EEVblog Electronics Community Forum
General => General Technical Chat => Topic started by: MagicSmoker on November 04, 2019, 04:58:35 pm
-
This thread: https://www.eevblog.com/forum/projects/insulating-washers-creepageclearance-issue-at-holes/ (https://www.eevblog.com/forum/projects/insulating-washers-creepageclearance-issue-at-holes/)
got me to wondering if there are any good books on helping the new EE get through UL (and other) safety agency requirements w/r/t creepage, clearance, insulation coordination, etc?
I have good books on designing for EMC compliance - Tim Williams and Harold Ott are two notable authors whose books on this subject I can highly recommend - but nothing I can recommend on designing for safety compliance, and a quick search on Amazon gave disappointing results (try searching for "product safety" under the engineering books section and despair at the deluge of NEC and NFPA-focused texts).
-
- Buy applicable standard
- Do what it says
- Profit
-
- Buy applicable standard
- Do what it says
- Profit
By this logic if I were to get arrested all I would need to do is go to a law library and read the relevant statutes to defend myself, eh?
Sometimes a little bit of plain English explanation (or whatever your lingua franca might be) helps one interpret the often-inscrutable legalese of regulatory standards.
-
read standard until your eyes bleed
-
Sometimes a little bit of plain English explanation (or whatever your lingua franca might be) helps one interpret the often-inscrutable legalese of regulatory standards.
Standards change often enough and apply in such specific combinations to any given field that trying to condense them into any other format is a loosing proposition. Big corporations have the resources to have internal standards which do this, but they still go back to the specific standards to audit a design before its released. So even if you had such a book you would still need to refer to the standard.
-
- Buy applicable standard
- Do what it says
- Profit
By this logic if I were to get arrested all I would need to do is go to a law library and read the relevant statutes to defend myself, eh?
If you don't understand law, you hire a lawyer, if you don't understand standards you hire a consultant who does.
-
Sometimes a little bit of plain English explanation (or whatever your lingua franca might be) helps one interpret the often-inscrutable legalese of regulatory standards.
I agree with you. While the law or regulation or standard is the last word, commentary, explanation and discussion go a long way in helping with understanding and implementation. There are books expounding on pretty much any regulated field because, as you say, it is not enough to read the regulation and understand the words; it takes a lot of training and education to get the full context which may be contained in other parts of the written regulations or even just be something commonly accepted.
It is not enough to tell a student the law is V = I * R. It takes a lot of explanation until the student really grasps and understands what it means.
Even with laws there are lawyers who specialize in specific fields because the law is so broad and even being a lawyer it is not enough to just "read the law". Practice and experience are a very important part.
Yes, safety standards are the last word but there is a lot of knowledge on how to best comply, best practices, where maybe you can get away with going marginal and where you should absolutely give a wide margin of safety, how the field is divided, etc. In theory you could just read the regulation and deduce a lot yourself, in practice someone with lots of experience can guide you, point out the most important things, etc.
Armed with that knowledge you are much better equipped to understand and implement the standard or regulation.
Often standards and regulations are written in very complicated ways. It takes a while to know even which part applies to your problem. Then you can get into a maze of sections referencing each other where you lose sight of what they are trying to say. (Except as referenced in section 2.D devices classified as "electrical" according to this section shall not need to comply with this section if their intended use shall be mainly domestic and not industrial as defined in appendix 5.)
Someone with knowledge and experience can be very helpful. Yes, you can hire them to do the work but you can also learn through books, colleagues, forums, etc.
-
Standards change often enough and apply in such specific combinations to any given field that trying to condense them into any other format is a loosing proposition. Big corporations have the resources to have internal standards which do this, but they still go back to the specific standards to audit a design before its released. So even if you had such a book you would still need to refer to the standard.
I thought I made it quite clear in my opening post that I was asking for book recommendations similar to what exist for EMC except aimed at (or at least better covering) the safety aspects of product compliance testing. I wasn't asking for a book that could substitute for the standards. Which segues nicely into your other perplexing reply...
If you don't understand law, you hire a lawyer, if you don't understand standards you hire a consultant who does.
I recommend hiring a compliance expert all the time, but that is obviously not mutually exclusive with wanting a book that could also be recommended that helps choose which standard(s) might apply, explains the basic idea behind a standard in non-legalese, and gives examples of how one might comply with it, or fail it (the latter being particularly instructive... especially when learned the hard way).
After all, my local UL testing lab says 50% of electronics products fail safety testing the first go around, so it seems rather obvious to me - and for some reason not to you - that just owning the standards isn't sufficient. And BTW, I own 14 different UL and EN standards just to deal with power supplies, traction drives, EV chargers and railway devices, which perfectly illustrates the old cliche that, "standards are great: there are so many to choose from!"
-
I agree with you. While the law or regulation or standard is the last word, commentary, explanation and discussion go a long way in helping with understanding and implementation. There are books expounding on pretty much any regulated field because, as you say, it is not enough to read the regulation and understand the words; it takes a lot of training and education to get the full context which may be contained in other parts of the written regulations or even just be something commonly accepted.
Yes, thank you for understanding what I meant, despite being the only person in this thread with a non-native-English-speaking country flag...
I mean, I have at least a dozen books on power electronics which I refer back to all the time yet not a single one dedicates more than a few pages to safety compliance issues. In the thread I linked to above the OP asked a simple question that did not have a simple answer: how does a mica (or similar) insulator between a metal-tab TO-220 and an exposed heatsink meet creepage requires between mounting screw and heatsink? UL 840 just gives creepage and clearance distances for various peak and RMS voltages and "Pollution Degrees" but a drawing or brief explanation of how this is kosher (or not) would be invaluable.
That thread - and my current experience with a company developing a bidirectional EV charger for grid support (the first product undergoing testing for a new UL standard, in fact - 9741) - drove home how little information there is on safety compliance compared to lots of books being available for EMC compliance. Or maybe it's just that EMC is a more specific search term on Amazon and I'm not looking hard enough.
At any rate, hopefully your response will steer this discussion onto a more productive path and maybe a good book recommendation will actually be made...
-
I haven't looked for books on safety regulation standards, and perhaps like many here I've collected a few on EMC over the years. The process of making a design fit through the complex path of standards is unfortunately often learned the hard way, that is, the expensive way.
General questions and statements that unfortunately may not apply to your question. Or, as Dave says, "for those of you playing at home":
1. Are you in a small company, meaning that you are responsible for everything, a medium company, with a few experienced people, or with a large company, with some sort of compliance group? The key here is whether or not the couple management levels above you have any awareness of standards in general, the potential costs it takes to achieve compliance, and the potential risks of not achieving compliance. Unaware management will freak out at the costs involved. $100 for a standard? $1000 for a pre-scan? $10000 for EMC? $100,000 for HazLoc? How will they take it?
2. Which categories of standards apply? Are there contradictory in requirements? Are the solutions to the problems contradictory? In what sequence should the various approval stages be completed? How do these interact with the processes implemented (either purposefully or frenetically) at the company? Is the difference between a directive and a standard understood?
3. In what fashion does the project planning and estimation process take into account the standards and approvals process? How much time and money is allocated? What data was used to arrive at those estimates? Did the people making the estimates have experience with it?
4. Are the product requirements achievable within the applicable standards? Do the target marketing regions have different standards that apply?
5. How complete was the design used for getting a quotation on an approvals process? Is the design actually being reviewed for approval complete? Is your organization prepared to do iterations on the design to achieve compliance?
6. A few hours of consultant's time at perhaps $500/hour is worth it if they get a clear idea of what to review. Many approval agencies can do consulting work, but those are kept strictly separate from an actual evaluation. Expect 1/3 new and useful information, 1/3 validation of what you already know, and 1/3 going into rabbit holes.
7. A quotation from an approval agency will give an idea of the scale involved. It should have a cursory list of applicable standards at the scale of the product. Others may apply at the scale of installation. Numerous requirements for the deliverables should be listed, several of which may not have been accounted for in the original project estimate.
Good luck! On more than one occasion I've suggested to Dave that this forum really needs some type of Standards / Approvals / Agency / etc. category.
-
And don't forget that if it is a product to be sold internationally you will want it to comply with American, European, Chinese, Japanese and other standards. Just look at a brick PSU for a laptop and you can feel dizzy just looking at all the approval logos.
-
And don't forget that if it is a product to be sold internationally you will want it to comply with American, European, Chinese, Japanese and other standards. Just look at a brick PSU for a laptop and you can feel dizzy just looking at all the approval logos.
And if you go for Korea, be prepared to fill out a 27B/6.
-
Regulatory is a huge industry, it's big money and largely closed off to the engineering community for the knowledge.
The agencies have their own internal training programs, mentors, experts to assist their certifiers but as an outsider, expect to pay over $350/hr to acquire knowledge on product electrical safety standards.
Agency Project Managers, Certifiers are all on the treadmill too, to bring in the cash, meet targets. It doesn't matter if the project fails- they earn the money either way.
Employers have spent over $250,000 on approvals for projects with me as a lead engineer and this is how I had to learn them.
I don't think a book would be worth it for several reasons, and find there are none out there.
The liability can be huge. Consulting services to companies for regulatory, or writing a book- there is risk an error or something misunderstood leading to either another product cycle $ (PCB changes, enclosure changes, new prototypes, testing etc.) or leading to an unsafe product - can happen.
When you sign up with UL/CSA/Intertek/Sira/Baseefa for example, there is a contract huge waiver that they are not liable for anything. Even if the agency screwed up and unsafe product resulted, they will tell you "remove the sticker". No refund even on the flawed approval costs, and no liability for the injured parties. You get stuck doing a product recall.
So the responsibility of giving knowledge about safety might be the first reason no books are out there?
-
- Buy applicable standard
- Do what it says
- Profit
By this logic if I were to get arrested all I would need to do is go to a law library and read the relevant statutes to defend myself, eh?
If you don't understand law, you hire a lawyer, if you don't understand standards you hire a consultant who does.
Then that defeats the whole idea of being a hobbyist or trying to start a business. Just hire someone! It's crap logic to go by that. Sometimes you want to learn how to do something yourself but just need some kind of resource (such as a book) to get started. Those official standards are very expensive and hard to get, and also not that clear to understand so you'd still end up wanting a book or something to guide you through it. I find they do the same with building codes. They tell you you must do things a certain way by law but then they don't actually make the info easily accessible.
Basically the only thing you can do because of this is to google stuff and just try to make it as best as you can. You're surely to overlook something though.
-
Then that defeats the whole idea of being a hobbyist or trying to start a business. Just hire someone! It's crap logic to go by that.
Not only that. Even if you are going to hire a pro in some field, you are going to communicate better and advance more if you have a good working knowledge of that field.
An engineer who has some knowledge of the law can communicate much better with a lawyer than one who doesn't.
Just like a lawyer who has some knowledge of engineering con communicate better with an engineer.
-
Standards change often enough and apply in such specific combinations to any given field that trying to condense them into any other format is a loosing proposition. Big corporations have the resources to have internal standards which do this, but they still go back to the specific standards to audit a design before its released. So even if you had such a book you would still need to refer to the standard.
I thought I made it quite clear in my opening post that I was asking for book recommendations similar to what exist for EMC except aimed at (or at least better covering) the safety aspects of product compliance testing. I wasn't asking for a book that could substitute for the standards.
EMC is largely the same thing across standards, don't emit more than X@Y Hz and don't fail/stop/error when exposed to P@R Hz despite the differences in specific measuring setups the underlying principles are common. Safety standards pull together all sorts of radically different details. A book couldn't present any coherent message as there are too many exceptions/conditions/interactions and the choice tree explodes, especially when trying to simultaneously meet several standards from different areas or regions. It would end up like the standards themselves with this is applicable expect when in categories of 5.2.1.a or methods from any of 3.21 excepting 3.21.5.
You could have a book full of common design patterns, but linking them back to when to use them is almost impossible. There are too many interactions and design variables. The "simple" example that brought you here of isolation of a semiconductor device already spits out numerous solutions that were mentioned in the thread:
different mounting with more creepage
protective bonding for heatsink
isolation of heatsink
Depending on the standards applying and specifics of the arrangement some or all of those may not be applicable. Just trying to discuss accessible parts gets off into substantially different definitions and requirements.
A book is a nice idea but is completely impractical. As I said, big organisations can make a good attempt at it; because they usually stick to a well defined set of regions and applicable standards for their products. The effort put into keeping those documents up to date is non-trivial, I can't see a large enough homogenous market that could support that sort of work being distributed as a book.
-
I haven't looked for books on safety regulation standards, and perhaps like many here I've collected a few on EMC over the years. The process of making a design fit through the complex path of standards is unfortunately often learned the hard way, that is, the expensive way.
Yep, same here. And if you only go through this process once every couple of years you either tend to forget some of the more arcane details or don't get a good feel for the often-Byzantine logic behind some of the decisions made. For example, my current client had UL rule against classifying their boards as Pollution Degree 1 even though an approved urethane conformal coating (HumiSeal 1A33) was applied. Having to go back and redesign for the much more onerous creepage and clearance of PD2 was a massive pain and quite a costly one, too. This was before my time, so not much context is available, but had they known that getting anything classified as PD1 is exceptionally difficult - even with a sealed enclosure and conformal coating or silicone potting used quite liberally - they probably wouldn't have tried in the first place.
2. Which categories of standards apply? Are there contradictory in requirements? Are the solutions to the problems contradictory? In what sequence should the various approval stages be completed? How do these interact with the processes implemented (either purposefully or frenetically) at the company? Is the difference between a directive and a standard understood?
Bingo! I'd also add that there is often considerable conflict between what is best for EMC compliance and safety agency approval. A classic example is the Y capacitance (ie - line to ground, for those not familiar with this lingo) that EMC wants to shunt common mode noise current to ground vs. what safety will allow to prevent making the ground conductor "too hot."
On more than one occasion I've suggested to Dave that this forum really needs some type of Standards / Approvals / Agency / etc. category.
Well this suggestion gets my approval.... <rimshot!>