Static electricity and discrete semiconductors

[OneRoomStudio]

As I got another parts shipment from Mouser today, it occurred to me that over the years, I have brazenly ripped open the antistatic packaging that diodes, transistors, regulators, etc. come in without any anti-static precautions at all - at this point I have handled many many thousands of components, and to my knowledge, I have never had one that was damaged by static. So I wonder - are the dangers of static electricity vastly exaggerated (to the point of every part coming in its own anti-static bag), or have I just been extremely lucky to date?

 

[Philbo King]

It depends...

It once took me a week to find why these $1200 FPGAs were dying. Then I acidentally captured an 800V transient spike on the oscope after rolling the chair back 10 feet and standing up in my polyester pants.
After I reported it, they suddenly spent a fortune on antistatic chairs...
It got me out of soldering hell. They were 168 pin grid array (thruhole) chips, and I probably changed it out 12 times before I discovered the problem.

 

[ccaudle]

You have probably been lucky to a certain extent, especially if you live in MN and handle parts in the winter.
Bipolar power parts like regulators do tend to be tougher, MOSFETS can get damaged on the gate pin pretty easily. Also some transistor damage doesn't necessarily kill the parts, it can just shorten the life or make them noiser than they otherwise would be.

 

[Brian Roth]

I don't recall where I read this.....maybe on this forum...someone stated that for a company like Mouser, it makes perfect sense to package all semiconductors in anti-static pouches. One reason...no need to keep two different types of packages (and "flagging" every SKU for anti-stat or "not").

Bri

 

[pucho812]

i have been told that discrete circuits are esd safe and chip circuits are not.
I never was in a position to test that and I could never find a definite answer from anyone I asked. The web isn’t much better in finding that info, any searching can yield unique results. As for me I try to use best practice with an esd strap as much as possible.

 

[Philbo King]

MOSFETs (anything MOS, actually) is susceptible to gate-to-channel punch through. It can happen at surprisingly low voltages because the gate insulator is so thin that the voltage gradient across it is quite high.

 

[PermO]

Broadcast systems I have worked on ESD is mandatory, so you have to wear the shoes and the special labcoat, the ESD floor is treated weekly, you can't just mop the floor.
Outside the room is an ESD test system, you stand on a metal plate and put your hands on the test device, if you get a red light the door will not unlock.
Everything you bring into the room must be ESD complient, no packaging materials are allowed in the room.

They have done a lot of research and with a microscope you can clearly see the effects of static discharge on components (integrated circuits).
The effects of these damages creep in over time, you can simply no longer trust your equipment, random faillures are not random, there's a reason for everything.

In broadcasting it's all about continuation of operation, there's a bunch of legal contracts saying blackouts should not happen at any time, and blackouts will be fined.

Most of these systems run from -48V PSU... When I asked why all these systems are powered from a negative rail they told me it slowed the corrosion of the big bare copper busbars compared to a positive rail PSU.

I would not worry about my resistors and my capacitors.

 

[ruffrecords]

Nearly 40 years ago there was a problem in the UK with the petrol pump system of a well known brand of fuel. There would be occasional complete system crashes at a petrol station. These were very serious because all the cumulative data was lost so they did int know how much money they had taken nor did they know how much fuel was left in the tanks. Long story short it turned out these outages only occured in the winter and were caused by a combination of pantyhose and fan heaters.

Cheers

Ian

 

[JohnRoberts]

ESD is real and it matters. Factory workers wear anti static (grounded conductive straps) to prevent static build up. These anti-stat straps generally have a high resistance in series so they don't expose workers to a dangerous shock current return path.

Modern ICs a a lot more robust that they were decades ago with protection circuitry built into every I/O pin.

I used to ship BBDs for my kit company plugged into conductive foam.

JR

 

[OneRoomStudio]

I absolutely agree with all that ESD is real and matters - I have always used precautions with ICs. My question is really about discrete semiconductors. Like pucho, I had been told that ESD doesn’t really affect discrete components much. Maybe it’s because I don’t work with many mosfets, and deal primarily with bjts, jfets, regulators, and diodes that I haven’t had issues.

I am intrigued by the notion of possible “unseen” damage in these components causing failure at a later time or shortening the lifespan in any way due to static. Has anyone experienced that or seen reference to it in modern jfets/bjts/diodes, etc?

 

[JohnRoberts]

I absolutely agree with all that ESD is real and matters - I have always used precautions with ICs. My question is really about discrete semiconductors. Like pucho, I had been told that ESD doesn’t really affect discrete components much. Maybe it’s because I don’t work with many mosfets, and deal primarily with bjts, jfets, regulators, and diodes that I haven’t had issues.

I am intrigued by the notion of possible “unseen” damage in these components causing failure at a later time or shortening the lifespan in any way due to static. Has anyone experienced that or seen reference to it in modern jfets/bjts/diodes, etc?

I recall back in the 70s mosfets were rare and notorious for being very easy to damage during handling.

Indeed bipolar devices are more robust but can still be harmed by static. I am not aware of any delayed failure modes, but there are subtle parameter shifts, so devices can still work but not completely as well as undamaged devices.

Work environment matters. Back in the early-mid 80s I had an office in an old brick factory building for some odd reason that building was terrible for static build up (perhaps our cheap office carpets). We also had an old shop vacuum cleaner that doubled as a van de graff generator. The static shocks I got after vacuuming my office carpet were painful. To reduce my pain from shocks, I connected a several Mohm resistor to the light switch plate ground screw and discharged myself through that resistor as needed. It worked

They sell anti-static sprays you can treat problematic carpets with. I never used one but ASSume that they help.

JR

 

[CalavoBob]

A pcb house I worked in discovered you can effectively lower static buildup in carpets by using a pump sprayer with a capful of Downey ( or Snuggle) fabric softener mixed in a gallon of water. We'd spray the carpets every 3 weeks.

 

[Admiral-dk]

I had an one year Internship at "JTAS Laboratoriet" in my youth 1979-1980 and was told about problems with ESD and components.
It started when the big long distance telephone exchanges became digital (they kind of where before that too), using First-Gen 4000 CMOS IC's.
It was observed that the first versions would die within 24 hours of a Module (Dual Europe size Cards - 6U Rack high) being exchanged from unboxing a brand new and plugging it in
So the Lab got a big amount of those Dead Modules back from the Technicians, located the dead IC's - opened those and put them under a Microscope. Here they could see some Very big Craters from the 'Punch-Through of the Electro-Static-Discharge'
This made Ericsson change the packing to including an Anti-Static bag + the Technicians where instructed in how to handle the Modules correct, when changing them .... => big improvement ...!
.... BUT not a total solution - now it took about a month after being replaced before the new Module had died
So the Lab did the same analyze on those -> Now the Craters where smaler than the with of the 'Tracks on the Silicon Substrate' -> the IC worked until the remainder of said Track burned over from the current running through it ....
This inspired some of the big IC manufactors to start adding Protection Diodes etc. to the I/O of the IC's in the 4000A and 4000B series - further reducing the problem .... But you still have to take care.

I was told that the Strap works nicely - but the problem can be solved without using it.
1) Always touch the 'GND' of the equipmet / PCB you're about to handle, before handling the PCB / IC / MosFet.
2) If you should give a Device (IC/Discrete) to another person - Do NOT put it into their hand !!!! - but drop it into their hand @ a distance about 200mm / 8" .... Why ? - well the device can't contain any charge that can dammage it + the distance between the persons are too big for a Zap between them.
There are a few more - but I have never worn a strap and had not had any IC's gone bad (from ESD).
Discrete MosFet Transistors - I had some, until we used 2N7002 - then many (probably 10-20%) => we changed to using the 2N7002CK that has internal Zener Protection of the Gate => problem solved.

Sorry for the rant ....

Per

 

[Newmarket]

i have been told that discrete circuits are esd safe and chip circuits are not.
I never was in a position to test that and I could never find a definite answer from anyone I asked. The web isn’t much better in finding that info, any searching can yield unique results. As for me I try to use best practice with an esd strap as much as possible.

Not 100% sure what is meant by discrete Vs chip circuits. But it seems to make little sense. In a crude sense the smaller the semiconductor junction area the more susceptible to ESD.

 

[Newmarket]

A pcb house I worked in discovered you can effectively lower static buildup in carpets by using a pump sprayer with a capful of Downey ( or Snuggle) fabric softener mixed in a gallon of water. We'd spray the carpets every 3 weeks.

Yes. Fabric softener has an antistatic effect. Similar for tumble dryer sheets. Can use them to reduce static noise when playing elec bass / guitar with unscreened plastic scratch plates and screws.

 

[C H]

You spray your guitarists with fabric softeners to discharge the static? And there was me thinkng it was because we were a bit smelly...

 

[pucho812]

Not 100% sure what is meant by discrete Vs chip circuits. But it seems to make little sense. In a crude sense the smaller the semiconductor junction area the more susceptible to ESD.

Take a mic pre. Mic pre with a 5532 or some other ic is a chip circuit. Take a mic pre made up of transistors and no ic chips that would be a discrete circuit.

 

[Newmarket]

Well that's just a bonus
But we can just program the guitar parts now and let the real musicians play live

 

[musipol]

Somewhat off topic, but I remember back in the day, I kept blowing up discrete FETs using an ungrounded soldering iron. The problem was ~60V AC on the tip and having the circuit gnd connected while soldering. I also recall blowing up 4000 series CMOS until I got a gnd strap. Even though those chips had protection networks, I managed to destroy them(with the help of carpeting).

 

[JohnRoberts]

Some outlets with floating EGC (like my house) can measure 60VAC (roughly one half mains voltage) on the safety ground but it is generally high impedance and low current. That said it might still be enough energy to damage sensitive components.

An old and not always safe strategy is to grab a bootleg ground. Neutral in typical domestic home wiring is bonded to ground at the utility panel/power drop, so nominally 0V. There are a couple dangers from "bootleg grounds".

#1 if that neutral wire feeding the outlet fails open circuit, the neutral current will have no return path and energize that safety ground including product chassis (not very safe).

#2 a far more dangerous problem is RPBG (reverse polarity bootleg ground). If that bootleg safety ground grabs the live line lead instead of 0V neutral, the chassis and safety ground are now energized. Outlets mis-wired RPBG can and have killed people.

[TMI about outlet wiring... OD1 /TMI]

JR