronics that had to meet EMC standards, the electronics had to withstand ESD discharges up to 15kV. Something as simple as a 1000pF cap at the input was often sufficient protection. Signals that couldn't tolerate that much capacitance were good candidates for a TVS diode. For power line transients, which are much longer in duration and contain much more energy, a larger TVS was usually the solution (perhaps in a SMB package). The datasheet will tell you how much energy a TVS package can handle. MOV's weren't used, and I get the feeling that they are more oriented towards mains power where lightning strikes are a risk.
"Doesn't happen often that lightning strikes near you." I live in Singapore and I've had a thunderstorm every afternoon for the last three days. Yesterday the lightning strike was so close you could hear the crackle before the bang. But at least the humidity stops you getting static shocks from the carpets in the office. Something you might be able to advise on is the problem that I have, which is when there is a lightning strike nearby my RCD trips, none of my neighbours seem to be affected. I've always thought it was because I have an induction hob and the coil is picking up the EMP from the lightning and creating a current flow in the neutral line, as normally the hob is switched off at a wall switch, (which I assume only isolates the live), and hence causing a differential L to N current flow. You could probably have used an electric fencing unit to create a known and constant voltage, or a car ignition circuit.
@@greatscottlab Yeah when I lived in SE Asia my UPS would last a few months before bulging from an overworked battery due to a lightning strike They never lasted longer than a few months. Plus, aside from Singapore, a lot of countries in SE Asia don't use ground. I added a ground for a house we built, not sure how reliable it is for a big short or overvoltage, but at least it seems to work according to the multimeter.
@@greatscottlab Nearly all power lines in the country (not that it's very big) are buried so I don't think that over voltage events occur very often. I've never had of over voltage protectors on domestic distribution boards.
One time lightning struck pretty close to my Mom's house. My brother got a small shock from the wired headphones he was wearing (plugged into a PC which was grounded). It also tripped a GFCI. Not sure why it tripped, I just chocked it up to high voltage doing funky things.
I lived in Florida, USA for a while. Lightning happened every other day. I lost a computer motherboard to a cable modem conducting a lighting strike from the cable system into it and thine through it into the computer because the lightning arrest system on the building failed. My next computer had optically isolated I/O ports and my replacement cable modem's coax went through a surge suppressor, as did the LAN cable that came off it, isolating any future lighting strikes from the network and my new computer with three layers of protection.
As a notebook repairman, I am very familiar with diode arrays, it is very common to find them protecting the USB, HDMI and other connectors on the boards, many versions of notebooks for home use differ from the corporate versions in that although they have the tracks, They do not have the chip installed, this small saving makes the notebook much more delicate, plus now all those connections go directly to the SOC, system on chip, practically killing the processor in the event of any incident.
Man, I've been a fan of yours long enough to know your real name. Haha, I think it's about 10 years that I have been a subscriber (when you still responded to messages on facebook within 24hrs, haha). I started watching you when I was in high school at 14 years old. Actually, you played a very big role in me studying BEng Electronic Engineering. I am now 24 working as an electronic engineer. I recently did an ADC protection design for one of our products where I used TVS diodes. I actually learnt what those are from you years ago. If I ever find myself in Germany, I'd really like to have lunch with you. I don't think you realise how much of an effect your videos have on some individuals. Much love from South Africa.
I'm familiar with GDT as a "Comgap". As in communications gap. Back in the landline days, they were a required component (in Canada at least) across tip and ring on a phone line. Every modem had one for example. It was to supress lightning strike level voltages from reaching the consumer. If it happened to protect a circuit, that was just a side effect. They were often used in combination with a MOV. Old-school rotary phones generally didn't need them as the ringer coil would take the punishment instead.
A thing I like to do to protect low current inputs is a high value series resistor and a capacitor to gnd behind it. The resistor limits the current and the capacitor just eats the charge, so there is nothing left that the internal protection can't handle. Great for button inputs, as this also covers debouncing or acts as a low pass for "slow" analog inputs
Thank you for the video! The MOV surprised me in this experiement, I didn't expect it will protect against sparks due to its slower reaction. I designed an off-line converter with an LNK-304 and lacked the mov. When my father plugged in a washing machine to the nearest socket the IC blown away. Never lack the MOV!
9:37 Quick correction, that isn't a thermal fuse, it's just a standard electrical fuse. A thermal fuse is tripped when the temperature of a component gets too high. For example on motors to stop the windings from burning, electric heaters to stop your house from burning, hotplates to stop the hotplate from burning and so on. The resettable thermal fuses are usually found on components like electric heaters, hotplates, dehumidifiers and other devices, though are a first line of defense, they reset after the temperature drops (though microcontrollers have taken their place in some applications) but they'll also be found as second line of defense devices on industrial electronics (or higher end consumer stuff) where you have a uC and temperature probe (thermistor or thermocouple) monitoring the temp, but if the uC crashes, the resettable thermal fuse cuts the circuit, then if that fails a thermal fuse (attached using thermal paste) blows, giving the device 2 chances to recover and one chance to need fairly simple repairs.
Hi GreatScott. Good video! I would like to add a few things: 9.44 the fuse in series with a varistor that you have drawn is not a thermal fuse, but an ordinary fuse, which would open when a varistor is damaged due to a high surge current. A thermal fuse is usually added in series to the varistor, so that if a varistor gets damaged and it conducts some small current it gets slowly hot - in that case a thermal fuse would open. some manufacturers even include a themal fuse in a varisto case, so the varistor has 3 pins. But you mentioned that the fuse blows when¸ an overvoltage event takes place, which is wrong. this would be a terrible design. actually we have to choose fuses, which have a high I^2t rating so that they DON'T blow during expected surge events. Your testing is only with ESD pulses, for which TVS diodes are best. Varistors and GDTs are used for surge pulses (indirect lightning strikes and inductive overvoltage events) which contain much more power. the 8/20us pulse is a typical indirect surge testing shape, which has 8 us of a rise time and 20 us of duration. there is also 10/1000us which has much more energy. ESD protection and surge protection are very different things. Usually for surge protection we use several of these protection components. for example we start with GDT, then varistor and then with the TVS diode. All three seperated with an inductor. Current and voltage on the device gets smaller with every component. 11.46 this is a thermal fuse and it opens when the varistor gets hot. the failure mode of a varistor is that with every surge pulse it conducts more leakage current. and when this leakage current is smaller than the current fuse in series, the varistor gets very hot and it ca cause a fire. that is why standards determine there has to be a thermal fuse there, which opens when varistor gets too hot. then it is time to change the protection module. Note that an ESD pulse has a rise time of 1 ns and duration of approx 50 ns. But a surge pulse has a rise time of 8 us and duration of 20 us or more. The problem with varistors is that the current through the device is increasing quite slowly with voltage rise. But a TVS diode's current increases much quicker with increasing voltage. at 7.58 you drew the correct V/I characteristic. but when you dig in a datasheet you see that the current rise in varistors is very slow. For example the varistor you showed MOV-20D681K conducts 1000A at 1600V and 1A at 800V. so it need additional 800V to get from 1A to 1000A. Usually a varistor is not used as a ESD protection device. I think it protected your circuit due to varistors parasitic capacitance, which conducts ESD current and not really due to the varistor voltage limiting.
This might be one of the most practical ESD - TVS based circuit protection video! I mean there are so many nuances of using TVS diode, routing them to the ground plane, ensuring low inductance to ground etc. This video gave me practical insight into the core fact that TVS diodes convert the ESD events into heat - boom seeing it practically save the circuit - made my day.
I wonder how effective spark gaps are, given they're pretty much free to add to a PCB. Our TV's power supply got quite a few of those, and after a lighting strike nearby, the cable TV box burned and it made the favor of killing the TV's motherboard through HDMI, along with the Switch dock on that same TV. Replacing the motherboard made it work again, but after a while the power supply also went bad, and I could see on the mains pins a tiny welding like dot on the metal casing (clearly an arc mark), along with a missing spike on the spark gap PCB track. It was already more than 10 years old at this point and it still survived that lighting strike.
Spark gaps definitely have a role in protecting electronic circuits from discharges of energy. In the world of color CRT displays, it was very common to have spark gaps integrated into the CRT connector socket as part of the protection for the video output circuits. The socket would also use a thin but wide grounding “wire” (strap) that would typically get connected to the outside coating on the CRT funnel, to have a low impedance ground return for any arcing that bridged the spark gaps.
When using a MOV on high side of power supply, you need to take account their failure mode: slowly conductive. So either couple them with thermal fuse close to them, or a combo MOV+ThFuse, or you can end up with a fire event in case of repetitive surges.
I would like to add some observations and know more. I have seen many SMPS with MOV for protection. First failed component I always saw is MOV with dark discolored and often cracked like 2 separate disks for each terminal. It happened over a few years in India and often with power surges. Specifically they open circuit after 3-5 power surges. Which is around 2 months of time. I tried observing this because another SMPS had a fuse installed instead of MOV with almost same rating. These SMPS were used to power small LED panels for a business display sign with a LED driver always.
I built a home energy monitor using CT clamps, and I learned they incorporate a TVS diode inside the clamp. Because without any load on the other end of the clamp, the 100A of current on my main line running through the core can generate theoretically millions of volts, and the whole thing just explodes in a puff of smoke without even touching any metal wires. I learned about their existence because I didn't understand any of this and unplugged my CT clamps from their load while current was running through them. The TVS diodes did their job, shunting the voltage to I think 22 volts, and making a hell of a racket, both noise and vibration, while they did so. Both clamps were unharmed and continue working accurately to this day.
As far as your microcontroller supply line goes, simply adding the 100n decoupling cap that should be there anyway will pretty much protect it against ESD. The pulse is high voltage but very low energy (standard ESD model often uses 150pF) and thus the decoupling cap will act as a capacitive divider and "eats" the ESD pulse.
Nice video, very informative. There is one thing I was missing. When adding such a protection to your pcb, you should pay attention to your layout and routing. You must route your wire through the pads of the tvs/mov to get the best results. And ideally they should be close to the connector.
We usually work in solar systems and we use surges all the time in both AC (grid input) and DC (pv input) but not these types it is similar to the red one but a glass insulator breaks open and allows everything to flow to ground instead of dissipating as heat and it is connected in parallel with the circuit. AGAIN IT IS USED FOR DIRECT LIGHTNING STRIKES AND CAN ONLY POP ONCE BEFORE DISPOSAL.
I have experienced two direct strikes on antennas fitted with lightning arresters and direct grounding wiring. The equipment survived without damage but proper protection is a bulky affair with thick cabling required to provide a low impedance path to ground. All the power circuit protection devices tripped, but no damage was done and everything worked after the tripped breakers were reset.
The GDT can create a lot of interference and scramble sensitive logic circuits digital logic since when the tube fires, the voltage across it collapses very quickly to a lower voltage, creating EMC surges. TVS diodes and MOVs tend to create much less EMC interference because they clamp the voltage and don't exhibit the voltage collapsing effect. So whenever we have to use a GDT (for example, when extremely low leakage current and stray capacitance is critical), we have to incorporate other protective devices in such a way to minimize the EMC surges.
The main reason the MOV is protected by a fuse in series is not because the fuse will blow in the case of an overvoltage event. As the MOV starts to age due to surges, its resistance decreases, which in turn increases the current flow through the MOV, leading it to heat up and you end up in a vicious circle that will end in the MOV bursting or cathing fire and the circuit, without the fuse, drawing way more amps than anticipated. In order to protect that from happening, the MOVs are now protected by a fuse.
Increasing the gap voltage doesn't change the voltage applied to the board by much since most of the extra voltage simply gets lost in the arc. It does increase the peak discharge current though. As for why the MOV manages to save the circuit despite having much higher voltage, that is down to your generator having pFs of output capacitance discharging into the MOV's nF of parasitic capacitance, forming a capacitive voltage divider that drags the discharge voltage low enough to save the input pin from excess ESD current. Also, IO pins have their own internal protection from the output FETs' body diodes that can usually handle human-body-model ESD as-is as long as the chip has a path from power+ground pins (depending on ESD polarity) to actual ground. Chips on an assembled grounded PCB can be surprisingly difficult to kill as Linus demonstrated in his Electroboom special.
Good video. I want also mention, for low speed input lines often simple ceramic capacitors are used to protect the input from ESD loads. Usually 10 to 100nF are used.
This video is really helpful. I suffered a power spike a couple years ago when the transformer outside my house blew up. lost a few appliances, namely my robot vacuum, looking at the board there were large scorch marks and some exploded components, after some research I discovered it was a varistor. easy enough to replace and still working to this day
At 6:12 there is a mistake: increasing the air gap doesn't increase the voltage across the board but only the voltage across the air gap, doing this just increases resistance and reduces the current.
TVS packs are the electronics tinkers best friend!! In many years since I first used one I do not think any of my projects have ever been without one. Though I have moved in the last few months to a new location there are so many things they want me to do here. First was the safety light in the stairs and then the active emergency LED lights in various rooms that are going through a minor upgrade. But the use of the TVS ane other types of protection devices is common in this lab. Peace
I used some components like the GDT to protect a lora antenna coax from high transient voltages, probably will not know if it really works until it doesn't. I think it was suggested to replace them occasionally as repeated spikes can make them fail. I imagine that a passing thunderstorm could induce significant spikes even without a lightning strike. I really appreciate the information you provide, Thank you.
We used to add MOVs to anything with an inductive load (PLC control systems working at 24VDC). Even contactors or solenoid valves can be a problem. Inductive spikes were a particular problem with high speed counters or servo-motors which use incremental encoders. An extra count here and there can really foul things up for the position of a pick and place robot.
Fun fact: some CRT display video output amplifiers used neon bulbs to protect against the CRT arcing through the video output electronics. Quite like a GDT as shown in this video, but you could see them flash occasionally while in action.
Thank you for the deep-dive and clarifying these parts. What is the best practice for TVS application in a digital circuit? Just on the power rails? Power rails and any connections that touch the outside of the product housing? What about cap touch LCDs that get touched (obviously) - their power rail too?
You can make a zenner diode go there, you can even create a zener diode serial array. This way you have a full wave antenna that bounces until it adds up to the next zoltage. So you wrap up full wave zeener diodes inside the blade antennta, at each stage, well it is mhz, so the audio is much louder in real time.
I've seen real weird stuff happen with lightning strikes, I'm sure it could be explained but oddness like a garden patio covered in tiny copper balls which were the leftovers of the coax from a direct strike on a TV antenna but the TV itself worked just fine, pretty much every piece of electronics in the house next door was dead, there's just so much power that it's incredibly difficult to defend against. Static is easier but it's always good to know that MOVs, VDRs and transient suppression devices like those tiny diodes degrade every time they're 'triggered', it's a real common fault for them to fail on USB ports in handheld devices for instance
Thank you for this video, i love almost every video Great Scott produces. MOVs are commonly used with a few other components in AC power supplies such as class Y capacitors. I would love to see these covered in similar videos. Please and thank you.
Are you sure about the annotation at 2:49? I've always heard OL as "Open Line", which also makes more sense than "overload", since nothing is overloaded, and two probes with no conductivity between them definitely constitutes an open line.
I've used MOVs before, but didn't know about TVSs. Thanks. I was wondering about your estimation of generator voltages. When I did experiments with other static generators, we estimated our gap voltages at 10kV/cm. 4 inches was 100kV. I'm not sure which is more accurate, yours or ours. Of course we were in a fairly dry climate. Also, did you notice a change in spark frequency when you increased the distance (voltage)? It seems to me that it should take longer to to build enough charge to jump a larger gap, implying that the frequency should go down. Did you notice that phenomenon?
I got the kV/cm rating back in university. It is also just guessing because it depends on lots of factors. And yes; the bigger the distance, the longer it took to charge up.
Thanks, sometimes I think you are looking over my shoulder with the timeliness of your videos. Right now, I was designing a circuit that needed this type of protection, and here you are with helpful advice.😀
@GreatScott -- I don't know how deep on this topic you're interested in going, but -- in light of your summary at 10:31, it would be interesting to explore how to use these components in combinations. (Along with "spark gap" which is essentially DIY GDT.) It might seem using these in some combination might handle multiple over-voltage/arc scenarios, but does the right protector step in at the right time or voltage to avoid the other protector getting overstressed?
Varistor are common in circuit specially in smps but there is problem. If there High enough voltage it will arc between the pads of circuit. I have literally fried 3 smps in similar way. When I opened it up the copper pad were the witness. So tvs diodes are perfect for such applications.
It probably should be mentioned that MOVs have a limited number of "hits" they can take. Or to put it another way, they get less effective each time they are activated. It would be very informative and intuition-building to see you demonstrate that effect!
You could say your catchphrases in German. This would be your really special feature, and truly authentic too! Thank you for what you are doing! Keep doing it, man
The cheaper one for solar should actually handle more current. It uses a multi layer carbon spark gap with movs to equalize the voltage across the disks via their capacitance. Mov protection is good for lower let through voltage but it is not going to handle current as well as a spark gap. Only issue with spark gaps / GDT is their follow on current, but you need a high current to maintain the arc, your solar wont produce that much current.
I dicovered TVS and varistors while reasearching for a DIY TIG welder, they seem to be used at the output of the Diode rectifier to avoid the HF/HV starter from blowing the diodes
Everyone should take a look at replacing all of these with the new devices TI has come up with. They are meant as TVS replacements so are named like TVS1400, TVS1401 etc. Where most protection devices are good for at best 100 hits these are tested at 5000 and still work🤯 They are especially good for automotive devices where repeated surges are normally experienced on a nearly daily basis.
Also do a web search for "combined SMA and TI TVS footprint" for an idea how they can be added into an existing design to allow easy transition without a complete redesign being needed.
FWIW, there are a lot of tutorials on how to add these to power strips, like one that your computer system is connected to. The better 'strips' have these, however seems to be pretty easy to add higher spec'd ones to your existing ones pretty cheaply. Wonder if there is a 'standard' which incorporates both the MOV and TSV in series to eke out most protection...
About 10 years ago I accidentally connected a 12v dc circuit to 240v ac, when it didn't work I soon realised my mistake and connected it back to the proper 12v dc again and it worked. I've never quite understood how the 240v didn't fry it. that same board is still working to this day.
...Re insulation tester being too safe (3'12") ... How about using insulation tester to charge a high voltage capacitor (maybe even your Leydon jar???), then use a switch (or flying lead) to connect your test circuit board? ( i.e. only connect test circuit once the capacitor voltage is at the desired destructive level) ...A neon in series might also add some visual effect?
When a MOV has been severely or repetitively compromised it deadrigates to a shorted component. This depending on its design location can cascade circuit damage.
MOV or TVS diode are not that different. MOV is a sintered material and has the same diodes but just an enormous amount of them in parallel and in series. If these are protecting they can degrade bit by bit until they short or break. The TVS diode will be gone in one single overload. The GDT's are typically used to protect against lightning and always combined with THS or MOV. ( Which is the reason it failed to protect your controller ) In old TV's there was a component which just provided a spark gap in air. just 2 wires 0,3 mm apart, serving as a GDT with the gas being air. Worth being mentioned is clipping diodes in a digital input circuit. Often integrated in the IC but if not, it can be placed externally. If lightning strikes directly, you can have all the protection possible : Nothing will save your circuit.
9:25 ".... dies immediately...." (shows a working circuit) 🙃🙂9:39 ".... that blows when overvoltage takes place", you mean current. Such fuse blows only when there is excessive CURRENT. 9:50 Aha! I spotted one (GPT) in an inverter, didn't know what it was until now. It was tied between the 230v output AC lines (N & L).
Hi Great Scott! Very interesting program. Great that you opened the two overprotection circuits. Interesting to see what they are doing to protect from the surge. HOWEVER, the really interesting teardown would be the OBO Bettermann Surge Protector. I have been wondering for many years what is the magic in there and why they are pretty much the only source for this class of surge protector and what they use in there to justify such a steep price. Last time I looked at them some 20 years ago they were over 300 € each. So, if your budget for a program stretches tgatbfar could you examine the OBO device. BTW the OBO device is the one that is use by pretty much all the railways around the world for serious surge protection. So, what is their magic sauce?
.... Another repeatable and controllable source of high voltage spikes would be the inductive spike from opening current path to an inductor (normally suppressed by flywheel diode)...
Reminds me of “load dump” in car electronics. Never really figured out what to do about it and figured if it happened I’d have bigger problems than my relay drivers failing. Fused MOS might have been a good solution.
Load dump has a finite amount of energy. And the voltage is also not so dramatic. Around 60V is the accepted load dump voltage. A semiconductor transient voltage suppressor can help. Do you want the circuit to reset automatically? Or are you OK with a fuse blown? The load dump in a motor vehicle is absorbed by many devices that are still connected when the battery is suddenly disconnected with the alternator at high current output... The engine is running and the over-voltage can be absorbed by lots of things. Not just your gadget. It's instructive to look at the circuits in well designed automotive devices to see how they handle the eventuality. A TVS will begin to conduct at a certain voltage. You can use that current to switch a hefty transistor to shunt a large current at over-voltage to ground, through a low impedance, protecting the circuit behind it from very high voltage. The transistor input should be separately protected unless it can tolerate the dump voltage. (This will cost more than a few cents!) Note that the load dump may not provide enough energy to blow a fuse if the protected device usually needs several amps. Other options include selecting load dump tolerant devices such as automotive voltage grade voltage regulators to protect the power in your circuit... But not the inputs. Inputs are generally high impedance (within their clamping voltage) and a TVS of appropriate voltage rating should do the rest by preventing excess load dump energy getting to the input pins of the chip. ESD protection is a bonus. Always keep the whole thing in perspective; not how you protect your microcontroller chip from the whole monster load dump! And don't spend $100 protecting a $10 circuit unless it's literally *vital* for the circuit to keep going.
I have had overvoltage(home) in sockets but the ptotecton has not blown due to low power demand aka no load but low power devices and those devices have got lots of damage, some insta out of commission others over time.
I use large GDTs on the antenna lines to protect my amateur radios. I live in the lightning capitol of the world, so that sort of thing is a requirement here unless you just like buying new radios all the time.
We once had a lightning strike cause some weird problems: other than tripping every signle breaker, the ethernet port stopped working on a computer, a signle line got burmed out on an led display of the elevator, and it partially broke a dvd player. Thankfully no serious damage occured
I have had a MOV absolutely blow it's guts up in a shower of sparks in a power supply for my PC, it was green in colour and ceramic and was within the power factor correction side of the circuit. I was having constant over voltage events of around 255V for a year and I think it wore out or something. All I did was flip the switch at the wall
When you mentioned the disadvantages of the TVS diode, you said leakage current and parasitic capacitance. But don't these also exist in normal wire? Maybe to a smaller degree?
I've been struck by lightning, I think. I was in the garage, but touching a metal security door frame and next thing I knew I was lying on the floor and my fingertips were blackened. This was at my grandma's home, but att my second childhood home the house kept getting struck by lightning and a lightning conductor did nothing. Computers and dialup modems had to be replaced often. Lightning behaves weird when there are ore deposits involved.
The leads should be kept short between the TVS and the component being protected, otherwise, the voltage spike across the leads because of the inductance might cause the TVS to clamp too late to save the component being tested.
Idea for next videos: Recently I found out that I can power my Black pill from my smartphone with simple TypeC-TypeC cable (yes, it's all about 5.1K resistors). Could you please research on it and investigate max power capabilities smartphone can have, which other embedded devices it can power, can it power small motor, can it be burned down, does smarphone has protection agains overcurrent and so on? For me it appeared like an interesting topic as I am currently developing 2 small projects where my devices are not required to run 24/7 but only when only man uses it. So I could get rid of wall power adapter or Li-ion battery
T7 component tester has a tiny triple TVS that blows if you forget to discharge the cap before testing. After replacing it twice, I gave up and use my other cheap tester with no problems. If you have one, you know what I mean.
This is probably the best video I've seen about a concept I knew about but hadn't delved into. You mentioned that a MOV is often alongside a fuse in a power supply that blows when such an event occurs. I'm wondering what the fuse adds in such a circuit beyond fragility. Is it for non-transient high voltage events?
Them huge capacitors on the spinning magnet wheel.... Memories of the simple schoolgirl engine & beyond. Bedini would be proud. Now just catch the collapsing spikes in a lead acid battery, to feed back to the capacitors. Not perpetual motion, but tuned right, a heat pump that can run for a surprisingly long time.
Haha, I just got a bunch of 6 leg micro packages in and spent hours manually soldering those legs without a breakout board! Broke 3 legs off before I could get one working and said f this and ordered break out boards.
You shouldn't directly solder the wires lol, add thin copper wires as extension and accommodation for twisting and turning. It's easy once you know the trick😉
Danke vielmals für das Video. Genau das war gerade Thema in unserem Repair-Café (der Kollege konnte die TVS nicht identifizieren) und ich frage mich selber schon länger, welche Optionen ich zur Absicherung einer Schaltung noch hätte. I will stay tuned :-) - Liebe Grüsse allen Klein und Gross
That was a really interesting video. What I'm wondering now is how I can determine where I would add a TVS in my circuit. How can I determine where it makes sense to protect from these voltage spikes? Do I just add a diode to every IC input?
The general rule I've seen is to add ESD protection anywhere a signal leaves the board in a manner which may be connected or disconnected while energized (based on the theory that the inductance of cables, etc. is likely to cause ESD beyond what ordinary decoupling capacitors will suppress). My rule is to add ESD protection to all signals which leave the board, at every connector of said signal if board space allows and at one if board space is tight, and to massively oversize ESD protection on power circuits or others in the same connector (this can save your board from being fried if you, say, incorrectly assembly a wiring harness providing power to your board and flip, say, the 5V and 12V rails, assuming you inspect your board with a thermal camera shortly after powering it up the first time). TVS diodes are cheap and very easy to integrate so I choose the safest option where possible. Then again, for cheaper boards this may not be worth it, and as others have mentioned here, good design practices surrounding decoupling capacitors will stop most ESD from being a problem. Different ICs also very significantly with respect to their internal ESD protection, some (such as low-cost microcontrollers) may be easily damaged while others (such as industrial-grade differential line drivers) may be extremely robust. Checking the datasheet is always a good idea if you're producing a quantity of boards where the cost of ESD protection would make a difference.
"One point twenty one jigga-watts! Great Scott!" Definitely interesting; I'll use TVS diodes in my projects more. MOVs are pretty common, I replaced a good bunch of these in various devices, which means they did a good job of protecting the circuit.
Won't the shunt capacitnce from the varistor across the supply rail reduce the peak surge from the spark? On a supply rail you can add capacitnce which in conjunction with the short time of the fake static distarge would provide enougbh protection.
I've had many components saved by TVS diodes. I have noticed an uptick in their usage even on bargain Chinese fake stuff and they seem to work most of the time except the cheap ones tend to blow up a lot easier then you have to replace it.
I had lighting strike the radio antenna of a cellular remote oil well monitor that I had built a long time ago. It skipped right over the lightning arrestor on the antenna cable and blew the ground track off the entire circuit, through the cellular modem, through the serial cable connecting the modem to the PLC, through the PLC circuit board, through the ground wire and into the electrical circuit on the engine skid that was supplying power to the automation panel. Was a hell of a thing. The ENTIRE circuit from top to bottom vaporized. There was nothing left of the tracks at all.
That sounds scary. Thanks for sharing your story :-)
"Well shit"
@@greatscottlab Well, it didn't happen while I was standing next to it. I got a call saying the install wasn't powering on. Remote trouble shooting didn't turn up anything so I had them pull the unit and send it back to me. I took it apart and found dust and vaporized copper.
What brand of PLC? That's not something that you see happen to a PLC very often.
The problem with voltage protection is that it's like walking under the hoover dam with your umbrella open. You should be good, right?
Gosh
I repair control circuits for some industrial equipment, and they regularly use both a TVS diode and GDT in parallel on each line. The TVS diode handles the lower voltage transients, while the GDT protects against higher incidents like lightning surges. It works very well, and it's easy to test and cheap to replace if it does blow.
Thanks for the feedback. Good to know.
Hearing the way each device worked that was my immediate thought "can you use them in parallel to protect against a wider range of voltages?". Good to hear I was on the right track.
If they're in parallel, couldn't a high voltage transient just take the TVS line?
I'm still learning this stuff
@@lostsauce0 So, electricity always takes the path of least resistance (more or less). Think of each of these devices as a conditional short, it's normally not shorted to ground, but under specific sets of voltages it triggers a short, at which point IT becomes the path of least resistance, and so the voltage spike will take that path rather than flowing through the higher resistance (sensitive) circuit.
Each device has a limited range of voltages and response times it's good for, past a certain point it will fry (or under a certain point it won't trigger), stop conducting, and then you're back to the circuit you're trying to protect being the path of least resistance which you don't want. By putting a bunch of them in parallel you're covering a wider range of potential voltages, and even if multiple of them get fried, it's cheaper and easier to replace these components than the rest of the circuit.
a TVS failing would mean that it does not conduct to ground, so a parallell device that also conducts to ground during a high voltage transient, would take over. Or what did you mean by "take the TVS line"?
Varistors are way slower than TVS diodes and aren't fit for fast transients such as ESD & EFT. Another thing to remember is Kirchoff's law, any current pulse will return to its source. Therefore instead of trying to suppress a pulse it is often more practical to divert it from the sensitive components using spark gaps etc
Thanks for the feedback :-) Maybe I should have included spark gaps as well. Well, I will find space for it in another video.
Seen a park gap do it's job in practice once (that is, after the fact not during it), on my TV's power supply. The comment I talk about it in case you wanna read: ruclips.net/video/MzxBBXpgwrE/видео.html&lc=UgydNWuE3TvVqRYRkm94AaABAg
The board does have an MOV there too I assume to protect against lower voltage surges instead of just lighting strikes or such high voltage stuff. Those two in combination sure did their job.
Spark gaps still leave residual overvoltage which needs to be handled. They are also not suitable for a low transient range. Neither are repeatable or reliable. Mov and tvs coupled with esd suppressor is the way to go.
Is a tvs diode not an esd suppressor? @@ecstazyrm
@demetrisswest What is the path of current in TVS in case of ESD. when circuit is not connected to anything. can u explain a little please.
these should be compiled to a playlist "things i should have known before starting electronics". this is a huge lifesaver. good work as always.
Here's a video idea: Signal integrity
When and how to use shielded wires, coaxial wires, differential twisted pairs, ect and avoiding issues like ground loops (like connecting the shield only on one end) ect
Yess. And things like 50 ohm terminations and signal reflections
@@ameliabuns4058 Oh yeah, definitely!
@@iamdarkyoshi although I think Scott's area of expertise is high voltages and mains stuff? so he might not be the best person for this as much as I love the way he explains things.
Based on his stair light PCB layout, running small-signal traces parallel with the PWM-regulated unfiltered LED power traces with the two LED power traces routed about as far apart from each other as possible for maximum EMI, I doubt Scott knows anything about signal integrity. He'd certainly gain a lot from at least looking into the basics.
@@teardowndan5364 well that could be phrased less harshly but yeah.
When designing automotive type electronics that had to meet EMC standards, the electronics had to withstand ESD discharges up to 15kV. Something as simple as a 1000pF cap at the input was often sufficient protection. Signals that couldn't tolerate that much capacitance were good candidates for a TVS diode. For power line transients, which are much longer in duration and contain much more energy, a larger TVS was usually the solution (perhaps in a SMB package). The datasheet will tell you how much energy a TVS package can handle. MOV's weren't used, and I get the feeling that they are more oriented towards mains power where lightning strikes are a risk.
"Doesn't happen often that lightning strikes near you." I live in Singapore and I've had a thunderstorm every afternoon for the last three days. Yesterday the lightning strike was so close you could hear the crackle before the bang. But at least the humidity stops you getting static shocks from the carpets in the office.
Something you might be able to advise on is the problem that I have, which is when there is a lightning strike nearby my RCD trips, none of my neighbours seem to be affected. I've always thought it was because I have an induction hob and the coil is picking up the EMP from the lightning and creating a current flow in the neutral line, as normally the hob is switched off at a wall switch, (which I assume only isolates the live), and hence causing a differential L to N current flow.
You could probably have used an electric fencing unit to create a known and constant voltage, or a car ignition circuit.
Thanks for the feedback. That sounds scary. And there I think overvoltage protection is a bit more important :-)
@@greatscottlab Yeah when I lived in SE Asia my UPS would last a few months before bulging from an overworked battery due to a lightning strike They never lasted longer than a few months. Plus, aside from Singapore, a lot of countries in SE Asia don't use ground. I added a ground for a house we built, not sure how reliable it is for a big short or overvoltage, but at least it seems to work according to the multimeter.
@@greatscottlab Nearly all power lines in the country (not that it's very big) are buried so I don't think that over voltage events occur very often. I've never had of over voltage protectors on domestic distribution boards.
One time lightning struck pretty close to my Mom's house. My brother got a small shock from the wired headphones he was wearing (plugged into a PC which was grounded). It also tripped a GFCI. Not sure why it tripped, I just chocked it up to high voltage doing funky things.
I lived in Florida, USA for a while. Lightning happened every other day. I lost a computer motherboard to a cable modem conducting a lighting strike from the cable system into it and thine through it into the computer because the lightning arrest system on the building failed. My next computer had optically isolated I/O ports and my replacement cable modem's coax went through a surge suppressor, as did the LAN cable that came off it, isolating any future lighting strikes from the network and my new computer with three layers of protection.
As a notebook repairman, I am very familiar with diode arrays, it is very common to find them protecting the USB, HDMI and other connectors on the boards, many versions of notebooks for home use differ from the corporate versions in that although they have the tracks, They do not have the chip installed, this small saving makes the notebook much more delicate, plus now all those connections go directly to the SOC, system on chip, practically killing the processor in the event of any incident.
Man, I've been a fan of yours long enough to know your real name. Haha, I think it's about 10 years that I have been a subscriber (when you still responded to messages on facebook within 24hrs, haha). I started watching you when I was in high school at 14 years old. Actually, you played a very big role in me studying BEng Electronic Engineering. I am now 24 working as an electronic engineer. I recently did an ADC protection design for one of our products where I used TVS diodes. I actually learnt what those are from you years ago. If I ever find myself in Germany, I'd really like to have lunch with you. I don't think you realise how much of an effect your videos have on some individuals. Much love from South Africa.
I'm familiar with GDT as a "Comgap". As in communications gap. Back in the landline days, they were a required component (in Canada at least) across tip and ring on a phone line. Every modem had one for example. It was to supress lightning strike level voltages from reaching the consumer. If it happened to protect a circuit, that was just a side effect. They were often used in combination with a MOV. Old-school rotary phones generally didn't need them as the ringer coil would take the punishment instead.
A thing I like to do to protect low current inputs is a high value series resistor and a capacitor to gnd behind it. The resistor limits the current and the capacitor just eats the charge, so there is nothing left that the internal protection can't handle. Great for button inputs, as this also covers debouncing or acts as a low pass for "slow" analog inputs
Thank you for the video! The MOV surprised me in this experiement, I didn't expect it will protect against sparks due to its slower reaction. I designed an off-line converter with an LNK-304 and lacked the mov. When my father plugged in a washing machine to the nearest socket the IC blown away. Never lack the MOV!
Thanks for the feedback :-)
9:37
Quick correction, that isn't a thermal fuse, it's just a standard electrical fuse.
A thermal fuse is tripped when the temperature of a component gets too high.
For example on motors to stop the windings from burning, electric heaters to stop your house from burning, hotplates to stop the hotplate from burning and so on.
The resettable thermal fuses are usually found on components like electric heaters, hotplates, dehumidifiers and other devices, though are a first line of defense, they reset after the temperature drops (though microcontrollers have taken their place in some applications) but they'll also be found as second line of defense devices on industrial electronics (or higher end consumer stuff) where you have a uC and temperature probe (thermistor or thermocouple) monitoring the temp, but if the uC crashes, the resettable thermal fuse cuts the circuit, then if that fails a thermal fuse (attached using thermal paste) blows, giving the device 2 chances to recover and one chance to need fairly simple repairs.
Hi GreatScott. Good video! I would like to add a few things:
9.44 the fuse in series with a varistor that you have drawn is not a thermal fuse, but an ordinary fuse, which would open when a varistor is damaged due to a high surge current. A thermal fuse is usually added in series to the varistor, so that if a varistor gets damaged and it conducts some small current it gets slowly hot - in that case a thermal fuse would open. some manufacturers even include a themal fuse in a varisto case, so the varistor has 3 pins. But you mentioned that the fuse blows when¸ an overvoltage event takes place, which is wrong. this would be a terrible design. actually we have to choose fuses, which have a high I^2t rating so that they DON'T blow during expected surge events.
Your testing is only with ESD pulses, for which TVS diodes are best. Varistors and GDTs are used for surge pulses (indirect lightning strikes and inductive overvoltage events) which contain much more power. the 8/20us pulse is a typical indirect surge testing shape, which has 8 us of a rise time and 20 us of duration. there is also 10/1000us which has much more energy.
ESD protection and surge protection are very different things.
Usually for surge protection we use several of these protection components. for example we start with GDT, then varistor and then with the TVS diode. All three seperated with an inductor. Current and voltage on the device gets smaller with every component.
11.46 this is a thermal fuse and it opens when the varistor gets hot. the failure mode of a varistor is that with every surge pulse it conducts more leakage current. and when this leakage current is smaller than the current fuse in series, the varistor gets very hot and it ca cause a fire. that is why standards determine there has to be a thermal fuse there, which opens when varistor gets too hot. then it is time to change the protection module.
Note that an ESD pulse has a rise time of 1 ns and duration of approx 50 ns. But a surge pulse has a rise time of 8 us and duration of 20 us or more.
The problem with varistors is that the current through the device is increasing quite slowly with voltage rise. But a TVS diode's current increases much quicker with increasing voltage. at 7.58 you drew the correct V/I characteristic. but when you dig in a datasheet you see that the current rise in varistors is very slow. For example the varistor you showed MOV-20D681K conducts
1000A at 1600V and 1A at 800V. so it need additional 800V to get from 1A to 1000A.
Usually a varistor is not used as a ESD protection device. I think it protected your circuit due to varistors parasitic capacitance, which conducts ESD current and not really due to the varistor voltage limiting.
You used the term “power”, but “energy” is a better term for these discharge events.
@@stevebabiak6997 true. Usually I’m checking max peak power capability in datasheets at a specific pulse time.
This might be one of the most practical ESD - TVS based circuit protection video!
I mean there are so many nuances of using TVS diode, routing them to the ground plane, ensuring low inductance to ground etc. This video gave me practical insight into the core fact that TVS diodes convert the ESD events into heat - boom seeing it practically save the circuit - made my day.
I wonder how effective spark gaps are, given they're pretty much free to add to a PCB. Our TV's power supply got quite a few of those, and after a lighting strike nearby, the cable TV box burned and it made the favor of killing the TV's motherboard through HDMI, along with the Switch dock on that same TV. Replacing the motherboard made it work again, but after a while the power supply also went bad, and I could see on the mains pins a tiny welding like dot on the metal casing (clearly an arc mark), along with a missing spike on the spark gap PCB track.
It was already more than 10 years old at this point and it still survived that lighting strike.
Thanks for the feedback. Interesting story. I should put spark gaps on my to do list for videos.
Spark gaps definitely have a role in protecting electronic circuits from discharges of energy. In the world of color CRT displays, it was very common to have spark gaps integrated into the CRT connector socket as part of the protection for the video output circuits. The socket would also use a thin but wide grounding “wire” (strap) that would typically get connected to the outside coating on the CRT funnel, to have a low impedance ground return for any arcing that bridged the spark gaps.
2:47 I always thought "O.L." on a multimeter in resistance measurement mode stands for "Open Line/Loop".
When using a MOV on high side of power supply, you need to take account their failure mode: slowly conductive.
So either couple them with thermal fuse close to them, or a combo MOV+ThFuse, or you can end up with a fire event in case of repetitive surges.
I would like to add some observations and know more. I have seen many SMPS with MOV for protection. First failed component I always saw is MOV with dark discolored and often cracked like 2 separate disks for each terminal. It happened over a few years in India and often with power surges. Specifically they open circuit after 3-5 power surges. Which is around 2 months of time. I tried observing this because another SMPS had a fuse installed instead of MOV with almost same rating. These SMPS were used to power small LED panels for a business display sign with a LED driver always.
I built a home energy monitor using CT clamps, and I learned they incorporate a TVS diode inside the clamp. Because without any load on the other end of the clamp, the 100A of current on my main line running through the core can generate theoretically millions of volts, and the whole thing just explodes in a puff of smoke without even touching any metal wires.
I learned about their existence because I didn't understand any of this and unplugged my CT clamps from their load while current was running through them. The TVS diodes did their job, shunting the voltage to I think 22 volts, and making a hell of a racket, both noise and vibration, while they did so. Both clamps were unharmed and continue working accurately to this day.
As far as your microcontroller supply line goes, simply adding the 100n decoupling cap that should be there anyway will pretty much protect it against ESD. The pulse is high voltage but very low energy (standard ESD model often uses 150pF) and thus the decoupling cap will act as a capacitive divider and "eats" the ESD pulse.
Nice video, very informative. There is one thing I was missing. When adding such a protection to your pcb, you should pay attention to your layout and routing. You must route your wire through the pads of the tvs/mov to get the best results. And ideally they should be close to the connector.
Small Neon Bulb can also serves as surge & lightning protector and suitable for small outdoor antennas. Its break over voltage is about 90v
We usually work in solar systems and we use surges all the time in both AC (grid input) and DC (pv input) but not these types it is similar to the red one but a glass insulator breaks open and allows everything to flow to ground instead of dissipating as heat and it is connected in parallel with the circuit. AGAIN IT IS USED FOR DIRECT LIGHTNING STRIKES AND CAN ONLY POP ONCE BEFORE DISPOSAL.
I have experienced two direct strikes on antennas fitted with lightning arresters and direct grounding wiring. The equipment survived without damage but proper protection is a bulky affair with thick cabling required to provide a low impedance path to ground. All the power circuit protection devices tripped, but no damage was done and everything worked after the tripped breakers were reset.
The GDT can create a lot of interference and scramble sensitive logic circuits digital logic since when the tube fires, the voltage across it collapses very quickly to a lower voltage, creating EMC surges. TVS diodes and MOVs tend to create much less EMC interference because they clamp the voltage and don't exhibit the voltage collapsing effect. So whenever we have to use a GDT (for example, when extremely low leakage current and stray capacitance is critical), we have to incorporate other protective devices in such a way to minimize the EMC surges.
The main reason the MOV is protected by a fuse in series is not because the fuse will blow in the case of an overvoltage event. As the MOV starts to age due to surges, its resistance decreases, which in turn increases the current flow through the MOV, leading it to heat up and you end up in a vicious circle that will end in the MOV bursting or cathing fire and the circuit, without the fuse, drawing way more amps than anticipated. In order to protect that from happening, the MOVs are now protected by a fuse.
Increasing the gap voltage doesn't change the voltage applied to the board by much since most of the extra voltage simply gets lost in the arc. It does increase the peak discharge current though. As for why the MOV manages to save the circuit despite having much higher voltage, that is down to your generator having pFs of output capacitance discharging into the MOV's nF of parasitic capacitance, forming a capacitive voltage divider that drags the discharge voltage low enough to save the input pin from excess ESD current. Also, IO pins have their own internal protection from the output FETs' body diodes that can usually handle human-body-model ESD as-is as long as the chip has a path from power+ground pins (depending on ESD polarity) to actual ground. Chips on an assembled grounded PCB can be surprisingly difficult to kill as Linus demonstrated in his Electroboom special.
I think you are the only one in this comment section who actually knows stuff.
Good video. I want also mention, for low speed input lines often simple ceramic capacitors are used to protect the input from ESD loads. Usually 10 to 100nF are used.
This video is really helpful. I suffered a power spike a couple years ago when the transformer outside my house blew up. lost a few appliances, namely my robot vacuum, looking at the board there were large scorch marks and some exploded components, after some research I discovered it was a varistor. easy enough to replace and still working to this day
At 6:12 there is a mistake: increasing the air gap doesn't increase the voltage across the board but only the voltage across the air gap, doing this just increases resistance and reduces the current.
TVS packs are the electronics tinkers best friend!! In many years since I first used one I do not think any of my projects have ever been without one. Though I have moved in the last few months to a new location there are so many things they want me to do here. First was the safety light in the stairs and then the active emergency LED lights in various rooms that are going through a minor upgrade. But the use of the TVS ane other types of protection devices is common in this lab.
Peace
Love this. Been researching ESD protection options and came across TVS diodes recently. This video really sold me
I used some components like the GDT to protect a lora antenna coax from high transient voltages, probably will not know if it really works until it doesn't. I think it was suggested to replace them occasionally as repeated spikes can make them fail. I imagine that a passing thunderstorm could induce significant spikes even without a lightning strike. I really appreciate the information you provide, Thank you.
We used to add MOVs to anything with an inductive load (PLC control systems working at 24VDC). Even contactors or solenoid valves can be a problem. Inductive spikes were a particular problem with high speed counters or servo-motors which use incremental encoders. An extra count here and there can really foul things up for the position of a pick and place robot.
why MOV and not reverse diode?
4:58 I disagree, you can find mosfets, comparators, regulators, ... In SC70 packaging
It's generally a bad idea to guess components by their footprint
Fun fact: some CRT display video output amplifiers used neon bulbs to protect against the CRT arcing through the video output electronics. Quite like a GDT as shown in this video, but you could see them flash occasionally while in action.
9:14 Why was the LED still blinking after you removed the protection and shocked the circuit again?
Thank you for the deep-dive and clarifying these parts. What is the best practice for TVS application in a digital circuit? Just on the power rails? Power rails and any connections that touch the outside of the product housing? What about cap touch LCDs that get touched (obviously) - their power rail too?
One of the most useful channels on YT thanks GreatScot this is really a great video.
You can make a zenner diode go there, you can even create a zener diode serial array. This way you have a full wave antenna that bounces until it adds up to the next zoltage. So you wrap up full wave zeener diodes inside the blade antennta, at each stage, well it is mhz, so the audio is much louder in real time.
If you ever saw
or will see
such an IC
on a PCB
then there is a big possibility
that it is a transient voltage suppressor IC
I've seen real weird stuff happen with lightning strikes, I'm sure it could be explained but oddness like a garden patio covered in tiny copper balls which were the leftovers of the coax from a direct strike on a TV antenna but the TV itself worked just fine, pretty much every piece of electronics in the house next door was dead, there's just so much power that it's incredibly difficult to defend against.
Static is easier but it's always good to know that MOVs, VDRs and transient suppression devices like those tiny diodes degrade every time they're 'triggered', it's a real common fault for them to fail on USB ports in handheld devices for instance
Thank you for this video, i love almost every video Great Scott produces.
MOVs are commonly used with a few other components in AC power supplies such as class Y capacitors. I would love to see these covered in similar videos. Please and thank you.
Are you sure about the annotation at 2:49? I've always heard OL as "Open Line", which also makes more sense than "overload", since nothing is overloaded, and two probes with no conductivity between them definitely constitutes an open line.
I've used MOVs before, but didn't know about TVSs. Thanks.
I was wondering about your estimation of generator voltages. When I did experiments with other static generators, we estimated our gap voltages at 10kV/cm. 4 inches was 100kV. I'm not sure which is more accurate, yours or ours. Of course we were in a fairly dry climate.
Also, did you notice a change in spark frequency when you increased the distance (voltage)? It seems to me that it should take longer to to build enough charge to jump a larger gap, implying that the frequency should go down. Did you notice that phenomenon?
I got the kV/cm rating back in university. It is also just guessing because it depends on lots of factors. And yes; the bigger the distance, the longer it took to charge up.
In the UK it's about 10Kv per cm as you say
Thanks, sometimes I think you are looking over my shoulder with the timeliness of your videos. Right now, I was designing a circuit that needed this type of protection, and here you are with helpful advice.😀
@GreatScott -- I don't know how deep on this topic you're interested in going, but -- in light of your summary at 10:31, it would be interesting to explore how to use these components in combinations. (Along with "spark gap" which is essentially DIY GDT.) It might seem using these in some combination might handle multiple over-voltage/arc scenarios, but does the right protector step in at the right time or voltage to avoid the other protector getting overstressed?
Varistor are common in circuit specially in smps but there is problem. If there High enough voltage it will arc between the pads of circuit. I have literally fried 3 smps in similar way. When I opened it up the copper pad were the witness. So tvs diodes are perfect for such applications.
It probably should be mentioned that MOVs have a limited number of "hits" they can take. Or to put it another way, they get less effective each time they are activated. It would be very informative and intuition-building to see you demonstrate that effect!
You could say your catchphrases in German. This would be your really special feature, and truly authentic too!
Thank you for what you are doing! Keep doing it, man
The cheaper one for solar should actually handle more current. It uses a multi layer carbon spark gap with movs to equalize the voltage across the disks via their capacitance. Mov protection is good for lower let through voltage but it is not going to handle current as well as a spark gap. Only issue with spark gaps / GDT is their follow on current, but you need a high current to maintain the arc, your solar wont produce that much current.
Also TVS are generally used for data lines because MOV capacitance is too high for higher frequency signals.
I dicovered TVS and varistors while reasearching for a DIY TIG welder, they seem to be used at the output of the Diode rectifier to avoid the HF/HV starter from blowing the diodes
Everyone should take a look at replacing all of these with the new devices TI has come up with. They are meant as TVS replacements so are named like TVS1400, TVS1401 etc. Where most protection devices are good for at best 100 hits these are tested at 5000 and still work🤯 They are especially good for automotive devices where repeated surges are normally experienced on a nearly daily basis.
Also do a web search for "combined SMA and TI TVS footprint" for an idea how they can be added into an existing design to allow easy transition without a complete redesign being needed.
It's very educational to have all of these in one video. Well done.
Ngl, this is the first time I've heard of a MOV, and the timing couldn't be more perfect!
Glad to hear it👍
FWIW, there are a lot of tutorials on how to add these to power strips, like one that your computer system is connected to.
The better 'strips' have these, however seems to be pretty easy to add higher spec'd ones to your existing ones pretty cheaply.
Wonder if there is a 'standard' which incorporates both the MOV and TSV in series to eke out most protection...
I installed on every circuit with expensive Elektronik, a surge protector to my outlets. Inside they have a couple MOVs and one GDT.
About 10 years ago I accidentally connected a 12v dc circuit to 240v ac, when it didn't work I soon realised my mistake and connected it back to the proper 12v dc again and it worked. I've never quite understood how the 240v didn't fry it. that same board is still working to this day.
...Re insulation tester being too safe (3'12")
... How about using insulation tester to charge a high voltage capacitor (maybe even your Leydon jar???), then use a switch (or flying lead) to connect your test circuit board? ( i.e. only connect test circuit once the capacitor voltage is at the desired destructive level)
...A neon in series might also add some visual effect?
When a MOV has been severely or repetitively compromised it deadrigates to a shorted component. This depending on its design location can cascade circuit damage.
MOV or TVS diode are not that different. MOV is a sintered material and has the same diodes but just an enormous amount of them in parallel and in series. If these are protecting they can degrade bit by bit until they short or break. The TVS diode will be gone in one single overload. The GDT's are typically used to protect against lightning and always combined with THS or MOV. ( Which is the reason it failed to protect your controller )
In old TV's there was a component which just provided a spark gap in air. just 2 wires 0,3 mm apart, serving as a GDT with the gas being air.
Worth being mentioned is clipping diodes in a digital input circuit. Often integrated in the IC but if not, it can be placed externally.
If lightning strikes directly, you can have all the protection possible : Nothing will save your circuit.
9:25 ".... dies immediately...." (shows a working circuit) 🙃🙂9:39 ".... that blows when overvoltage takes place", you mean current. Such fuse blows only when there is excessive CURRENT. 9:50 Aha! I spotted one (GPT) in an inverter, didn't know what it was until now. It was tied between the 230v output AC lines (N & L).
Wait a moment... That isn't ElectroBoom job?
He didn't do this video topic yet. So I had to do it ;-)
@@greatscottlab when a collab? Or is too much a trip on German hotels?
@@s0ygecko Forbidden on grounds of safety.
@@s0ygeckothe only things that will trip are the RCDs...
@ElectrBOOM can reclaim his job? 🤣
Hi Great Scott!
Very interesting program.
Great that you opened the two overprotection circuits. Interesting to see what they are doing to protect from the surge.
HOWEVER, the really interesting teardown would be the OBO Bettermann Surge Protector. I have been wondering for many years what is the magic in there and why they are pretty much the only source for this class of surge protector and what they use in there to justify such a steep price. Last time I looked at them some 20 years ago they were over 300 € each. So, if your budget for a program stretches tgatbfar could you examine the OBO device.
BTW the OBO device is the one that is use by pretty much all the railways around the world for serious surge protection.
So, what is their magic sauce?
.... Another repeatable and controllable source of high voltage spikes would be the inductive spike from opening current path to an inductor (normally suppressed by flywheel diode)...
I never knew about transient voltage suppression components. Thanks for walking us through-it was very interesting!
Spark gaps on PCBs are also commonly found in the primaries of power supplies for this purpose. Thought they deserve an honorable mention :))
9:52 i thought - Gate Drive Transformer
Best DIY electronics channel on RUclips.
Reminds me of “load dump” in car electronics. Never really figured out what to do about it and figured if it happened I’d have bigger problems than my relay drivers failing.
Fused MOS might have been a good solution.
Load dump has a finite amount of energy. And the voltage is also not so dramatic. Around 60V is the accepted load dump voltage. A semiconductor transient voltage suppressor can help.
Do you want the circuit to reset automatically? Or are you OK with a fuse blown?
The load dump in a motor vehicle is absorbed by many devices that are still connected when the battery is suddenly disconnected with the alternator at high current output... The engine is running and the over-voltage can be absorbed by lots of things. Not just your gadget.
It's instructive to look at the circuits in well designed automotive devices to see how they handle the eventuality.
A TVS will begin to conduct at a certain voltage. You can use that current to switch a hefty transistor to shunt a large current at over-voltage to ground, through a low impedance, protecting the circuit behind it from very high voltage. The transistor input should be separately protected unless it can tolerate the dump voltage. (This will cost more than a few cents!) Note that the load dump may not provide enough energy to blow a fuse if the protected device usually needs several amps.
Other options include selecting load dump tolerant devices such as automotive voltage grade voltage regulators to protect the power in your circuit... But not the inputs. Inputs are generally high impedance (within their clamping voltage) and a TVS of appropriate voltage rating should do the rest by preventing excess load dump energy getting to the input pins of the chip. ESD protection is a bonus.
Always keep the whole thing in perspective; not how you protect your microcontroller chip from the whole monster load dump! And don't spend $100 protecting a $10 circuit unless it's literally *vital* for the circuit to keep going.
I have had overvoltage(home) in sockets but the ptotecton has not blown due to low power demand aka no load but low power devices and those devices have got lots of damage, some insta out of commission others over time.
I use large GDTs on the antenna lines to protect my amateur radios. I live in the lightning capitol of the world, so that sort of thing is a requirement here unless you just like buying new radios all the time.
9:52: Global Descriptor Table refernce???
We once had a lightning strike cause some weird problems: other than tripping every signle breaker, the ethernet port stopped working on a computer, a signle line got burmed out on an led display of the elevator, and it partially broke a dvd player. Thankfully no serious damage occured
Thanks for the feedback.
About these TVS, if I remember correctly bigger surges may dge the diode but still protect the circuit.
I have had a MOV absolutely blow it's guts up in a shower of sparks in a power supply for my PC, it was green in colour and ceramic and was within the power factor correction side of the circuit.
I was having constant over voltage events of around 255V for a year and I think it wore out or something. All I did was flip the switch at the wall
Thanks for another clear and informative video Scott
When you mentioned the disadvantages of the TVS diode, you said leakage current and parasitic capacitance. But don't these also exist in normal wire? Maybe to a smaller degree?
I have learned something new, never heard of HV protection and I am into electronics for more than 20 years.
I've been struck by lightning, I think. I was in the garage, but touching a metal security door frame and next thing I knew I was lying on the floor and my fingertips were blackened.
This was at my grandma's home, but att my second childhood home the house kept getting struck by lightning and a lightning conductor did nothing. Computers and dialup modems had to be replaced often. Lightning behaves weird when there are ore deposits involved.
The leads should be kept short between the TVS and the component being protected, otherwise, the voltage spike across the leads because of the inductance might cause the TVS to clamp too late to save the component being tested.
Could you use the insulation tester to charge a high voltage capacitor, and then output through a known resistance to limit current?
Agree
Idea for next videos:
Recently I found out that I can power my Black pill from my smartphone with simple TypeC-TypeC cable (yes, it's all about 5.1K resistors). Could you please research on it and investigate max power capabilities smartphone can have, which other embedded devices it can power, can it power small motor, can it be burned down, does smarphone has protection agains overcurrent and so on?
For me it appeared like an interesting topic as I am currently developing 2 small projects where my devices are not required to run 24/7 but only when only man uses it. So I could get rid of wall power adapter or Li-ion battery
T7 component tester has a tiny triple TVS that blows if you forget to discharge the cap before testing.
After replacing it twice, I gave up and use my other cheap tester with no problems.
If you have one, you know what I mean.
Thanks for the feedback.
This is probably the best video I've seen about a concept I knew about but hadn't delved into. You mentioned that a MOV is often alongside a fuse in a power supply that blows when such an event occurs. I'm wondering what the fuse adds in such a circuit beyond fragility. Is it for non-transient high voltage events?
Hey GreatScott!! I would love to see a video about ferraite cores.. how do they work and when/how to use them 😊 i think its a quite intersting topic
Them huge capacitors on the spinning magnet wheel.... Memories of the simple schoolgirl engine & beyond.
Bedini would be proud. Now just catch the collapsing spikes in a lead acid battery, to feed back to the capacitors.
Not perpetual motion, but tuned right, a heat pump that can run for a surprisingly long time.
Haha, I just got a bunch of 6 leg micro packages in and spent hours manually soldering those legs without a breakout board!
Broke 3 legs off before I could get one working and said f this and ordered break out boards.
Oh boy.....
You shouldn't directly solder the wires lol, add thin copper wires as extension and accommodation for twisting and turning.
It's easy once you know the trick😉
@@gentleclash exactly what I did. :)
Danke vielmals für das Video. Genau das war gerade Thema in unserem Repair-Café (der Kollege konnte die TVS nicht identifizieren) und ich frage mich selber schon länger, welche Optionen ich zur Absicherung einer Schaltung noch hätte.
I will stay tuned :-) - Liebe Grüsse allen Klein und Gross
I'm a fan of Sidactors, which behave like SCRs/Triacs/crowbars. They reduce the time at excessive voltage, and minimize their own power diss.
Mouser has been the best supplier I've dealt with. Customer service set the bar unbelievably high. I've dealt with them and allied for about 20 years
That was a really interesting video.
What I'm wondering now is how I can determine where I would add a TVS in my circuit. How can I determine where it makes sense to protect from these voltage spikes? Do I just add a diode to every IC input?
The general rule I've seen is to add ESD protection anywhere a signal leaves the board in a manner which may be connected or disconnected while energized (based on the theory that the inductance of cables, etc. is likely to cause ESD beyond what ordinary decoupling capacitors will suppress).
My rule is to add ESD protection to all signals which leave the board, at every connector of said signal if board space allows and at one if board space is tight, and to massively oversize ESD protection on power circuits or others in the same connector (this can save your board from being fried if you, say, incorrectly assembly a wiring harness providing power to your board and flip, say, the 5V and 12V rails, assuming you inspect your board with a thermal camera shortly after powering it up the first time). TVS diodes are cheap and very easy to integrate so I choose the safest option where possible.
Then again, for cheaper boards this may not be worth it, and as others have mentioned here, good design practices surrounding decoupling capacitors will stop most ESD from being a problem. Different ICs also very significantly with respect to their internal ESD protection, some (such as low-cost microcontrollers) may be easily damaged while others (such as industrial-grade differential line drivers) may be extremely robust. Checking the datasheet is always a good idea if you're producing a quantity of boards where the cost of ESD protection would make a difference.
Great video, Scott. Packed with loads of interesting info. Thanks for sharing.
"One point twenty one jigga-watts! Great Scott!"
Definitely interesting; I'll use TVS diodes in my projects more. MOVs are pretty common, I replaced a good bunch of these in various devices, which means they did a good job of protecting the circuit.
Just woke up and learnt something about overvoltage protection. This will be a great day!
Won't the shunt capacitnce from the varistor across the supply rail reduce the peak surge from the spark? On a supply rail you can add capacitnce which in conjunction with the short time of the fake static distarge would provide enougbh protection.
It would be interesting to see you build some "crowbar" circuits for smaller electronics so we could see how they also work in protecting electronics.
I've had many components saved by TVS diodes. I have noticed an uptick in their usage even on bargain Chinese fake stuff and they seem to work most of the time except the cheap ones tend to blow up a lot easier then you have to replace it.
at 9:25 I dont get it. you said the MC died but it was blinking fine?
Look at the current flow. It was internally shorted and slowly dying.
Excellent video , could you put all 3 in parallel? That would give speed and high energy?
At 3:02 , did anyone recall the classic Mehdi graph of PVF?