Is ESD overhyped?

Yes, damage due to ESD (electro-static discharge) is real. Just because you haven't seen it isn't much evidence of anything. Some obvious possibilities why you haven't observed the problem are:

1. You failed to create a proper ESD event.
2. You didn't measure the results properly.
3. The built-in ESD protection of whatever you were subjecting ESD to worked, and kept the device from getting damaged.

posted about 4 years ago

CC BY-SA 4.0

Olin Lathrop‭

7964 reputation 18 327 951 466

Real life example. The machine shop of the company where I work has HAAS VF2 milling centers. At 3.5 tons, you'd figure a big enough to warrant some care and respect.

The favored method of data transfer is USB stick.

On a winter day, the cnc programmer picks up enough static on the short walk from his desk to the machine, that it would discharge upon inserting the USB stick. In the older of the machines (mid 2000's generation), this discharge would sometimes reset the whole thing, as if you flipped the power switch. Ruining whatever was being cut, which could be multiple parts in the setup, and if a skinny drill or tap was engaged at that moment, the sudden stop could break that.

That is the kind of thing the EU's stricter ESD requirements are meant to prevent.

posted almost 4 years ago

CC BY-SA 4.0

Pete W‭

350 reputation 9 7 49 118

ESD is a very real concern on mass-produced goods with an expectation of quality and durability (i.e. those items for end-user consumption with warranty). So real in fact that there are numerous international standards dealing with it, including:

• IEC 61000-4-2
• ISO 10605-2008

Design houses will mandate ESD protection on all sensitive ports if these standards are to be met, and during product qualification an ESD simulator (or "gun") capable of producing kilovolts of discharge will be used to test every exposed surface on the product to ensure nothing gets fried if someone shuffles their feet or rolls across the floor in their office chair (while wearing a stylish Cashmere sweater) and grabs your product by something other than its handle.

There will be extensive ESD controls in the production facility as well - workers will wear ESD gowns or smocks, ESD shoes or heel straps, and will use ESD wrist straps when working with product. Trays carrying product (and parts) will be static-dissipative. In extreme circumstances, humidification will be fully controlled.

ESD is really about probabilities. Will one incidental handling with no ESD controls cause an incident? Maybe. Will thousands of incidental handlings without ESD controls cause your product quality and yield to plummet? Most definitely.

posted about 4 years ago

CC BY-SA 4.0

Adam Lawrence‭

141 reputation 0 2 14 0

In addition to permanent damage, which is what you are concerned about, incorrect behavior is also possible. One design I sort of inherited had a major ESD problem. It was long ago and I don't remember the exact details, but there was an input (maybe a reset input or an interrupt for a microprocessor) that was either floating or impropertly terminated. Developers with early access to prototypes reported that it would sometimes reboot when they set it on a carpet and walked around it.

This type of problem can be fixed with a strong pullup or a a few capacitors.

In my experience, the problems that you encounter with production designs in ESD testing (which may not be the same as ESD in the field) usually don't involve permanent damage to silicon. Usually it is incorrect behavior caused by a state transition on an input.

So ESD is not a myth. But maybe the problem is not what you imagine. Also, industry processes are in place to eliminate ESD damage. This includes ESD protection on individual parts and even individual transistors like your IRF510. So it may be that the reason you think it is a myth is because those processes and design details are effective at preventing permanent damage from ESD.

posted almost 4 years ago

CC BY-SA 4.0

mkeith‭

61 reputation 0 1 6 0

ESD precautions can save a lot of tedious failure analysis.

A very inconvenient thing about ESD is that a failure can be attributed to ESD only by ruling out every other possible cause. ESD failure is established by exclusion. One checks every other possible failure hypothesis, and none of them pan out. Only then one can say in a not too confident tone: "Hm... Then it ought to be ESD. We can't think of anything esle."

... I had never seen its [ESD] effects myself.

It's possible that you had failures due to ESD. But you didn't see the ESD discharge. So, you didn't attribute the failure to ESD.

I have seen failures due to ESD first-hand.

posted about 2 years ago

CC BY-SA 4.0

Nick Alexeev‭

541 reputation 4 22 54 47

Something that hasn't been mentioned in other posts yet is that ESD events needn't be utterly destructive. Most ESD events are "microevents". That is, they don't damage a part by rendering it non-operative, but they simply degrade its characteristics.

This is more evident in parts used for their peculiar characteristics or when a part is operated near its limits. Some examples just off the top of my head:

• A high-frequency low-noise BJT could have its noise figure worsened, and you would notice it only when it is employed in very sensitive circuits. For example in the first stage of a smartphone antenna amplifier, where a degraded part would make the phone have more problems receiving signals in low coverage areas.

• A BJT could have its gain reduced, but you would notice it only in very demanding applications. For example if you use a BJTs required to have, say, hFE>400 in an amplifier whose gain is meant to be 10 thanks to negative feedback, you won't notice much difference between a BJT with hFE=800 and a "degraded" one with hFE=600.

• An high efficiency LED could have its lifetime shortened, and you won't notice any effect once you hit, say, the thousands hours of continuous operation. The only effect could just be that a lighting fixture that was designed for 10000hrs will fail after 8000hrs. This is going to matter only for the poor fellow that purchased that fixture, unless the ESD damage affected a whole batch of LEDs and therefore hundreds of fixtures, in which case your nifty fixture will be considered crap by the market, and your company could earn a bad reputation.

• A MOSFET could have its gate leakage current increased, but that won't affect your circuit unless that MOSFET is used as an input stage of a sensitive measurement circuit. So when that MOSFET is used as a switch to drive a relay it won't usually matter, but if that MOSFET is connected to an input line of a GPIO pin of a MCU in a ultra-low-power application, you could have much increased battery consumption.

• The oxide layer of a MOSFET gate could have been only "scratched" by the ESD event, this meaning that the next event is more likely to do destructive damage. Or this could also mean that the maximum GD voltage rating worsened (say, from 30V to 25V). And you won't notice until you use your stock of MOSFETs in an application that comes closer to that limit.

All these kind of "light" ESD events are more subtle and may cause a lot more headaches than destructive events. The fact that these are real problems can be demonstrated by statistical analysis at industrial level, e.g. take two groups of 1000 MOSFETs, expose one to mild ESD events and test them extensively, comparing the results to the first group (the control group).

All this also to say that most of the times an hobbyist is unlikely to observe the effects of these most common ESD events. Your bunch of BC550C BJTs or 2N7000 MOSFETs could rattle in those plastic screw boxes for years (yes, I use them too) and never show any relevant problem. Of course I would never use them to build the safety control circuit for a shiny new Tesla coil experiment.

posted over 1 year ago

CC BY-SA 4.0

1y ago

Lorenzo Donati‭

316 reputation 10 18 60 201

Just to let you know: if one day you have to use the LM337 negative voltage regulator, protect everything with TVS or Zeners. This is the most failing component I've ever seen (and I'm not the only one who thinks so, see youtube). I suspect strongly it is very sensitive to ESD. In my last application, after having replaced the LM337 for the 66th time, I protected it heavily and it finally stoped to fry.

posted about 4 years ago

CC BY-SA 4.0

4y ago

coquelicot‭

704 reputation 17 33 90 142

Yet another example. If using MOSFETs only for power switching applications, it can seem like these are "robust" and "immune" to ESD events. While the package clearly states "ESD sensitive", devices never seem to fail from ESD - why?

Because power MOSFETs have a very large gate and other capacitances - sometimes as high as hundreds of nano-Farads. So an ESD event is complete (a few microseconds) long before one of these pins has a chance to exceed the max rating of the device. The whole device just charges slightly, thanks to the capacitance.

The venerable 2N7000 series of MOSFET however (low power, fast switching) have very tiny capacitances, so are thus very "sensitive" to ESD events. Simply because a "zap" to any pin allows the voltage to rise quickly without capacitively coupling to the other pins. Out of all common jellybean parts, these might be the most sensitive, and can certainly fail just from being handled without proper ESD protection. As the other answers indicate, "failure" can be anywhere from "total short" to "slight degradation of performance." In general, the faster the MOSFET, the lower the capacitances, so thus the more the ESD susceptibility. Same for IGBTs and any other silicon - lower capacitance = higher susceptibility.

I have a chair which generates static electricity (triboelectric effect) whenever I get up from it (in dry air.) I can get up, walk over to an electronics workbench with ESD mat, and "zap" a board sitting on it (due to the capacitance of the board.) This one zap could certainly destroy something. That could happen if I were careless.

I could instead walk over to it, touch both hands to the mat, then touch the board with no issue whatsoever. The charge would be quickly drained from my body due to the slight conductivity of my skin and the mat. So part of "ESD Protection" is being cognizant of what ESD is, how / when / where it is generated, and how to ensure you won't accidentally zap something. Even consider that your backside clothing could still have a charge on it (especially true if wearing something like a jacket), and backing up to the bench could touch this to the board and cause damage - even after discharging yourself.

So yes, it is best under any condition to sit down at an ESD-safe workstation, wear the high-resistance-grounded wrist/ankle strap, and then perform the ESD-sensitive work. This gives relatively good protection from most conceivable ESD events. Barring that, just being aware and careful goes a long way also. If assembling one-off prototypes or the occasional hobbyist board, a full-fledged ESD workstation is impractical. Likewise, large-volume PCB assembly houses all have and use ESD workstations, as +2% failure rate due to ESD (really, operator laziness/ignorance) would cut into their already tiny profit margin.

posted over 1 year ago

CC BY-SA 4.0

rdtsc‭

141 reputation 0 4 14 1

Many ESD problems are "latent" that is they do not sofort show up....some time later things will stop working :)

posted over 1 year ago

CC BY-SA 4.0

dekker‭

67 reputation 4 1 11 0

I know of a designer that didn't add ESD protection on a mass manufactured product in the field. They had to replace almost 10 thousand units, because they kept popping in certain low humidity and when people used them in certain ways. So you'd rather be safe than sorry. An expensive error for a couple cent part.

posted over 1 year ago

CC BY-SA 4.0

Hawk2020‭

66 reputation 2 2 8 1