SSTC pt. 9 - Theory (why your transistors fail)

Today, some theory of snubber networks and a halfbridge Solid State Tesla Coil controlled by IR2153 chip, but most of it applies to any half bridge or full bridge (H-bridge) SSTC or any other switching device like switching power supplies or inverters, induction heaters or welding inverters. What destroys the MOSFETs or IGBT transistors. Do your transistors keep blowing? Design it properly and burn no more silicon :).
The first episode:
• SSTC pt. 1 - intro
The next episode:
• SSTC pt. 10 - Gate dri...
Please support this channel on Patreon:
/ diodegonewild

[@kolinevans9127]

Your cat is ultra engaged with snubber networks and miller capacitance.

[@zaraak323i]

"If your transistors are blowing it's either because you have no snubber networks or your transistors are from eBay!" Truer words were never spoken! Thanks for the laugh.

[@Betterstartnow]

Your cat is jealous of the time you spend on the electronics! Love it. This is good stuff man!

[@nakrul987]

cat needs petting

[@1pcfred]

Or it wants to be fed. Mine gets real chumy when I forget to feed him.

[@tomashrabe9874]

lets make another 10 second video, when your cat says: your transistors are blowing up, because they are from ebay :D

[@DiodeGoneWild]

It's almost 2 months since I ordered some FGA60N65SMD and HGTG30N60A4D from Ebay. Maybe Ebay transistors don't blow up... because they never arrive :).

[@johnflower8384]

@@DiodeGoneWild What if high power schottky diode is used in antiparallel connection with the C-E of the IGBT, or in parallel with the internal diode of MOS?

[@Dzejwor_Gromowladny]

@@johnflower8384 Nothing ;) it will gona blow anyway because you have two diodes in parallel - one extremely fast and one extremely slow ;) and recovery time of circuit like this is pretty much the same like slow diode :) With mosfet you need a shotky in series with mosfet and fast recovery diode in parallel with whole this series circuit. With igbts things are diferent - they usualy have good quality parallel diode that is a separate silicon structure independend from IGBT itself and boned internaly to C & E of the switch.

[@gorjy9610]

​@@DiodeGoneWild you ordered them at wrong time. Everything I order before 2 months never came. Everything buy'd before and after (let's say before 2,5 months and after 1,5) ist delivered. My personal conspiracy theory is that since about that time China started to export masks and all that crap they (who are they? Nobody know, it's a conspiracy theory after all) just throwed away all post in a garbage to make room for face mask shipment.

[@akkudakkupl]

Well I blew up a genuine Semicron halfbridge 1200V module, so the source is only half the problem...

[@manuelengel3680]

6:55 What an angry cat. Why does it not make the video itself if it knows everything better?! Jokes aside, this was a great and informative video for me, especially given that I am far away from being an expert.

[@buffplums]

You have a brilliant way of explaining things with plenty of diagrams, you are a very clever gentleman sir thank you. Your knowledge has taught me a lot and helped me understand some of the problems I’m experiencing in my BOOST converter design

[@Georgy-fg3bg]

"Everything You Always Wanted to Know About SSTC But Were Afraid to Ask. " Qualified analysis. Excellent explaining. Congratulation!

[@glmnet]

Didn’t watch video yet. But already know the answer: “the cat ate my transistors”

[@maxsilvester1327]

*the cat sat on my transistors

[@PapaMav]

The cat puked up a hairball on the transistors.

[@blowingchannel4424]

Now full schematic and explanation of how your cat works ;)

[@AlexanderBukh]

well, one thing i can say for sure, it resonates

[@k33per03]

"Random ideas in random oorrrrrdeeeeeerr" ...the best design philosophy.

[@eDoc2020]

Thank you for this video series. I am learning a lot about the tricky aspects of power switching electronics.

[@marcogeri]

This is simply a great explanation on mosfet switching, snubbers, zvs and other advanced power electronics topics. Very nice job indeed, you are very good in trying to explain with simple words a very complex subject; congratulations!

[@alfagulf]

Wow, learned so many things in less than 20 minutes. Well done, great job. 👍👍👍

[@user-js7ip7hf1u]

Your videos are very instinctive . Thanks for sharing

[@electronic7979]

Helpful video 👍

[@edgeeffect]

A while ago I watched Alexander Hartdegen's videos about a DIY MOT spot welder. He stressed the use of snubber networks to protect the timer/switching circuit. Watching your SSTC series, I almost understand what this is about now. :)

[@gacherumburu9958]

Thanks for the info! Very educational..👍

[@clubunity9342]

Cat in the starting : May be daniel will need my help.

[@kriswillems5661]

I love your channel. I am an electronics engineer, but I am still learning from you. Great accent too. Love it.

[@johnconrad5487]

thanks for making this good video.

[@dominikgula1840]

@DiodeGoneWild Thank You!!!

[@hoggif]

Way back when I designed one, I had tank capasitors near transistors. I used polypropylene caps very near transistors as the tank. I measured very low voltage spike with mosfets. I used only mosfets parasitic capasitance to make zero voltage switching or only small parallel capasitors in parallel with mosfets to make tank with stray inductance and mosfet+parallel capasitance. I had pretty heavy drive on mosfets. It took me a while to figure out miller capasitance than needed pretty high drive for mosfets and/or some level shifting like outputting +12/-15V after driver transformer I used. A bit lower off voltage was the key with my device. I think you explain thing very clearly. That would have saved me about 100€ or more current money at the time I did mine. Very solid design I think.

[@oniruddhoalam2039]

Why did three people dislike? Nobody explains such stuff so nicely as Danyk.

[@maxsilvester1327]

And nobody has a funnier cat

[@walterschlupp154]

Cheap Yihua dc power supply ps-1502dd keeps blowing output transistor. Do you think there is some simple solution to it? Thank you.

[@cezarcatalin1406]

High power zeners and two ceramic capacitors in series with two high current antiparallel diodes in parallel with resistors is the best snubber network I ever found. It does everything: - absorbs voltage spikes - clamps voltages to safe levels for transistors - doesn’t allow ringing - limits the reverse charging loses

[@WolfgangMahringer]

Can you post a drawing what you mean? Thanks!

[@kamranjavaid]

Your cat is the master of snubber network

[@christopherthumm4046]

Please do an episode about your cat !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! He or she deserves a special episode

[@oniruddhoalam2039]

Diode, can you please make a video on the disadvantages of using cheap digital oscilloscopes you reviewed.

[@dannyperry8070]

how did you determine the value of the parallel capacitors?

[@richardnanis]

I just dicovered your improvement with the Gate-Current-Amplifiers at the end of the video? Could you tell us more about this trick? Is the switching time reduced as indicated on the paper? This would be a great effect and lead to less power loss in the mosfets/igbts. I could try to use this in my royer converter for building an induction heater.

[@Miata822]

More great stuff but I'm starting to get intimidated. I have long wanted to build my own VFD to power 3-phase motors from a DC source. These are commercially available, of course, but it would be more satisfying to build it myself. I am starting to realize what a large task that will be. Your H-bridge circuit, though, is wonderful. I'm sure I will find uses for it, perhaps a Jacob's Ladder for my mad science lair.

[@drdrums1]

Your cat reached resonance at the end.

[@ohmscorner3635]

Great content as usual. I learn in each single video you made. #ohmscorner

[@kemalkurt5257]

Maybe you can make some VTTC after finishing SSTC because i think vacuum technology is really intersting and needs more work (i know you have made VTTC back then ) if it doesnt bother you could make one in this channel

[@marksasahara1115]

Waaaiiiit a minute. This is just another damn cat video. LOL! Come for the electronics, stay for the cat. Keep up the great work, man!

[@buffplums]

That’s interesting about the body diodes not clamping the overshoot. This is exactly what I’m experiencing. I didn’t want to use snubber network as it traps all that wasted energy but I guess if you have loads of parasitic oscillation yo have to capture it otherwise the IGBT will die.

[@richardnanis]

Great theory as always! I lost the red thread somewhere in the middle. I know the ebay-transistor phenomenon :-) Just had a power transistor blow up in my royer converter on an induction heater. It was a IRFP4768 blowing up at only 10A while it is constructed for handling 66A @ 100deg C. Hmmm - just ordered new ones from LCSC - they should be real ones i hope... Even simple circuits can have their secrets - but royer converters or half bridge/ full bridge supplies are everything but simple...

[@johnconrad5487]

It is interesting that you made a Royer using MOSFETs. Royer is usually good for Bipolar.

[@richardnanis]

@@johnconrad5487 Why do you think that? Many Royer converters are made from mosfets - or sometimes igbt's. The last is better for higher voltage as mosfets deliver high current but rarely high voltage and high current at the same time. The backdraw of igbt's is they are slower and have higher input capacitance - so they cant be switched as fast as mosfets. But they can handle 100 - 200 kHz nowadays.

[@Antagon666]

@@johnconrad5487 Best thing is to buy 50 IGBTs at once, but 5 of them are fake, I tried pair of fake ones in my DRSSTC, it ended up converting the IGBT into a flamethrower.

[@ElIsraelJijijiji]

Can the IR2153 supply a 500khz Frequency???

[@haroonsyedmohammed8430]

Please tell us about yourself! From where do you get this knowledge, I bet no textbook have this thing. Your knowledge is very inspiring.

[@laharl2k]

Haroon Syed Mohammed There are lots of books about electromagnetism. They are very long though

[@haroonsyedmohammed8430]

I think this is power electronics stuff.

[@laharl2k]

​@@haroonsyedmohammed8430 The tesla coil is like the name implies a coil, that falls under electromagnetic theory, and if you see the whole device, its no more than a spacial kind of transformer. Switch mode power supplies use transformers and/or inductors to convert power. The tesla coil uses a transformer. Transformers are just the theory put to practice

[@laharl2k]

@glyn hodges I completely agree, i also read lots of books about electromagnetism but even if i understood the theory i couldnt figure out how to use it to make a transformer. It wasnt until i found a book specifically on switch mdoe power supplies that i understood how it worked and of course it explained things that werent in the theorical books :/

[@laharl2k]

glyn hodges No clue about the authors. Im care more about what people say rather than who says it. The pdfs must be in my digital library but thats on my linux partition and im not restarting the pc until next year, but i remember i got them all from libgen.io searching for magnetism and electeomagnetism. There wasnt a lot that were relevant so it should be some of those. The name was something about "switch mode power supply designs" iirc

[@atose_offline5463]

How can I send you a file?

[@SeanBZA]

How about snubbers with a diode, and also a much faster diode across the IGBT so that you do not have to use the slow body diode. Allows more energy to be absorbed in the snubber, but does not have a high turn on current spike to discharge the capacitor, as most of it's charge has already been dissipated in the resistor.

[@nightrous3026]

Seems to be going quite well. Expect some trial and error. We all make mistakes and forget a connection somewhere sometimes.

[@williamsquires3070]

You have to add a snubber network to slow the rate at which the cat sits on the schematic. 😊

[@ag6371]

The cat is in your thumbnail

[@DiodeGoneWild]

I was thinking hard what to use as a custom thumbnail, and then the cat appeared in the autogenerated one and that was it...

[@followtherules4808]

0:43. Too much theory. Time to flee. 😂😅😆

[@maxsilvester1327]

I see your comment twice (this is number one)

[@ElIsraelJijijiji]

If the transistors blow they blow the chip right??

[@gulshanbarange7968]

Diodegonewild Do a Question and answer (Q &A) video

[@Martinsp16]

great

[@rnbv3418]

Danny, what is your level of formal education? Just curious. Your expertise is amazing.

[@DiodeGoneWild]

in the field of electronics, none :)

[@drobotk]

Well, I don't know if it's idiotproof. I actually killed the chip somehow. At 25V it worked absolutely fine, nothing was hot or anything. Then I raised the voltage to about 50V and poof! Chip is gone. I guess maybe I do need to add that resistor after that UF diode? What do you think might have been the problem? I'm powering the chip with a separate 12V power supply. I have all of the snubbers just like you showed. My mosfets (2SK2698) didn't blow up. The chip isn't even in short circuit, it just doesn't oscillate anymore. Let me know what you think Dan, and thanks!

[@DiodeGoneWild]

Are you sure the diode was UF4007, not 1N4007? Do you have 100nF ceramic capacitors close to the power pins? There should be one between pins 1 and 4 and the other between pins 6 and 8. Plus also an electrolytic capacitor between pins 1 and 4. It's also necessary that power connections and control connections of emittters/sources of the transistors are not shared. They should split right at the transistor pins. Some IGBTs even have a Kelvin emitter for this! Did you use zeners for the gates? They clamp the voltage and also limit negative overshoots. Also the connections to the sources and gates should be as short as possible and go close to each other, not to make a big inductive loop. If you use wires, they may be twisted for less inductance of the loop.

[@drobotk]

@@DiodeGoneWild Oh, that makes sense. I totally forgot about the decoupling capacitor for the chip. Also, I used a BA159 diode not a UF4007, but I have seen you have used it there in your older coil. Anyways, I already ordered a new chip. I'm gonna treat it better this time haha. Thanks for your diagnosis :)

[@DiodeGoneWild]

Yes, those 100nF capacitors are quite important. And it worked with BA159, but UF4007 is even faster.

[@makoveliprod]

At high frequencies you must used more fast diodes, like a HER157, don't use 1n4007. Perhaps 1n4007 were re-marked in UF4007 :) Also, bootstrap capacitor must be high quality, and gates resistor have not too low resistance. And put a 15V zeners between gate and source of mosfets.

[@johnconrad5487]

@@drobotk BA159 is not a fast diode. Check trr = 500nS compared to UF4007 which is 75nS

[@evabaroni6693]

Igbt body diodes also aren't the fastest or most robust, particularly in earlier generations. eBay and AliExpress parts are generally rubbish. Some time you get lucky with the left overs of a production run where the OEM purchased via official channels.

[@qzorn4440]

i hope kitty has 9 lives on the workbench...🥰 meow... very interesting video...😊

[@tetraederzufrequenztesla2342]

Good video Friend. Definitely, the supervisor engineer Cat is agree with you, but he always wait for a good test... How can we build a circuit tesla coil that generate a magnetic scalar wave?

[@1pcfred]

This guy is building a new woodpecker over the horizon radar system. Duga 2 electric boogaloo!

[@minazulkhan8287]

live long and continue making videos

[@konstantinkokkorakis1907]

Lol I love the cat you are so nice

[@bachichunduryify]

Yes

[@gyulakurucz8331]

Hello Mr.Diode (Danyk?)! Currently Iam designing an adjustable (0-40V, 0-10A) SMSP, using tl494 chip at 100khz frequency, using half bridge topology. Unfortunately under small load (6V,5A) the FETs exploded. Someone recomended me to use series capacitor, which makes resonant circuit with the leakage inductance of the primary. I didnt used snubbers. Can you check my schematic and comment it please!

[@makoveliprod]

With high load it works nice? What kind of drivers are you used to control mosfets? Transformer? Better to look at similar schematic solutions in computer SMPS and cheap chinese SMPS.

[@gyulakurucz8331]

@@makoveliprod I used 6n135 high speed optocoupler, then a uc(or tc idk)4429 fet driver to restore the signal to squarewawe. I used the optocoupler to isolate the ic from the fet, because it has direct feedback to able to adjust the output to 0v.

[@gyulakurucz8331]

@@makoveliprod I havent tried it under higher lod, because as I said the fets (irf740, original) exploded 😃

[@makoveliprod]

@@gyulakurucz8331What power supply method do you use for mosfet drivers with optocouplers and tl494? Perhaps at low load some of these circuits does not get enough power? When TL494 shutdown (no or low vcc), both mosfets is off?

[@makoveliprod]

@@gyulakurucz8331Use ~100W lamp in series with mains to protect mosfets from exploded

[@wingnut2893]

Two capacitors in parallel add up to the double capacitance. Two equal capacitors in series result in half the capacitance of one of them. Unlike resistors...

[@raajandear7284]

I like it cat

[@astromancer]

"Your transistors are from ebay" I'm going to use it as an insult from now on.

[@Dzejwor_Gromowladny]

In my opinion two very important things: first - any solidstate tesla coil will never work stable with fixed frequency generator on primary because secondary's resonant ferquency is NOT a fixed thing ;) when disharge grows it adds a lot of capacitance to the topload and resonant ferquency drops a lot and the changes are greater the smaller the resonator is - secondary feedback is a MUST in SSTC;) Second thing: bootstrapped gatedriver with fixed deadtime like IRxxxx ICs is not a good choice for any sstc type because it wont work stable in this condidtions and have trouble with proper work with secondary feedback topology :).

[@fumthings]

wow, 50 percent theory, 50 percent cat video for the win.

[@wolfgangboettcher3126]

Siehste und schon geht es weiter

[@g-gon8869]

What is your cat's name?

[@rickjljr11]

I would suggest you take your scope and verify all the statements you make about current and voltage in the circuit. I think you'll be surprised.

[@overbuiltautomotive1299]

my chinessey tig welder quit they sent me weird schematics lol with no labeling on it just drawing basicly no values on it .uh gerrr darn them lotos tig200 ac/dc

[@leozendo3500]

This video should be renamed half-bridge passive ZVS ZCS technology.

[@mandovibrazrocha2295]

Cat. Nice companion.

[@remontlive]

Like for you

[@gordonwelcher9598]

KZbin payment should be: $$$ per 1000 views * number of cats in video.

[@borislicic]

Nikola Tesla didn't make schematic and drawings, he makes a machines out of his head and never made a mistake.

[@schwarszendeust]

And of course here is my cat

[@lucassamuel6069]

Here is my caaat**

[@andricode]

Here is my caaaat**

[@miaudottk9080]

How did you teach the human to act as if explaining electronics, meow?

[@nowayjerk8064]

more cat please thank you.

[@sfp4197]

The engineer cat can learn me ? I killed mosfets wort 200$ at least i bout them at 50$

[@JBW19951]

0:27 I guess the inductor has about 50 uCats. 🤔

[@astromancer]

Ok, this might be a little bit too advanced for me.

[@Clasiczzz]

Like for the innocent cat ❤️

[@khelifinouredine1842]

In this circuit there is a dead time ,, so that the mosfet does not burn

[@petehiggins33]

9:47

[@AlexanderBukh]

tabby cat being tabby

[@nsfeliz7825]

so your cat is an electrical engineer?

[@saschakolb7102]

Anyone may notice, that analog technology is far more exciting than the digital technology. With Analog you handle with physical borders. With digital technology you just handle borders other persons limiting you.

[@Kungfupanda839]

Please do a video about your cat 😍

[@n0madfernan257]

clicked for the cat-pacitor... i get out now

[@handlesrtwitterdontbelivethem]

this cat is also in my home

[@perolozac01]

Seems like your cat is determined to stop you from building this. Maybe she knows something...

[@davey2k12]

Good point eBay used transistors lmfao 😂

[@DiodeGoneWild]

Used ones are the better option on Ebay. You can get ones with a lower current rating and a changed marking.

[@davey2k12]

@@DiodeGoneWild special Chineseium design 👍😀

[@tukangunduh]

Too many theory... 🤣😀😁😆

[@devenyi73]

Cat need to speak more. It is funny.

[@TechHowden]

3rd

[@johnkubik8559]

Showing your cat in 5-10% is cute, having it hiding schematics in technical explication is really annoying. You're talking about about all your connection lines been inductive yes they are, in the 100Mhz+ it would be a major problem but we are in the 100Khz range, think in term of transmission lines on your PCB, strait and keeping the return path next to the outgoing path not just as an inductance, it has a capacitive and impedance factor to take in account.

[@DiodeGoneWild]

100kHz sine wave would be fine. The sudden changes in a square wave are the problem.

[@sheetalmekhale]

It's not the frequency, it's the rise and fall times, in MOSFETs they can be in the order of few nanosecond's, the gate resistor is used to slow things up.

[@Trinitrophenylmethylnitramines]

Your circuit blown up because you don't have a Catpacitor

[@ourchannel8215]

🤣😂🤣😂🤣😂 from ebay...

[@kamote6375]

1st

[@drobotk]

Not really, I was here yesterday

[@maxsilvester1327]

@@drobotk how?

[@drobotk]

@@maxsilvester1327 Patreon

[@gabrielhacecosas]

I really hate that cat 🔪🔪🐱🔨🐆

[@sfp4197]

If you vant to plug it in to the wall the cat should make your circuit

[@sailorondatea]

eww u have transistors from eebayyy...😒

[@followtherules4808]

0:43. Too much theory. Time to flee. 😂😅😆

[@maxsilvester1327]

I see your comment twice (this is number two)