Chassis Ground layout for better ESD and EMI compliance

EMI

make

won't,

I meant that I've never seen a standard for our market, well, other than the UL and CE stuff for the mains inputs. I looked up information on UL creepage numbers. It appears I can get away with 2mm (1206 caps).

That's the nice thing about standards; there are so many to choose from.

Likely never. They *do* know about TH parts, though. I'd think the lead inductance of a TH part would be much worse, too.

Reply to
krw
Loading thread data ...

won't,

For this board I'm also in a big space crunch. I need to fit an isolated driver channel into about .6"x1".

Heat, mostly, I think. Chip-scale transformers aren't wonderful. I decided not to use the ADuM5201 because the ADM2587 turned out to be simpler. It supplies 15mA, which is more than enough with RS-422 transceiver swept under the covers.

Reply to
krw

[...]

There you go, that's exactly what I meant. CE/UL are the usual standards in the civilian world.

Usually not, the application or the target market point to a particular standard and then you don't really have a choice. If more than one market is to be served with your product it is possible that you have to adhere to more than one of the standards. But you can't wiggle out of a standard just to pick a more lenient one.

Doesn't have to be. After all, that's the way they made UHF TV-tuners in the 60's and those worked well above 300MHz ;-)

If you or the powers-that-be absolutely don't want to have through-hole there are L-bracketed long hi-rel SMT caps. But they'll cost ya. Even for a hi-rel client where cost isn't all that critical we opted for through-hole after looking at pricing. Because we'd have needed 50 of them (electrolytic avoidance).

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

This is the kind of thing that makes me wish books like AoE had a chapter on EM fields and waves. I asked Win Hill for it but no answer, so I guess the next edition will just have updated opamps. Not worth buying.

A coax is a waveguide, and non-conducting RF waveguides (no central conductor) are grounded at both ends. As far as the E field alone is concerned, I would guess it only needs to be held at ground potential from one end.

--

Reply in group, but if emailing add one more
zero, and remove the last word.
Reply to
Tom Del Rosso

I thought you meant standards for the market. There are none and few of our products go through UL or any like certification.

We don't do medical or aviation. There are no such standards.

The reality is that 300MHz is only the fundamental. We've had issues with the

100+ harmonic (not that I expect to even look at 30GHz).

Price here likely isn't much of a problem. *I* would hate it, but management would likely not blink. They would *not* like TH parts. We already have too many of them (connectors).

Reply to
krw

Every market has standards. The UL/CE deals with safety, EMC and all that. That's what I meant, there are very few markets where you don't have to adhere to that. Alien spacecraft maybe, but only of those guys don't have trial lawyers ;-)

Then there's the other standards such as signal levels. Or the RS422 you just mentioned, because that's also a standard.

But you just mentioned UL/CE which is where you'll find (or already found) the answer to your creepage distance question. If your product can be handled by a living being, could cause a fire or any other harm, there will be a standard to adhere to. I have seen mfgs blissfully unaware of that but this does not relieve them of liability.

IME you almost have to migrate from board level to system level EMI mitigation once your concerns are spectra above half a gigeehoitz.

Ok, if the caps are not too large SMT may be ok. Make sure you get the proper ratings on the caps. If the barrier is a safety requirement that's one of the first thing the engineers at the notified body will likely want to see. Regular caps often do not have an AC rating so are typically a red flag to them. Y and X caps do have such ratings.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

creepage

UL/CE, safety, yes, only where our customer demands it. EMC because governments demand it. The specific voltages or creepage requirements, nope, except on the mains side.

It's RS-422 because that's the standard we chose to use. It wasn't selected by someone else. We "own" both ends.

For the voltage protection *we* decide is useful. There is no standard for such.

On the mains side UL says something about this, sure. On the other end of the box, nope.

now,

IME, you have to start with the board to have any hope of system level compliance. In this particular box, there is little else other than the board and a steel box.

Any potential differences are in the grounds between boxes a couple of thousand feet away (which includes mobile units). Of course someone can mix the black and green wires, but I'd expect smoke somewhere.

Reply to
krw

the

creepage

Then you are ok, as long as your product specs don't claim a certain breakdown immunity.

[...]

now,

the

That is true, especially since you aren't bound by too much in safety requirements. A steel box offers nice system level shielding opportunities though.

Then I'd be careful. This sort of ground loop isolation function often falls under some standard. This is why many Ethernet transformers have breakdown voltage specs and often agency ratings.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.