Author Topic: Simplified guidelines for device certification?  (Read 1447 times)

0 Members and 1 Guest are viewing this topic.

Offline nardevTopic starter

  • Supporter
  • ****
  • Posts: 411
  • Country: ba
Simplified guidelines for device certification?
« on: August 18, 2019, 01:43:40 pm »
Hello guys.

I believe there is a lot of documents and aspects regarding a device certification.

However, I just want to get some idea about most common problems and safest ways to avoid them for device design.

So let's imagine a device/pcb with buck converter, bunch of relays for a common, small power/current/voltage 2Kw, microcontroller, touch keypad on same pcb, maybe some shift registars, common LEDs and of course bunch of a connectors.
P.s. I would not try to add antenna and other radio like segments to this example as I believe it would make things way too complex for a simplified talk.

What are most common issues?
Isolations of used and unused GPIO pins?
Space between relays, lines, parts?

I'm aware that somebody could talk days on this generalized question but as I said, "what is the most common" ?

Thnx
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Simplified guidelines for device certification?
« Reply #1 on: August 18, 2019, 02:30:42 pm »
The keyword is filtering and overvoltage protection: every incoming and outgoing signal should have some form of filtering and overvoltage protection. A simple series resistor of 1k Ohm may already do the job for GPIO pins. Also keep high current/high frequency loops small and keep them away from other circuitry. Nowadays I have a common mode choke at the input and output of any isolated DC-DC converter. This ensures that the DC-DC converter can't push noise into system ground.

But there are many ways to mess things up; there isn't a generic recipe.
« Last Edit: August 18, 2019, 02:33:50 pm by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: nardev

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #2 on: August 18, 2019, 03:04:37 pm »
So what are you trying to certify for? I can give you my best practice but it will make your products expensive if they are comercial because i have to pass military.
 

Offline Mr. Scram

  • Super Contributor
  • ***
  • Posts: 9821
  • Country: 00
  • Display aficionado
Re: Simplified guidelines for device certification?
« Reply #3 on: August 18, 2019, 04:47:34 pm »
So what are you trying to certify for? I can give you my best practice but it will make your products expensive if they are comercial because i have to pass military.
I'd be interested to hear about military certifications. Both military and space applications seem to pose interesting challenges.
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #4 on: August 18, 2019, 05:10:32 pm »
So what are you trying to certify for? I can give you my best practice but it will make your products expensive if they are comercial because i have to pass military.
I'd be interested to hear about military certifications. Both military and space applications seem to pose interesting challenges.

I don't know if there are any specific certificates. But there are standards that will be written into specifications and contracts. The theory is that each contractor tests their devices and then the whole vehicle is tested with the expectation that it should pass as all the components have.

However the standards are a lot more stringent than automotive so if I design something "that has to pass EMC" it may be over designed for something that wants to pass auto (reg 10 in the EU) or comercial testing.
 

Offline bjbb

  • Contributor
  • Posts: 44
  • Country: us
Re: Simplified guidelines for device certification?
« Reply #5 on: August 19, 2019, 06:14:17 pm »
This venue has not, in my experience, provided correct and usable replies to compliance and regulatory questions. There was a recent article on Hackaday, while not wholly incompetent, was incomplete. But may be a place for the layperson to start. If you are a professional designer working alone doing a project for public consumption, now is the time to seek a Notified Body, NRTL, SCC etc per whatever third-party accreditation system that has been set up by your government.

In general, a start place would be to look at your country's harmonized standards based on CISPR32 and 35, and IEC62368-1, as a base-line reference for EMC and product safety requirements, respectively.

But regulatory stuff cannot be done by the untrained and inexperienced hobbyist - interpreting standards and providing design assessments are typically done by experienced engineers that inhabit the bizarre intersection of physics, chemistry, and statutory law.
 
The following users thanked this post: nardev

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #6 on: August 19, 2019, 08:25:45 pm »

But regulatory stuff cannot be done by the untrained and inexperienced hobbyist - interpreting standards and providing design assessments are typically done by experienced engineers that inhabit the bizarre intersection of physics, chemistry, and statutory law.

I would not be so downbeat. If you are not dealing in high powers passing EMC can be a doddle. No one looks at a standard, scratches their chin and grey beard and decides on how to design the circuit.

Over time you learn techniques that you know are effective for certain standard tests. Understanding the basics of filtering and how the noise propagates is a good start. one can never design to pass a certain test. You simply weigh up the cost/benefit of what you can use and apply as much as you think will get you through the test.
 

Offline nardevTopic starter

  • Supporter
  • ****
  • Posts: 411
  • Country: ba
Re: Simplified guidelines for device certification?
« Reply #7 on: August 19, 2019, 09:59:25 pm »
So what are you trying to certify for? I can give you my best practice but it will make your products expensive if they are comercial because i have to pass military.

That is a good question, but intentionally i wanted to avoid saying that. I guess, tome things are equally important for
 fcc and some other asian certifications. Also for CE sign.
 

Offline Mr. Scram

  • Super Contributor
  • ***
  • Posts: 9821
  • Country: 00
  • Display aficionado
Re: Simplified guidelines for device certification?
« Reply #8 on: August 19, 2019, 10:13:20 pm »
I don't know if there are any specific certificates. But there are standards that will be written into specifications and contracts. The theory is that each contractor tests their devices and then the whole vehicle is tested with the expectation that it should pass as all the components have.

However the standards are a lot more stringent than automotive so if I design something "that has to pass EMC" it may be over designed for something that wants to pass auto (reg 10 in the EU) or comercial testing.
Can you give examples of how military design might differ? The sector fascinates me but it's obviously not the most transparent sector.
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #9 on: August 20, 2019, 07:07:21 am »
Can you give examples of how military design might differ? The sector fascinates me but it's obviously not the most transparent sector.

More filters, oh, and more filters. You just go to every extreme that you reasonably can. We had a device fail just from the clock output of a µC that drove another chip although it was a plastic enclosure, so understand that it's the edge rate that couses the trouble. So you have to start using screened metal enclosures which can become expensive. At the moment my employer is in love with anodising stuff as the finish it great and very durable. But we use aluminium so you have to have a full faraday cage, so I have to work out ways of getting things to be electrically connected like re milling the anodise to get some clean metal again. Everything much be electrically connected, so say you have metal connectors, these need to be electricaly connected to the enclosure or that is still a hole in the side and that connector body poking through is now an aerial because any two parts that are not tightly bonded can have two different potentials which = An aerial.

You have to understand the fact that you have radiated and conducted emissions and both need catering for. Radiated is minimised with screened/faraday cage enclesures and 4 layer PCB's so that you have a ground layer in close proximity with the signal layer. Dave demonstrated this in one of his videos on that build it yourself computer where he added a GND and VCC planes to a 2 layer PCB and the radiated emmisions went down by 15dB. Conducted emissions need to be addressed as anything you conduct out onto cables can then be radiated. So just suitable filtering galor.

Tim is the best EMC person on here. I've just been tinkering very unscientifically.
 
The following users thanked this post: Mr. Scram

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #10 on: August 20, 2019, 07:09:57 am »
I made sure I studied the attached module in my HNC course. Sadly I am told they are removing this module.
 
The following users thanked this post: nardev

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #11 on: August 20, 2019, 12:45:54 pm »
Of course there is always the very practical book by Wurth Elektronik: https://simonselectronics.co.uk/product/triology-of-magnetics-wurth-elektronik-book/
 
The following users thanked this post: nardev

Offline nardevTopic starter

  • Supporter
  • ****
  • Posts: 411
  • Country: ba
Re: Simplified guidelines for device certification?
« Reply #12 on: September 11, 2019, 11:12:33 am »
Of course there is always the very practical book by Wurth Elektronik: https://simonselectronics.co.uk/product/triology-of-magnetics-wurth-elektronik-book/

Man you are killing me. I need to go bact to Univ  |O |O |O
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Simplified guidelines for device certification?
« Reply #13 on: September 11, 2019, 11:36:55 am »
why do you need to go back to university? it's easier to learn what you need rather than what other people think you need.
 
The following users thanked this post: nardev

Offline nardevTopic starter

  • Supporter
  • ****
  • Posts: 411
  • Country: ba
Re: Simplified guidelines for device certification?
« Reply #14 on: September 11, 2019, 11:37:51 am »
why do you need to go back to university? it's easier to learn what you need rather than what other people think you need.

Hmm.. usually it was like that  :-+
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf