Cases and ESD concerns

Agreed, there are good reasons to use metallic cases. I'll do manual heat conduction if I managed to get one.

@Mike,

There is no inherent problem with a plastic case - look around you and count up how many small electronic gizmos are packaged in metal and how many in plastic. Plastic cases do not spontaneously  create static charges on their insides and do give the parts inside considerable protection. When you do an 8kv contact static discharge test using  a 'spark gun' quite thin plastic boxes (1mm thick) will prevent a spark breaking though. A plastic box is fine to protect an RPi like device from static but does nothing about emissions or susceptibility of/to EM radiation.

In the case of an RPi (which has been tested for CE marking) susceptibility is not a big issue (it's not intended for use in safety critical applications) and it's low power and certification mean you can be relaxed about emissions.

(The dielectric strength of ABS is about 18kV/mm (http://www.machinist-materials.com/comparison_table_for_plastics.htm))

Michael Kellett

"which has been tested for CE marking"

for clarification it should be noted that the RF testing that was done

was not successful with regard to residential standards.

"it's not intended for use in safety critical applications"

[citation required]

Hello Coder27 -

what's all this "citation required" stuff - are you going all formal on us 

Obviously the RPi is not suitable for safety critical applications - it's normal to assume that a part is NOT suitable unless it can be proved to BE suitable and even at their most enthusiastic I haven't heard the RPi Foundation make any such claims.

As for the CE testing - all EMC testing is a bit hit and miss (small changes in cables can make big differences) and the main difference between the industrial and residentail standards is that the former requires less susceptibility but allows more emisssion, in the context of boxes I think my comments still stand -  and without actual testing it would be hard to be sure if any specific arrangement of cables and box had made it (the RPi) better or worse but I don't think it's likely that anyone will notice.

Michael Kellett

Clearly like everone said above the RPI is not intended in anyway to be suitable for Safety Critical Applications.  But as Coder27 said you dont see anything on the market not using some form of plastic case.  If manufacturers are using these they have done some form of testing that cleared plastics cases for use.  just my opinion.

@Michael

I've been thinking the same thing for years, as far as consumer products with plastic cases. I've often wondered if they have metallic shielding inside; I'm guessing most do not.  It kind of baffles me, considering the yearly training I have to take, preaching about proper shielding and all that.

And yes, I realize that plastic won't spontaneously create static; but static is created by two materials rubbing together, or coming in contact and then separating.  And most plastic, I believe, is very prone to that phenomenon.  So a plastic case moving around various surfaces, my hand rubbing on it, or even the RPI itself rubbing on the inside of the case, can create an electric field which can charge up nearby objects.  What you say about plastic not allowing a charge to penetrate it may very well be true; but can it prevent an electric field outside the case from charging up the RPI inside the case?  If not, then when you eventually open the case, the charge built up on the RPI will discharge into me, even if I'm wearing a properly grounded wrist strap, which would cause as much damage as me zapping it with my finger after shuffling my feet on the carpet.

Perhaps the bottom line is that if loads of hobbyists are using plastic cases without much problem, I just need to not worry about it.  I suppose I can take comfort in the fact, if I destroy it or reduce its useful life, it's only $35!

Nate,

you wrote:

"Clearly like everone said above the RPI is not intended in anyway to be suitable for Safety Critical Applications."

What's clear to you may not be clear to everyone, so I'd be

interested to hear what specifically you think the main suitability

issues are.  Such as, do you have a problem with Linux?

or with firmware verification? or some hardware issue?

polyfuses susceptible to overheating?  oscillators that

don't always oscillate?  or whatever?

There are lots of people who have discussed using RPis

for space applications, real-time industrial process control

applications, health-care monitoring and display applications, etc.,

and as far as I know, such applications have been encouraged

rather than discouraged by the foundation.

@Coder27

A Safety Critical system is one where failure or malfunction may endanger human health or life. An application in space, industrial process control or health care monitoring may or may not be safety critical.

The issue with using the RPi in a Safety Critical applicaton is that (AFAIK)  neither the hardware or software design was undertaken with any particular safety critical application in mind and so has not been produced in accordance with any generally accepted standard. This does not mean that it is not safe - but that it can't be proven to be safe and therefor would not be used without additional work (which might be staggeringly expensive (eg re-writing ALL the code from the ground up)).

Try http://en.wikipedia.org/wiki/Safety_Integrity_Level

for a starter into the minefield of Safety Critical systems - it's far too big  a subject to cover in any detail here.

Michael Kellett

Well this has gotten way off topic from my original question about ESD.  Anybody have any more thoughts on that subject?

My thoughts on the subject (as an electrical engineer by original training) is that every device has the potential to be affected by ESD, and by EMC, and by acoustic and thermal sensitivity, and many other environmental factors, and every device also has the potential to be the cause of harm in all of these categories.  The engineer is a science-based practitioner who understands all of these issues, is able to measure them, and can produce a device with sufficient immunity and low enough emissions to operate successfully in its expected environment.

Everything is a compromise.  (Engineering is often described as an Art of Compromise.)  The Foundation produced a device with a certain amount of immunity and emission but traded off against cost, and there is no doubt whatsoever that more immunity and less emission in all categories could be obtained through casing of the board.  Note that many other factors come into this equation, and it's not a joke to consider "paperclip immunity" etc as well.

I strongly recommend to anyone who cares that they case their Pi.  As long as they know what they are doing, their Pi will end up with a better immunity/emission specification than without it.  Note that this will not be easy though, because the Pi operates at far too high a thermal point to encapsulate without giving thought to heat dissipation.  It's not much use being totally immune to everything and producing zero emissions if it results in board meltdown.

Morgaine.