Author Topic: Episode suggestion: ESD is a myth  (Read 20563 times)

0 Members and 1 Guest are viewing this topic.

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Episode suggestion: ESD is a myth
« Reply #50 on: November 29, 2017, 02:39:48 pm »
If you have broad experience of electronics assembly around the world, in different climates, it would most interesting to hear your experiences of what those environments are like.

Many years I used to work for a company that have own fully equipped SMT assembly floor +/- same size as in EEVblog video and briefly visited another, much bigger one. From environmental control point of view they are all the same: controlled temperature 24/7 all round the year both in assembly and components storage places. No freaking open windows for god's sake. Big factory even have controlled humidity so it does not float around when rain comes.

To put it in Laymans Terms: when your pick&place machine alone costs 10x more than ventilation/condition system for whole factory floor then it is insane stupidity to invest in manufacturing equipment but save on climate control.

In case you wonder why it is so important to have control over temperature & humidity in electronics manufacturing (SMT assembly) floor - it's very important ;) If failure rate of boards that comes out of SMT assembly suddenly becomes high, then QA experts of customer will visit factory. If they do not see that factory follows environmental or ESD safety standards (yes, factory *must* have temp&humidity monitors/loggers all over the place), then lawyers of customer will tear it apart.
« Last Edit: November 29, 2017, 04:50:40 pm by ogden »
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8605
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #51 on: December 01, 2017, 03:03:09 pm »
If you have broad experience of electronics assembly around the world, in different climates, it would most interesting to hear your experiences of what those environments are like.

Many years I used to work for a company that have own fully equipped SMT assembly floor +/- same size as in EEVblog video and briefly visited another, much bigger one. From environmental control point of view they are all the same: controlled temperature 24/7 all round the year both in assembly and components storage places. No freaking open windows for god's sake. Big factory even have controlled humidity so it does not float around when rain comes.

To put it in Laymans Terms: when your pick&place machine alone costs 10x more than ventilation/condition system for whole factory floor then it is insane stupidity to invest in manufacturing equipment but save on climate control.

In case you wonder why it is so important to have control over temperature & humidity in electronics manufacturing (SMT assembly) floor - it's very important ;) If failure rate of boards that comes out of SMT assembly suddenly becomes high, then QA experts of customer will visit factory. If they do not see that factory follows environmental or ESD safety standards (yes, factory *must* have temp&humidity monitors/loggers all over the place), then lawyers of customer will tear it apart.
You have narrowed "electronics assembly" to a single topic - stuffing PCBs. This is a small part of the overall assembly space in an integrated assembly plant, and these days it is often subcontracted, so it doesn't occupy the final assembly plant at all. What about all the other assembly work? Putting displays, multiple PCBs, connectors and other modules into a final product takes lots of space, and tightly controlling the atmosphere would be a waste of time for most of it, unless the local climate is quite troublesome.

In the assembly plants I have been in, in the UK and Asia, PCB stuffing takes a much smaller space than, say, the final burn in area. Everything beyond goods inward, and up to the case being screwed/glued/welded shut requires a proper static control regime. Burn in is typically something like 48 hours running, followed by probing to ensure everything is functioning to spec. That means burn in happens before the case is closed, to facilitate the post soak probing. You need to manage static all the way through. If you make 10k or 20k TVs a day, that takes a lot of space. The only environmental control worth bothering about during soak testing, for most products, is to avoid a condensing atmosphere and excessive dust buildup.

Environmental control is seldom an all or nothing issue. Most factories control specific areas in specific ways. For example, if there is any paint spraying, that requires a moderate quality clean room to avoid particulates spoiling the finish; it requires efficient capture of waste paint; and if there is the risk of a condensing atmosphere it probably requires humidity control. You don't apply that kind of practice to the whole plant, though. By only applying it in specific areas, the energy costs are kept under control.

The requirements for PCB assembly depend a lot on the climate. In a place like Singapore, where the humidity hovers around condensing for much of the year, humidity control is important throughout an electronics assembly plant. The places where really tight environmental control is needed during PCB assembly, it is ensured by the equipment itself. Most SMD reflow and vapour phase soldering systems are completely enclosed, with a well controlled atmosphere inside. Old through hole wave soldering lines were far less controlled. The use of a dry nitrogen atmosphere within an SMD soldering system is pretty common practice now. Its not expensive to continuously separate nitrogen from the air these days, as long as your requirement is not for perfect purity. That dry inert atmosphere, combined with preheating, helps dry out the components before they reach the high heat of the soldering process. Eliminating most of the oxygen eases the demands on the flux, and results in a less tarnished final product.
« Last Edit: December 01, 2017, 03:07:57 pm by coppice »
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #52 on: December 01, 2017, 03:37:01 pm »
You have narrowed "electronics assembly" to a single topic - stuffing PCBs.

Moreover, the topic of ESD control seems to have been narrowed to assembly. There's also service, repair, calibration and site service to consider. Those four happen in all sorts of environments, from immaculate cal labs to dirty/hot/cold/wet/dry factory floors.

Even managed operating environments have ESD concerns - I used to have to manage rack upon rack of computers, network equipment and telephone switches in an environment where connections were being made and broken on a daily basis, boards swapped, whole chassis swapped. Plenty of room there for ESD incidents. I never had any confirmed ESD kills, but did have quite a few suspicious deaths.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5173
  • Country: us
Re: Episode suggestion: ESD is a myth
« Reply #53 on: December 01, 2017, 11:48:21 pm »
You have narrowed "electronics assembly" to a single topic - stuffing PCBs.

Moreover, the topic of ESD control seems to have been narrowed to assembly. There's also service, repair, calibration and site service to consider. Those four happen in all sorts of environments, from immaculate cal labs to dirty/hot/cold/wet/dry factory floors.

Even managed operating environments have ESD concerns - I used to have to manage rack upon rack of computers, network equipment and telephone switches in an environment where connections were being made and broken on a daily basis, boards swapped, whole chassis swapped. Plenty of room there for ESD incidents. I never had any confirmed ESD kills, but did have quite a few suspicious deaths.

ESD susceptibility varies widely with assembly level.  In a single component there is very little capacitance to absorb any ESD, and often only a single trace and/or component in the ground path.

At the circuit card level there is far more capacitance, and often multiple paths to ground, series resistance and other mitigating factors.  Often these features are deliberately designed in to make the boards more robust.

These factors continue at each level of assembly up to the finished and delivered product.  Any manufacturer who has a warranty or reliability requirements has done the work to make sure that normal ESD pulses applied to normally accessible parts of the equipment are non damaging.

How much problem you have personally had with ESD will vary with your environment and with the type of work you do.  Board swapping entails relatively low risk because it is hard to touch most boards to remove them without touching a grounding element of the chassis, and because the boards inherently are less susceptible to damage.  A new board being pulled from packaging has more risk of being exposed than one being pulled from a chassis, but still benefits from the generally lower susceptibility of board level products.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6927
  • Country: ca
Re: Episode suggestion: ESD is a myth
« Reply #54 on: December 01, 2017, 11:59:51 pm »
I have GaN LED's in series with my workbench ESD mat ground, and it tells me everything, they light up with any ESD or AC hum. Great for your workbench, I've caught bad grounds on wiring and equipment with the LED's.
Humidity gets very low during Canadian winter and ESD is pretty harsh. I can make 30kV getting up from my office chair, and that will kill a semi.

For many years, ESD was scoffed at in the service and repair industry. I knew to "touch ground" and discharge static before swapping a board on a computer, as nobody would spend money on wrist straps and bags. Other techs had ESD-related callbacks and problems, the VN10K is the most sensitive-to-static component ever made. They die about a week or two after being subjected to ESD.
 

Offline Mr. Scram

  • Super Contributor
  • ***
  • Posts: 9810
  • Country: 00
  • Display aficionado
Re: Episode suggestion: ESD is a myth
« Reply #55 on: December 03, 2017, 10:02:25 am »
I have GaN LED's in series with my workbench ESD mat ground, and it tells me everything, they light up with any ESD or AC hum. Great for your workbench, I've caught bad grounds on wiring and equipment with the LED's.
Humidity gets very low during Canadian winter and ESD is pretty harsh. I can make 30kV getting up from my office chair, and that will kill a semi.

For many years, ESD was scoffed at in the service and repair industry. I knew to "touch ground" and discharge static before swapping a board on a computer, as nobody would spend money on wrist straps and bags. Other techs had ESD-related callbacks and problems, the VN10K is the most sensitive-to-static component ever made. They die about a week or two after being subjected to ESD.
Can you tell more about those LEDs and how they work?
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Episode suggestion: ESD is a myth
« Reply #56 on: December 03, 2017, 11:20:38 am »
You have narrowed "electronics assembly" to a single topic - stuffing PCBs.

Indeed. This is part where temperature and humidity shall be controlled. I indicated that I talk about SMT assembly and you were ok with that. Suddenly you are not? :)

Quote
What about all the other assembly work?

When components are on the PCB, then temperature & humidity is not an issue anymore indeed. Workers need comfortable temperatures instead :)

Quote
The only environmental control worth bothering about during soak testing, for most products, is to avoid a condensing atmosphere and excessive dust buildup.

Burn in is not assembly, so it does not count here in this context anyway.

Quote
Most SMD reflow and vapour phase soldering systems are completely enclosed, with a well controlled atmosphere inside.

:palm:  Nitrogen in the oven just reduces oxidation. It's not humidity control.

Avoidance of popcorn effect is why you need humidity control in the component storage and basically in whole SMT assembly floor: https://en.wikipedia.org/wiki/Moisture_sensitivity_level

When temperature outside oven changes by more than 5 degrees oC - you must change temperature profile (preheat) of the oven. Viscosity of solder paste also changes due to temperature. Better just read some specs of solder paste and component temperature profiles to see that it is indeed insane stupidity to not control temperature around board assembly :)
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8605
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #57 on: December 03, 2017, 12:06:54 pm »
You have narrowed "electronics assembly" to a single topic - stuffing PCBs.
Indeed. This is part where temperature and humidity shall be controlled. I indicated that I talk about SMT assembly and you were ok with that. Suddenly you are not? :)
Quote
What about all the other assembly work?
When components are on the PCB, then temperature & humidity is not an issue anymore indeed. Workers need comfortable temperatures instead :)
When the components are on the PCB the environmental requirements are little different than any other point in assembly, apart from the inside of the reflow or vapour phase system. You need to control humidity, if condensation is a potential issue for you, and you need to control static.
Quote
The only environmental control worth bothering about during soak testing, for most products, is to avoid a condensing atmosphere and excessive dust buildup.
Burn in is not assembly, so it does not count here in this context anyway.
How is burn in not relevant, when it requires the same environmental conditions as the rest of assembly - a non condensing atmosphere and good static control? Do you exclude test and other key functions of the typical assembly shop from your thinking? If so, why?
Quote
Most SMD reflow and vapour phase soldering systems are completely enclosed, with a well controlled atmosphere inside.
:palm:  Nitrogen in the oven just reduces oxidation. It's not humidity control.
Now you are just being a jackass. I said DRY nitrogen is used. Did you miss the word DRY on purpose?
Avoidance of popcorn effect is why you need humidity control in the component storage and basically in whole SMT assembly floor: https://en.wikipedia.org/wiki/Moisture_sensitivity_level
It would be a very poorly run line which keeps components outside their well controlled long term storage conditions (sealed packaging, or low humidity storage area) for long periods, where they could absorb enough moisture deep into their surfaces to be a problem. Popcorning is largely a problem of poor long term storage, and avoiding it doesn't require a tightly controlled atmosphere. It just requires a rather dry one. The key thing in the assembly line is to ensure a non-condensing atmosphere at all points, with no downward temperature shocks leading to sudden condensing.
When temperature outside oven changes by more than 5 degrees oC - you must change temperature profile (preheat) of the oven. Viscosity of solder paste also changes due to temperature. Better just read some specs of solder paste and component temperature profiles to see that it is indeed insane stupidity to not control temperature around board assembly :)
What kind of horrible reflow systems have you used? They can't cope with a stream of boards intermixed at different incoming temperatures, but they cope with a slowly changing incoming temperature just fine. Part of their specifications is te incoming temperature range they can adapt to, and how quickly it can change. Their thermostatic control ensures the temperature of the boards are stabilised, and the dry atmosphere ensures a rapid dispersion of any residual moisture.

The solder paste viscosity is usually only critical during the actual reflow, where the reflow machine tightly controls the temperature, and therefore the viscosity. I have never seen it affect the PnP systems. Low viscosity between the PnP and reflow machines used to often be a cause of small parts being shaken out of place. That lead to a huge rise in the use of glue spotting. People tend to just manage their lines better these days, and reduce shaking.
 

Offline AF6LJ

  • Supporter
  • ****
  • Posts: 2902
  • Country: us
Re: Episode suggestion: ESD is a myth
« Reply #58 on: December 03, 2017, 01:07:26 pm »
I have visited this thread from time to time and I find myself wondering when the shape of the Earth will become a topic of debate..........
Oh wait;
Th3ere is already a thread on that.
 :palm: :palm:
Sue AF6LJ
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6927
  • Country: ca
Re: Episode suggestion: ESD is a myth
« Reply #59 on: December 04, 2017, 04:56:11 am »
I have GaN LED's in series with my workbench ESD mat ground, and it tells me everything, they light up with any ESD or AC hum. Great for your workbench, I've caught bad grounds on wiring and equipment with the LED's.
Humidity gets very low during Canadian winter and ESD is pretty harsh. I can make 30kV getting up from my office chair, and that will kill a semi.

For many years, ESD was scoffed at in the service and repair industry. I knew to "touch ground" and discharge static before swapping a board on a computer, as nobody would spend money on wrist straps and bags. Other techs had ESD-related callbacks and problems, the VN10K is the most sensitive-to-static component ever made. They die about a week or two after being subjected to ESD.
Can you tell more about those LEDs and how they work?
My silly LED's light up with even a few pF stray (mat) capacitance from AC hum, and showed me two bad ESD grounds on my bench. GaN green LEDS are super sensitive, even 1uA you can see.

I had a worn out duplex receptacle that made no connection to GND pin, and an open-circuit snap-clip on the ESD mat.

I tried adding a protective diode-clamp across the LED's but leakage just made them not work.

Try build it. Totally worth it to "see" ESD.
« Last Edit: December 04, 2017, 05:11:30 am by floobydust »
 
The following users thanked this post: Mr. Scram

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #60 on: December 04, 2017, 01:41:26 pm »
I tried adding a protective diode-clamp across the LED's but leakage just made them not work.

Try using just the base-collector junctions of small signal BJTs as diodes, or a couple of diode wired low leakage JFETs such as 2N4117s.

You did stack the diodes so that the Vf added up to more than the Vf of your LEDs? Obvious I know, but we've all missed something stupid like that at some point.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6927
  • Country: ca
Re: Episode suggestion: ESD is a myth
« Reply #61 on: December 06, 2017, 05:50:51 am »
This is what I tried. It clamped hard about 4.3V leaving maybe 40mA for the LED's.
The problem is the clamp starts conducting enough to hog LED current, well below the actual clamping point.
So at 2.5V, it was like 120uA.
The diodes are a poor choice, but I wanted something strong enough to take a high current hit.
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #62 on: December 06, 2017, 03:20:46 pm »
I've just tried a quick simulation with 5 stacked 2N4117 JFETs, diode wired, and they only hog about 3uA at 2.5V BUT they are only rated for a forward gate current of 50mA so just plain aren't up to the job. I don't think you'll find a PN junction with the right combination of characteristics for this job. If you can find a gas discharge tube with a low enough voltage rating you might be onto a winner with that.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6927
  • Country: ca
Re: Episode suggestion: ESD is a myth
« Reply #63 on: December 08, 2017, 01:58:34 am »
I think a neon lamp or GDT may work as a clamp for the LED's.  The ESD pulse rise-time is so fast.
Littelfuse CG75 gas discharge tubes rated ~75VDC ramp impulse 100V/sec. but rated 400-650V surge impulse breakdown, which is high.

The LED's could take the impulse, say 100mA with a series resistor, or another resistor to limit the whole ESD discharge current?

 

Online wraper

  • Supporter
  • ****
  • Posts: 16797
  • Country: lv
Re: Episode suggestion: ESD is a myth
« Reply #64 on: December 08, 2017, 12:14:51 pm »
IMO if there is already 1 meg resistor in series, you could try connecting small cap in parallel which will smooth the pulse. And likely there is no ESD protection needed with 1 meg resistor at all. If there is say 50kV discharge, current will be limited to 50mA (unless there is flashover in resistor).
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #65 on: December 08, 2017, 03:20:16 pm »
IMO if there is already 1 meg resistor in series, you could try connecting small cap in parallel which will smooth the pulse. And likely there is no ESD protection needed with 1 meg resistor at all. If there is say 50kV discharge, current will be limited to 50mA (unless there is flashover in resistor).

Dielectric breakdown in dry air is ~3MV/m so the distance for a flashover at 50kV would be about 17mm.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6927
  • Country: ca
Re: Episode suggestion: ESD is a myth
« Reply #66 on: December 08, 2017, 06:52:12 pm »
This is really a way to "see" ESD and I need it to be reliable. I can't tell you how many companies I've seen with old trashed bench wiring and nothing grounded properly, or non-ESD floor wax, or worn out wrist straps, yet they are working with delicate electronics. They think ESD prevention is a nuisance, something very informal.

One company had repeated product MOSFET failures about 2-4 weeks after PCB stuffing. The MOSFET controlled a gas-solenoid and shorted, leaving fuel on.  Of course explosions resulted  :palm: The MOSFET circuit was reviewed, the failed MOSFETS sent back to the semi manufacturer for electron microscopy, and all failures were said due to ESD damage of the gate junction. The company's manufacturing plant had loosey goosey ESD policies, pretty much nothing. It was a debacle. ESD is not a myth, it's just invisible.


These ESD mat/wrist-strap resistors are vanilla 1/4W through-hole parts, so many kV just jumps across them. They are to protect against mains electrocution.
Without the 1MEG resistor, I'm going to try some high level ESD hits and see how it does. Something that can take beyond 61000 Level 4 machine hits
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8605
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #67 on: December 08, 2017, 06:58:08 pm »
This is really a way to "see" ESD and I need it to be reliable.
Well you could just buy some of the little alarm boxes made for this job. I've never bought them myself, so I don't know their price, but they aren't very complex. They are used in most labs, so I assume the production volume is high enough for the price to be low.
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Episode suggestion: ESD is a myth
« Reply #68 on: December 08, 2017, 07:21:34 pm »
These ESD mat/wrist-strap resistors are vanilla 1/4W through-hole parts, so many kV just jumps across them.

Just for the benefit of those not in the know:

The construction of most standard small metal or carbon film resistors involves a spiral groove cut into the film. This means that the breakdown voltage of these parts is mostly dictated by the size of the gap in these spirals, not the overall size of the part. That's why when dealing with high voltages old fashioned carbon composite are frequently specified as in those the element that would have to undergo breakdown is near to the full length of the part. The maximum withstand voltage on a typical 1/4 carbon film resistor is 500V, the withstand voltage on a 1/4W carbon composite resistor is 6kV.

If you want the very best isolation in the face of high voltages use resistors with specified withstand voltages and use them in strings  constructed so that they are never individually subjected to voltages above their rated withstand voltages (e.g. 3 x 333k in series in place of 1M, and so on).
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf