What Does ESD Do To My Circuit and How Can I Protect Against It?

[Kevin Darrah] is risking the nerves on his index finger to learn about ESD protection. Armed with a white pair of socks, a microfiber couch, and a nylon carpet, like a wizard from a book he summons electricity from his very hands (after a shuffle around the house). His energy focused on a sacrificial 2N7000 small signal MOSFET.

So what happens to a circuit when you shock it? Does it instantly die in a dramatic movie fashion: smoke billowing towards the roof, sirens in the distance? [Kevin] set up a simple circuit to show the truth. It’s got a button, a MOSFET, an LED, and some vitamins. When you press the button the light turns off.

He shuffles a bit, and with a mini thunderclap, electrocutes the MOSFET. After the discharge the MOSFET doesn’t turn the light off all the way. A shocking development.

So how does one protect against these dark energies out to destroy a circuit. Energies that can seemingly be summoned by anyone with a Walmart gift card? How does someone clamp down on this evil?

[Kevin] shows us how two diodes and a resistor can be used to shunt the high voltage from the electrostatic discharge away from the sensitive components. He also experimentally verifies and elucidates on the purpose of each. The resistor does nothing by itself, it’s there to protect the diodes. The diodes are there to protect the MOSFET.

In the end he had a circuit that could withstand the most vigorous shuffling, cotton socks against nylon carpeting, across his floor. It could withstand the mighty electric charge that only a grown man jumping on his couch can summon. Powerful magics indeed. Video after the break.

32 thoughts on “What Does ESD Do To My Circuit and How Can I Protect Against It?

  1. If you really want to know how tough such protection can be, check out the DO-160 Aerospace requirements for lightning induced transients. Short summery: one of the worst transients spikes is +/- 1600V with 1 ohm output inpedance for 200us, 50 times in 1 minute.

    For automotive specifications, check ISO 7637.

    Transient protection is about more than just ESD, but the basic nethod is the same: clamp voltage, limit current.

    1. More than ten years ago, I worked for a government sub-contractor that was mostly in the defense industry. I was working on an antenna system for fighter aircraft at the time. I had to take this antenna system into a radio chamber (one of those huge metal rooms with funny angled foam on all of the walls). The funny thing was that this was in Florida, where the humidity means that static electricity is almost never a problem. However, in this chamber, somebody had placed wood planks for walking over top of this foam stuff, and just walking on the wood would generate enough static to create 3/4″ sparks when I touched metal. I was actually terrified at handling the antenna in that kind of environment. Still, everything survived — we had some beefy protection diodes on there.

  2. This is a great into to the real world of electronics. Most micros, logic chips, etc have this protection build in but its almost always a good idea to add it to almost everything no matter if it doesnt like a transistor or does. This configuration is classic but you can do it other ways too. Just think of it this way. Using diodes to shunt current to a low impedance path like one of your power rails or ground based on the different polarity of the shock. BAT54S is a good one for this and some thin film bigger package resistors.

  3. If you look hard enough, you can find discrete MOSFETs with built-in ESD protection (zener clamps) on the gate – at least for power devices, not sure they exist for small-signal. Why aren’t such self-protecting devices more popular?

    1. Reverse-biased diodes (which includes Zeners when they’re not conducting) act like varicaps. Small ones, usually only a few pF, but when you’re working with high-ish frequency stuff, those few pF can get you in trouble.

    2. Because they can be a pain in the neck if you don’t need them there. If you don’t expose the chip to off-board connections, ESD shouldn’t be a problem unless someone is poking around on the board. And having the diodes there means that you’ve got a few pF extra of capacitance, and you’ve got a parasitic path back to the power supply that can back-power the chip, which sucks if you’re trying to shut sections of the circuit off.

    1. One of the worst was SainSmart. I ordered TFT displays, 2.4GHz transceivers, etc. and they arrived with the pins stuck in white foam, in one case literally packing peanuts, wrapped with packing tape and stuck in plain clear Ziploc bags. Not a one of them worked 100%, some were DOA.

      Between the static charge created by the tape and foam, truck tires, and all those rubber belts along the way, I’m not surprised they were damaged.

      I gave them holy-hello about it, they are much better now.

  4. so many datasheets for parts i use list est protection as a main feature. makes me wonder if those pc builders being all anal about their writs straps know its for naught. i for one have never used one, and i cant recall ever running into an issue caused by esd.

    1. Sometimes ESD causes small dendrites to grow between traces in the silicon, sometimes ESD causes a small crack in a trace in the silicon. Either of these may not be noticed when they first occur, but the damage has begun; and dendrites can become shorts, and cracks can become open circuits.
      You may not know if a component failure was caused by ESD damage two or three years ago…

    2. At the factory, if the customer has any failures at all, their initial reaction is to blame us for all of the problems.
      If we didn’t use ESD wrist straps, they would claim the cause must be ESD and must be our fault, even if clearly they sawed the PCB in half or something.

      A few ESD wrist straps are much cheaper than arguing a battle like that.

      1. I have run….
        10x assembly runs over 6 months. 5x at facility A and 5x at facility B using the same reels and they have almost identical processes yet facility A does not have almost any ESD protection and facility B has warehouse wide ESD protection with leg straps etc. I got 20% failure from facility A and 3% from facility B. The boards looked identical yet facility A failures were bad chips due to unknown circumstances. Its possible that they damaged it another way but I prefer to use a facility that has ESD protection because of this. The reality is ESD can be an issue for some products yet others its not as important.

    3. As long as you touch something grounded before handling electronics and are not sitting on a chair with synthetic fibers, you are *probably* OK. ESD protection is standard in the I/O pads on all chip libraries.

      Still, with that being said, I once killed a motherboard by an ESD event into a USB cable. Killed the motherboard completely.

      1. Let’s examine what you just said. As long as you touch something grounded, you are OK. But an ESD event through the ground shielded USB cable killed a motherboard. Anyway, I think many people find the whole “touch something grounded first” a mystery. I have known people who would walk outside, touch a metal fence post, then walk back inside to work on their project. Thinking that touching grounded metal somehow magically protects them.  Steve Greenfield AE7HD http://www.linkedin.com/in/stevenjgreenfield

        1. Its all about potential difference. Grounding is only valid if the item you are soon to touch is also at ground potential. It would be acceptable if you were floating in the air to hold a 115kV line because the potential difference between you floating and the 115kV line is zero. (neglecting initial bonding woes) So as an example if someone has a project in their hands and you want to touch it too you need to be at the potential they are NOT ground. By touching then you are now at the same potential.

          1. Exactly. Better to touch the outer casing and keep contact while working, than to ground yourself and -not- ground the case. Grounding wires to wristbands have 1M resistors built in, partly to protect you from being electrocuted by a live circuit, and partly to limit the current of any static discharges. This reduces induced currents, too.  Steve Greenfield AE7HD http://www.linkedin.com/in/stevenjgreenfield

      2. I once killed a brand new cheap TP-Link router with ESD. I was doing the initial setup with a LAN cable plugged into my laptop and the router sitting on the floor by the easy chair. When I finished with all the settings, I unplugged the Ethernet cable from the laptop and stood up from the chair while still holding the end of the cable. I felt a spark, and that was the end of ALL of the wired LAN ports on that router.

        The port that was hit directly is dead. The others sometimes work a bit but are very unreliable. If you refresh the router page in the browser about 20 times you might finally get all of the page to load.

  5. When describing ESD to other people, I usually say something along the lines of, “You’ve got a part there rated for 3.3v which can tolerate a maximum of 5v. What do you think will happen when you smack it with the 15kv generated by your fuzzy socks?”

  6. Absolute worst ESD sensitivity I ever encountered was on discrete FETs used on Fluke 8620 bench meter. Seemed like no matter what I did, I always blew some up. Full ESD antistatic station with wrist strap, desk pad, ion blower, conductive lab coat and lab humidifier. Sensitive little suckers!

  7. Really good video and clear walk through of the experiment. I have a home made security system and the sensors coming in from outside switches cause my controller board to heat up after a lightning strike. I tried adding some series resistors but now I can see from your video why they are having no effect. I have seen the arrangement of two backwards diodes before, but your explanation of how they shunt the current to the power rails is clear. I can now see a solution to my problem rather than just trying random components and then waiting for a lightning storm! Appreciate you burning your fingers to educate us :-)

    1. ESD and lighting strike are really two different things in terms of severity. To survive DO-160 they do a lot more than just clamping with a diode + resistors. In your case, you also have the option of replacing long wires (like it seems you have in your setup, by long i mean a few metres) by a wireless connection. Long cables act as antennal, and are often the main input for esd in a system. You can try to protect the system with diode clamps sure, but they can be tricky to size, so why not removing the main exposed input, gonna be easier ;)

    2. You might be experiencing a latch-up. It is a nasty phenomenon going on in CMOS IC’s, which all contain a giant intrinsic SCR between their power rails. It remains off in normal operation, but a solid zap turns it hard on, causing excessive supply current – though the device might return to normal after power cycling.

  8. I was expecting a little more information here. I’m used to run tests according to IEC 61000-4-2. The problems coming from ESD are not just the high voltage applied directly to the components, you could have mentioned something about the induced fields that you generate when you discharge. A lot of circuits do get into trouble because the pretty high frequency pulse generated and coupled into your signal transmission lines causes glitches. That reset-pin on your microcontroller you’ve only pulled up with a pretty high value resistor is a good target, we always add at least a 100pF and a >100nF cap right next to the controllers reset pin to shunt these coupled HF events to ground and keep the reset pins stable.

    1. I fully agree, but it is a nice stepping stone for all the hackaday visitors that are not familiar with (or fail to acknowledge or chose to ignore) the importance/horror of ESD. I see an opportunity for hackaday to write a series of articles to address this issue. Hackday has written more educational stuff that really helps the beginner and refreshes the mind of the expert. I really think that those articles should be “bundled” in a corner of the website and could act as an aid to everyone who starts with electronics.

    2. As Jan suggests, while lacking in real technical meat, it’s a good starting point to explain the issues and things to be thinking about to protect devices but something that HaD could really build on and go in to some high levels of detail.

      If one is just starting out (or only have something like this video for guidance), something that needs to be watched out for is that not all ESD protection is created equal and it’s easy to be misled into thinking that because a part states some level of ESD immunity that it’ll be fine. I’ve just been finding this out for myself while doing some 61000-4-2/HBM testing of some I2C isolators that are specified to be safe to +/-8kV, still die pretty quickly in the design so I’ve been zapping a variety of different ESD protection diodes to find the best/cheapest/smallest additional logic that will keep it all working.

      1. I fully agree and would like to see some more articles with some deeper informations about what actually can happen with your circuits or inside your components to follow up on the simple “intro”. Damage that does not immediately show, but maybe shifts your component performance out of spec, or leads to premature failures later on. Or something about capactivie and inductive coupling of EMI-events inside your circuits (and why it’s a good idea to put protection devices at the right place and hook them up with more than a 2mil wide track).

        And yes, you have to be very careful with manufacturers listing ESD-protection in the datasheets. The protection usually only applies to some of the IOs, and if you have a fine pitch component and apply a few kilovolts to the “protected” signal port, the ESD can just “jump” over to some unprotected pins of the same component and zap it dead anyway… And always remember that the current that should flow thru the protection diodes inside your IC wants to go somewhere. If your VCC/GND connection is bad (high impedance, no capacitor directly next to the IC,…), the diodes can not really do theyr work because the current can’t (or does not want) to go anywhere.

  9. Another nasty thing about ESD is that it doesn’t always cause an obvious, immediate failure. You may zap something and it powers up and behaves fine on the bench or an idea environment, but decide to go into its failure mode at 30,000FT at few degrees above zero or in the middle of the desert well between rest stops. Best practice is to treat all active components and populated boards with actives as ESD sensitive regardless of if they are or not. It’s a pain and I wouldn’t do that at home or a situation where it doesn’t matter too much, but in a lot of industries it matters a lot.

  10. i know this sounds hard to believe, but i had a 486 based system with flash-memory bios and auto-recovery, static took out the bios, dont know if it was damaged or just corrupted. the auto-recovery greeted me with a request for a floppy with the correct bios. too bad the internet did not have a copy of the bios at the time as back then people didnt upload things just to be the one who uploaded the most and my model wasnt up there.

    and i know it was static damage because it happened when i reached into the case while on , causing a crash (non-cached bios), and on reboot gave an error.

    lukily the damage was limited to the bios and all other data/hardware were salvaged.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s