Chassis Ground layout for better ESD and EMI compliance
Mark
are one side, Ethernet conns on the 2nd side and RS232 on the 3rd side
of the board. All are through hole connectors and on top layer with
separate chassis ground area on each side for connectors shields. So I
have three chassis ground islands on th board for three sets of these
connectors. If we have to connect the chassis ground to the signal
ground through 1M Resistor and 1nf/2KV capacitor, which of the
following is better and why.
1) keep the 3 chassis ground areas separate and connect each of them
to the signal ground through 1M Res and 1nf/2KV cap. In this case
there will be 3 pairs of 1M and 1nf in parallel. Is this an issue?
2) Have one chassis ground covering all three sides of the board and
connect to the signal ground with 1M Res and 1nf/2KV cap.
3) What is the better place to connect the chassis ground and signal
ground? Is it at the power supply return?
4) I have seen one reference design on the net with power supply
return and connector shields connected together in a copper area and
this copper area connected to the signal ground through a ferrite
bead.
It is an 8 layer board with two signal ground planes and chassis
ground on all layers beneath the connectors stitched together at
multiple places with vias.
Thanks
mark
John Larkin
wrote:
Why isolate the PCB ground from the chassis? The best thing to do is
bolt the (single!) PCB ground plane to the chassis as many places as
possible, and ground the connector shells to the chassis, too.
John
Mark
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Thu, 19 May 2011 10:44:57 -0700 (PDT), Mark <markjsu...@gmail.com>
Thanks John. Isolated the chassis and signal grounds for better ESD
immunity.
-M
John Larkin
wrote:
>On May 19, 11:03 pm, John Larkin
But you're making it worse!
The pattern is familiar: isolate grounds for some reason, then get
cold feet and add networks between the isolated ground and the chassis
ground to try to make them the same potential.
I have a customer who is doing that to me now: insisting on isolation,
and insisting on a network to bypass the isolated ground to chassis
ground. The details are even worse. Makes no sense.
This same customer insists on shielding all cables, but grounding the
shields at one end only! Nominally the source end; but the cables have
signals going in both directions!
I tried to give them an analogy they could understand:
Imagine the end of a garden hose. Imagine a bunch of twigs sticking
out the end. Grab the end of the hose and shake it. What happens to
the twigs?
John
anospam
"Mark" <markj...@gmail.com> wrote in message
news:ceada41c-ba14-4357...@h36g2000pro.googlegroups.com...
try using a small choke, it provices DC path, but limits AC noise.
Best place is near PS, and one only if you can to prevent ground loops
The 1 meg are more for static bleed.
Joerg
The areas where the screws are should be compliantly plated. Nickel,
gold, whatever reduces dissimilar metal effects and thus corrosion. And
_no_ solder, a very common mistake. Solder tends to creep under pressure
and the mechanical force gradually sags. Worst case the screw works
itself out during a lengthy and rough transport.
Another common mistake is to have the clearance too small. Then a washer
may partially ride on the soldermask, not a good thing.
--
Regards, Joerg
http://www.analogconsultants.com/
"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Joerg
[...]
> I have a customer who is doing that to me now: insisting on isolation,
> and insisting on a network to bypass the isolated ground to chassis
> ground. The details are even worse. Makes no sense.
>
> This same customer insists on shielding all cables, but grounding the
> shields at one end only! Nominally the source end; but the cables have
> signals going in both directions!
>
> I tried to give them an analogy they could understand:
>
> Imagine the end of a garden hose. Imagine a bunch of twigs sticking
> out the end. Grab the end of the hose and shake it. What happens to
> the twigs?
>
They grow roots into the hose?
<duck and run>
--
SCNR, Joerg
John Larkin
wrote:
>John Larkin wrote:
>
>[...]
>
>> I have a customer who is doing that to me now: insisting on isolation,
>> and insisting on a network to bypass the isolated ground to chassis
>> ground. The details are even worse. Makes no sense.
>>
>> This same customer insists on shielding all cables, but grounding the
>> shields at one end only! Nominally the source end; but the cables have
>> signals going in both directions!
>>
>> I tried to give them an analogy they could understand:
>>
>> Imagine the end of a garden hose. Imagine a bunch of twigs sticking
>> out the end. Grab the end of the hose and shake it. What happens to
>> the twigs?
>>
>
>They grow roots into the hose?
>
><duck and run>
Recognize this?
ftp://jjlarkin.lmi.net/Zeit.jpg
John
k...@att.bizzzzzzzzzzzz
Ground early and often! ...except where you can't. :-(
k...@att.bizzzzzzzzzzzz
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
BTW, John, what's your take on this (I have the problem, in spades):
http://www.analog.com/static/imported-files/application_notes/AN-0971.pdf
Joerg
Can't read the upper word but this is funny. They had to cancel one day
of the Amgen tour. This Sunday we had snow here in the foothills, at
elevation 1400ft. In the middle of May!
Oh, yeah, how could I forget. Hot is global warming and cold is always
just weather. Right?
--
Regards, Joerg
k...@att.bizzzzzzzzzzzz
Of course not! Hot is "global warming" and cold is "global climate change".
Joerg
If you must use these for safety reasons: Capacitive stitching has about
the same effect at the test lab as a super-loud muffler has on the
sheriff who just stopped a kid for speeding ;-)
Then one also has to worry about the reverse path. What if some large EM
spike gets in?
No way to use optocouplers instead? It'll be cheaper, too.
John Larkin
wrote:
>John Larkin wrote:
>> On Thu, 19 May 2011 13:52:12 -0700, Joerg <inv...@invalid.invalid>
>> wrote:
>>
>>> John Larkin wrote:
>>>
>>> [...]
>>>
>>>> I have a customer who is doing that to me now: insisting on isolation,
>>>> and insisting on a network to bypass the isolated ground to chassis
>>>> ground. The details are even worse. Makes no sense.
>>>>
>>>> This same customer insists on shielding all cables, but grounding the
>>>> shields at one end only! Nominally the source end; but the cables have
>>>> signals going in both directions!
>>>>
>>>> I tried to give them an analogy they could understand:
>>>>
>>>> Imagine the end of a garden hose. Imagine a bunch of twigs sticking
>>>> out the end. Grab the end of the hose and shake it. What happens to
>>>> the twigs?
>>>>
>>> They grow roots into the hose?
>>>
>>> <duck and run>
>>
>>
>> Recognize this?
>>
>> ftp://jjlarkin.lmi.net/Zeit.jpg
>>
>
>Can't read the upper word but this is funny.
The sticker says Zeitgeist Racing Team. Or maybe Zeitgeist Drinking
Team. It was kinda hard to photograph.
"Warm Beer * Cold Women"
They had to cancel one day
>of the Amgen tour. This Sunday we had snow here in the foothills, at
>elevation 1400ft. In the middle of May!
Lots of snow, and some rain, in Truckee last few days.
ftp://jjlarkin.lmi.net/CABIN.JPG
>
>Oh, yeah, how could I forget. Hot is global warming and cold is always
>just weather. Right?
I just got an email from Alpine Meadows. They are re-opening for
skiing for the 4th of July weekend. They got 72 feet of snow this
year.
All the reservoirs are full, and the melt has just begun. Long, hot
showers. So much for the idiot climate predictions.
John
Joerg
Some day we'll have to meet there for a burger and Chimay Triple.
>
> They had to cancel one day
>> of the Amgen tour. This Sunday we had snow here in the foothills, at
>> elevation 1400ft. In the middle of May!
>
> Lots of snow, and some rain, in Truckee last few days.
>
> ftp://jjlarkin.lmi.net/CABIN.JPG
>
>> Oh, yeah, how could I forget. Hot is global warming and cold is always
>> just weather. Right?
>
> I just got an email from Alpine Meadows. They are re-opening for
> skiing for the 4th of July weekend. They got 72 feet of snow this
> year.
>
> All the reservoirs are full, and the melt has just begun. Long, hot
> showers. So much for the idiot climate predictions.
>
A few years ago that was different but today I couldn't name one person
in this whole neighborhood who still believes AGW is true. Bill Sloman
could go door to door and stomp his feet on the ground, it won't change
their minds. Ok, this ain't San Francisco but it's still California.
George Herold
> John Larkin wrote:
> > wrote:
>
> >> John Larkin wrote:
> >>> On Thu, 19 May 2011 13:52:12 -0700, Joerg <inva...@invalid.invalid>
>
> - Show quoted text -- Hide quoted text -
>
> - Show quoted text -
Aren't we near the bottom of the 11 year solar cycle? In five or six
years there will be a different tune again.
We're having record rains, here in Western NY. So muddy, that driving
the tractor means leaving 6" (15cm) tracks in the yard.
George H.
Chieftain of the Carpet Crawlers
>try using a small choke, it provices DC path, but limits AC noise.
>
>Best place is near PS, and one only if you can to prevent ground loops
>
>The 1 meg are more for static bleed.
Maybe next time you should try actually reading the post.
John Larkin
Apparently huge places, like lots of India and China, are feeding
their people from agriculture based on over-pumping aquifers. And they
are running out. If anything ever revives our balance-of-trade fiasco,
it will be food.
John
George Herold
>
> - Show quoted text -
Hmmm, Isn't there some aquifer in the mid-west that our farmers are
pumping dry?
Too much water in the Mississippi and not enough in Texas.
George H.
John Larkin
wrote:
>
No, it increases noise.
>
>Best place is near PS, and one only if you can to prevent ground loops
Ground loops are good. It means you're keeping everything grounded.
John
Spehro Pefhany
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>
>
>Apparently huge places, like lots of India and China, are feeding
>their people from agriculture based on over-pumping aquifers. And they
>are running out. If anything ever revives our balance-of-trade fiasco,
>it will be food.
>
>John
Not only China and India.. water is being mined all over. For example,
the Ogallala aquifer too: "some estimates say it will dry up in as
little as 25 years" .. it's dropping 36" per year, 3/4 as much as the
water table in north China.
http://en.wikipedia.org/wiki/Ogallala_Aquifer
It's apparently enough worldwide to measurably increase sea levels
even without warming.
Libya is an interesting case:
http://news.bbc.co.uk/2/hi/science/nature/4814988.stm
Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
sp...@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
k...@att.bizzzzzzzzzzzz
"Capacitive" stitching?
>Then one also has to worry about the reverse path. What if some large EM
>spike gets in?
Not *too* concerned about that. I'm more worried about the huge current
switching at 180MHz and its effect in the test chamber. ;-)
>No way to use optocouplers instead? It'll be cheaper, too.
No. What powers the optocoupler? I really need an isolated 422 interface.
Lots of them.
Bill Sloman
> John Larkin wrote:
> > wrote:
>
> >> John Larkin wrote:
> >>> On Thu, 19 May 2011 13:52:12 -0700, Joerg <inva...@invalid.invalid>
Score one for Exxon-Mobil's lying propaganda. Joerg seems to live in a
neighbourhood where no-one can read the scientific literature for
themselves, or read critically enough to realise how nonsensical the
Exxon-Mobil-inspired propaganda actually is. This comes close to
saying that they haven't got enough minds between them for it to be
worth trying to change them.
John Larkin's susceptibility to denialist propaganda is easy to
explain - he doesn't even understand the basic facts about Darwinian
evolution and DNA, and still thinks that because the weather is
chaotic, climate has to be be too, which clearly indicates that he
wasn't paying any attention in any of his science classes that weren't
obviously relevant to electronics - but it seems a bit odd that an
entire neighbourhood could be just as weak in science.
--
Bill Sloman, Nijmegen
Bill Sloman
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Thu, 19 May 2011 16:46:23 -0700, Joerg <inva...@invalid.invalid>
> wrote:
>
>
>
>
>
>
>
>
>
> >John Larkin wrote:
> >> On Thu, 19 May 2011 13:52:12 -0700, Joerg <inva...@invalid.invalid>
Except that you can't find any climate prediction that says California
wasn't going to get lots of snow last winter.
You can find all sorts of idiot journalists saying that anthropogenic
global warming has this or that implication for current weather, but
actual climate predictions for the last few years and the next few
years, by actual climate scientists, published in peer-reviewed
journals are rather thinner on the ground.
The one example that comes to mind
http://www.agu.org/pubs/crossref/2010/2009JD013568.shtml
http://eprints.ifm-geomar.de/8738/1/2009JD013568-pip.pdf
actually predicted colder winters - in Europe - just before we had two
cold winters in succession.
John Larkin doesn't read carefully enough to distinguish between idiot
journalism and serious scholarship, and seems to think that the
defects of the former debunk the latter. Not a particularly sensible -
or insightful - attitude.
--
Bill Sloman, Nijmegen
Tom Del Rosso
>
> I tried to give them an analogy they could understand:
>
> Imagine the end of a garden hose. Imagine a bunch of twigs sticking
> out the end. Grab the end of the hose and shake it. What happens to
> the twigs?
I think they fall out, but how is that an analogy?
--
Reply in group, but if emailing add one more
zero, and remove the last word.
Joerg
Yup. Quote "Input-to-output ground plane stitching".
>> Then one also has to worry about the reverse path. What if some large EM
>> spike gets in?
>
> Not *too* concerned about that. I'm more worried about the huge current
> switching at 180MHz and its effect in the test chamber. ;-)
>
Wait until them thar flashin' shows up in them clouds and the computer
gets all hairbrained :-)
But it take an incredible field-strength to upset these couplers.
>> No way to use optocouplers instead? It'll be cheaper, too.
>
> No. What powers the optocoupler? I really need an isolated 422 interface.
> Lots of them.
I have mostly powered them via little transformers driven with a few
100kHz tapped off somewhere. Additional motivators were the requirements
of 2nd source and low cost.
John Larkin
All they have to do is count cords of wood.
John
whit3rd
> I have 2 USB, 2 Ehthernet and 2 RS232 connectors on the PCB. USB conns
> are one side, Ethernet conns on the 2nd side and RS232 on the 3rd side
> of the board.
Of these, only the USB has a grounded shield connection; a chassis
ground connection hear that USB shroud is called for.
The PCB ground separation is intended to keep small signals from
coupling through ground 'bounce'; neither RS232 nor Ethernet nor
USB are sensitive to a few millivolts of common-mode (ground bounce)
noise, so what's the problem?
Bill Sloman
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Fri, 20 May 2011 04:14:24 -0700 (PDT),BillSloman
>
> >On May 20, 3:07 am, Joerg <inva...@invalid.invalid> wrote:
> >> John Larkin wrote:
> >> > On Thu, 19 May 2011 16:46:23 -0700, Joerg <inva...@invalid.invalid>
> >> > wrote:
>
> >> >> John Larkin wrote:
> >> >>> On Thu, 19 May 2011 13:52:12 -0700, Joerg <inva...@invalid.invalid>
> >> >>> wrote:
>
>
> >> > All the reservoirs are full, and the melt has just begun. Long, hot
> >> > showers. So much for the idiot climate predictions.
>
> >> A few years ago that was different but today I couldn't name one person
> >> in this whole neighborhood who still believes AGW is true.BillSloman
> >> could go door to door and stomp his feet on the ground, it won't change
> >> their minds. Ok, this ain't San Francisco but it's still California.
>
> >Score one for Exxon-Mobil's lying propaganda. Joerg seems to live in a
> >neighbourhood where no-one can read the scientific literature for
> >themselves, or read critically enough to realise how nonsensical the
> >Exxon-Mobil-inspired propaganda actually is. This comes close to
> >saying that they haven't got enough minds between them for it to be
> >worth trying to change them.
>
> All they have to do is count cords of wood.
You really don't get it do you. The problem is not that the weather
isn't conforming to your ideas of what anthropogenic global warming
means for your local climate, but that you've never gone to the
trouble of finding out what the climate experts actually have to say
about the implications of anthropogenic global warming.
The kindest way of explaining this is that you lack the wit to realise
that journalists wanting to make something of a local weather story
will "explain" whatever it was in terms of anthropogenic global
warming, when its just weather - as any climate scientist would have
told them if they had gone to the trouble of asking. Sadly, that
doesn't make for the kind alarmist story that journalists like to
peddle to titivate the gullible public, of which you are a depressing
example.
--
Bill Sloman, Nijmegen
Ecnerwal
John Larkin <jjla...@highNOTlandTHIStechnologyPART.com> wrote:
> All they have to do is count cords of wood.
...and they know what's happening all over the globe, eh?
Try counting cords of wood that used to be houses in Alabama, for a
somewhat different metric. But, of course, that's not one tiny area of
California, and one tiny area of California is, according to you, the
entire globe.
All the rain/snow which has made the Mississippi exceedingly full also
didn't fall in California, so it didn't happen on your "globe", either.
Got it. Or your warped view is that global warming somehow means
automatically drier, despite the fact that warm air carries more
moisture from the ocean and dumps it in the mountains around you as more
snow when it rises and cools. Or in Ohio to run down and flood
Louisiana, as the case may be.
Nobody but Faux "News" thinks "a warmer climate" == "everywhere
uniformly a little bit warmer" - in point of fact, "more chaotic" has
been both predicted and observed - individual areas hotter, colder,
wetter, drier, less predictable, more violent storms (see Alabama).
Unseasonably hot April weather causing fruit trees to flower and be
nailed by unseasonably late frosts are perfectly in line with "warmer on
average" even though one involves a period of economically damaging
cold. Two weeks of 30 degrees above average and two nights of 10 degrees
below average will do that just fine, and still be "much warmer on
average" - as will "your part of California having an Antarctic climate"
and the rest of the country (or globe) being 1 degree (or less) warmer.
--
Cats, coffee, chocolate...vices to live by
Please don't feed the trolls. Killfile and ignore them so they will go away.
John Larkin
Is your problem radiated 300 MHz stuff, failing radiated EMI specs?
We use tons of these couplers, but we don't officially EMI test our
VME modules, since they are just boards that go into racks.
We sometimes wave a spectrum analyzer around a board and see what's
radiating. Things are getting scary. We buy some "programmable"
crystal oscillators that have an internal PLL that runs at a couple of
GHz, and they radiate a lot. The ADI couplers are indeed noisy. A 270
MHz ARM, or an FPGA with a 1 GHz serdes PLL, are probably bad; we'll
be doing that combo soon, and I'll snoop it for fun.
We sometimes worry that such stuff will get into opamps and get
rectified into noise and weird offsets, but haven't seen that as a
problem so far. The RF seems very local to the chips, usually.
If you're failing EMI, you might try using the Silabs isolators. Or
the TI parts. I think both have drop-ins for the ADUM parts.
John
John Larkin
Of course I have. I just think they have a decent probability of being
wrong. Sciences not subject to experimental verification tend to get
things wrong [1]. And they should avoid making specific predictions
(as in our region headed for hot-and-dry) lest reality contradict.
But implications? They seem to be uniformly dismal. I can see how that
would appeal to you.
John
[1] Superconductivity continues to confound the condensed-matter boys.
See the latest Science mag.
John
John Larkin
<MyName...@ReplaceWithMyVices.Com.invalid> wrote:
>In article <68uct69vek1n0v9c7...@4ax.com>,
> John Larkin <jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>> All they have to do is count cords of wood.
>
>...and they know what's happening all over the globe, eh?
>
>Try counting cords of wood that used to be houses in Alabama, for a
>somewhat different metric. But, of course, that's not one tiny area of
>California, and one tiny area of California is, according to you, the
>entire globe.
Cite?
>
>All the rain/snow which has made the Mississippi exceedingly full also
>didn't fall in California, so it didn't happen on your "globe", either.
>Got it. Or your warped view is that global warming somehow means
>automatically drier, despite the fact that warm air carries more
>moisture from the ocean and dumps it in the mountains around you as more
>snow when it rises and cools. Or in Ohio to run down and flood
>Louisiana, as the case may be.
Google "climate change california drought" for lots of hilarious
"emotive imagery", like...
http://www.theclimatehub.com/will-climate-change-make-drought-the-norm-in-california
>
>Nobody but Faux "News" thinks "a warmer climate" == "everywhere
>uniformly a little bit warmer" - in point of fact, "more chaotic" has
>been both predicted and observed - individual areas hotter, colder,
>wetter, drier, less predictable, more violent storms (see Alabama).
>Unseasonably hot April weather causing fruit trees to flower and be
>nailed by unseasonably late frosts are perfectly in line with "warmer on
>average" even though one involves a period of economically damaging
>cold. Two weeks of 30 degrees above average and two nights of 10 degrees
>below average will do that just fine, and still be "much warmer on
>average" - as will "your part of California having an Antarctic climate"
>and the rest of the country (or globe) being 1 degree (or less) warmer.
Now you are confusing weather with climate.
John
Joerg
> On Fri, 20 May 2011 11:38:16 -0400, Ecnerwal
> <MyName...@ReplaceWithMyVices.Com.invalid> wrote:
[...]
>
>> All the rain/snow which has made the Mississippi exceedingly full also
>> didn't fall in California, so it didn't happen on your "globe", either.
>> Got it. Or your warped view is that global warming somehow means
>> automatically drier, despite the fact that warm air carries more
>> moisture from the ocean and dumps it in the mountains around you as more
>> snow when it rises and cools. Or in Ohio to run down and flood
>> Louisiana, as the case may be.
>
> Google "climate change california drought" for lots of hilarious
> "emotive imagery", like...
>
> http://www.theclimatehub.com/will-climate-change-make-drought-the-norm-in-california
>
Quote "As global warming melts the mountain snow that supplies fresh
water to the densely populated Californian region ..."
ROFL! What are they smoking over there? But wait. East Anglia? Wasn't
that where ...
[...]
John Larkin
wrote:
>John Larkin wrote:
>> On Fri, 20 May 2011 11:38:16 -0400, Ecnerwal
>> <MyName...@ReplaceWithMyVices.Com.invalid> wrote:
>
>[...]
>
>>
>>> All the rain/snow which has made the Mississippi exceedingly full also
>>> didn't fall in California, so it didn't happen on your "globe", either.
>>> Got it. Or your warped view is that global warming somehow means
>>> automatically drier, despite the fact that warm air carries more
>>> moisture from the ocean and dumps it in the mountains around you as more
>>> snow when it rises and cools. Or in Ohio to run down and flood
>>> Louisiana, as the case may be.
>>
>> Google "climate change california drought" for lots of hilarious
>> "emotive imagery", like...
>>
>> http://www.theclimatehub.com/will-climate-change-make-drought-the-norm-in-california
>>
>
>Quote "As global warming melts the mountain snow that supplies fresh
>water to the densely populated Californian region ..."
Exactly. The snow melts every year. There are no glaciers around here.
We get a lot of our water from the melting snow. He says it like it's
a bad thing for snow to melt.
What would be bad is if it didn't melt, namely we had another ice age
and all the snow piled up, like it did 20K years ago.
72 feet of snow at Alpine Meadows this season. Maybe a little more to
come; it snowed again last night. How's the weather at your place?
>
>ROFL! What are they smoking over there? But wait. East Anglia? Wasn't
>that where ...
"Emotive imagery." Used to be called "science."
John
Joerg
It peaks above 70F today. But only today, the weekend forecast is rain
and cold. Somehow the 2011 summer will either be short or not happening :-(
Heating well into May we have now blown through five cords of wood this
season, half a cord was scavenged from next year's supply so we may have
to order extra. 10 years ago less than two cord sufficed.
>> ROFL! What are they smoking over there? But wait. East Anglia? Wasn't
>> that where ...
>
> "Emotive imagery." Used to be called "science."
>
A lot of this "science" sounds more like a ruse to press more taxes out
of us. To feed some gravy train.
Joerg
> John Larkin wrote:
>> On Fri, 20 May 2011 09:28:32 -0700, Joerg <inv...@invalid.invalid>
>> wrote:
>>
>>> John Larkin wrote:
>>>> On Fri, 20 May 2011 11:38:16 -0400, Ecnerwal
>>>> <MyName...@ReplaceWithMyVices.Com.invalid> wrote:
>>> [...]
>>>
>>>>> All the rain/snow which has made the Mississippi exceedingly full also
>>>>> didn't fall in California, so it didn't happen on your "globe", either.
>>>>> Got it. Or your warped view is that global warming somehow means
>>>>> automatically drier, despite the fact that warm air carries more
>>>>> moisture from the ocean and dumps it in the mountains around you as more
>>>>> snow when it rises and cools. Or in Ohio to run down and flood
>>>>> Louisiana, as the case may be.
>>>> Google "climate change california drought" for lots of hilarious
>>>> "emotive imagery", like...
>>>>
>>>> http://www.theclimatehub.com/will-climate-change-make-drought-the-norm-in-california
>>>>
>>> Quote "As global warming melts the mountain snow that supplies fresh
>>> water to the densely populated Californian region ..."
>> Exactly. The snow melts every year. There are no glaciers around here.
Forgot to mention, we do have those. For example Mt.Shasta, and its
glaciers are growing ...
Rich Grise
> John Larkin wrote:
>>
>> I tried to give them an analogy they could understand:
>>
>> Imagine the end of a garden hose. Imagine a bunch of twigs sticking
>> out the end. Grab the end of the hose and shake it. What happens to
>> the twigs?
>
> I think they fall out, but how is that an analogy?
>
to ground at the point that's closest to the hub of the star. Or maybe
closest to the signal source. Or toss a coin. ;-P (I wonder if there's
a rule of thumb?)
But I have no idea how that relates to shaking sticks out of a garden hose.
Cheers!
Rich
Bill Sloman
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Fri, 20 May 2011 08:33:41 -0700 (PDT),BillSloman
So where is the prediction?
>I just think they have a decent probability of being
> wrong. Sciences not subject to experimental verification tend to get
> things wrong [1]. And they should avoid making specific predictions
> (as in our region headed for hot-and-dry) lest reality contradict.
The problem is that you haven't pointed to any such prediction, so it
looks exactly as if you are wasting our time by being rude about an
prediction that doesn't actually exist.
> But implications? They seem to be uniformly dismal. I can see how that
> would appeal to you.
They don't. I'd be much happier if anthropogenic global warming just
went away without us having to rebuild our energy economy, but since
we do seem to be running out of oil and natural gas anyway, that
particular cloud does have a silver lining.
>[1] Superconductivity continues to confound the condensed-matter boys.
> See the latest Science mag.
Since investigating super-conductivity does involve experiments, I
don't see what kind of point you think you are making here. And you
seem to have missed the point that anthropogenic global warming
involves more than just a rise in the earth's average temperature
(which does happen to be going on at the moment) but also changes in
the temperature gradients up through the atmosphere as well as a
variety of other measurable changes.
You make a song and dance about Ca;ifornia weather not comforming to
your ideas of what global warming theory predicts (not that you can be
bothered to point us to any such prediction) but you can't be bothered
to find out any of the other predictions it makes, even though some of
them are going to be more easily detectable.
Anthropogenic global warming is a falsifiable hypothesis, despite the
fact that you go around claiming that it isn't. If you put as much
energy into finding some predictions it actually did make - rather
than inventing predictions you would have liked it to have made - you
might come over as something less of brain-washed propaganda victim.
--
Bill Sloman, Nijmegen
John Larkin
Around 10 years ago, the press here was full of official predictions
that California was becoming hot and dry, because of AGW. Since then,
it's been mostly cool and wet.
AGW people can always keep extending their time scales and still
claiming they are right. Cute trick.
And there was Viner's (CRU) 2000 prediction:
within a few years winter snowfall will become “a very rare and
exciting event”. “Children just aren’t going to know what snow is,” he
said.
Yeah, right.
John
John Larkin
<ri...@example.net.invalid> wrote:
>Tom Del Rosso wrote:
>> John Larkin wrote:
>>>
>>> I tried to give them an analogy they could understand:
>>>
>>> Imagine the end of a garden hose. Imagine a bunch of twigs sticking
>>> out the end. Grab the end of the hose and shake it. What happens to
>>> the twigs?
>>
>> I think they fall out, but how is that an analogy?
>>
>I'm wondering the same thing. If it were my call, I'd connect the shield(s)
>to ground at the point that's closest to the hub of the star. Or maybe
>closest to the signal source. Or toss a coin. ;-P (I wonder if there's
>a rule of thumb?)
Tossing a coin is about as good as most people get.
>
>But I have no idea how that relates to shaking sticks out of a garden hose.
When you shake the hose, the twigs shake too.
When a shield is ungrounded, and there's noise induced onto the
shield, the wires inside will have the same noise. Shake the shield
and the wires inside shake too.
John
Rich Grise
> On Fri, 20 May 2011 10:15:15 -0700, Rich Grise
>>
>>hose.
>
> When you shake the hose, the twigs shake too.
>
> When a shield is ungrounded, and there's noise induced onto the
> shield, the wires inside will have the same noise. Shake the shield
> and the wires inside shake too.
>
no current path, so how can any signal be induced, except at, say, 1/4
wave? I think I remember hearing (reading) that that's the rationale for
leaving one end flapping in the breeze in the first place.
Thanks,
Rich
Joerg
If the cables are non-diff that'll result in noise capacitively coupled
in. With diff it can be ok. But sometimes grounding at both ends is not
advised for other reasons, such as erroneous large currents. "Dude,
don't that smell like pee vee cee around here?"
John Larkin
If the shield is not very short, it has serious inductance back to
ground. And there are a zillion frequencies where the impedance spikes
up. Not just 1/4 wave, but 5/4, 9/4, etc. And nearby frequencies where
Z is high but not infinite. Every motor controller, light dimmer,
relay contact, radio station, cell phone, ESD zap, and lightning bolt
will shake the ungrounded end. And the wires inside.
Take a long cable. Ground the shield a distance away, in another room
or another building. Scope the voltage on the shield at the end where
it's ungrounded, against local ground.
Repeat for conductors inside the shield.
Now, ground the shield at the scope end.
John
John Larkin
wrote:
>Rich Grise wrote:
>> John Larkin wrote:
>>> On Fri, 20 May 2011 10:15:15 -0700, Rich Grise
>>>> But I have no idea how that relates to shaking sticks out of a garden
>>>> hose.
>>> When you shake the hose, the twigs shake too.
>>>
>>> When a shield is ungrounded, and there's noise induced onto the
>>> shield, the wires inside will have the same noise. Shake the shield
>>> and the wires inside shake too.
>>>
>> Ah, there's the rub. Since only one end of the shield is connected, there's
>> no current path, so how can any signal be induced, except at, say, 1/4
>> wave? I think I remember hearing (reading) that that's the rationale for
>> leaving one end flapping in the breeze in the first place.
>>
>
>If the cables are non-diff that'll result in noise capacitively coupled
>in. With diff it can be ok.
Depends on how much common-mode noise you can stand. Seems to me that
less is better than more.
But sometimes grounding at both ends is not
>advised for other reasons, such as erroneous large currents. "Dude,
>don't that smell like pee vee cee around here?"
That would be unusual, two grounds with enough voltage diference, at a
low enough impedance, to heat up a cable shield.
John
Les Cargill
<dsnip>
>
> Anthropogenic global warming is a falsifiable hypothesis, despite the
> fact that you go around claiming that it isn't.
Not even close...
--
Les Cargill
Bill Sloman
http://www.independent.co.uk/environment/snowfalls-are-now-just-a-thing-of-the-past-724017.html
It's actually what a reporter - Christopher Onions - wrote after
talking to David Viner. They were talking about English winters, not
Californian winters, and it looks as if Viner was extrapolating from
what had been happening over the previous few years, which - in
retrospect - looks more like a local warm period driven by the
Atlantic Multi-Decadal Oscillation
As a prediction about California, it's a non-event. It was never
published in a peer-reviewed journal and in fact sounds more like
something created by selective quotation than any kind of formal
prophecy.
You've also picked up Tim Lenton's comments about the increased
likelyhood of drought in California. The clip doesn't mention any
dates.
What he may have had in mind might be this
http://web.mit.edu/newsoffice/2011/drought-risk-climate-change.html
"Mean changes in the number of drought months relative to the 20th
century baseline for 2036-2065 along the A1B (moderate emissions)
scenario."
This is actually a prediction, but you are going to have to wait a few
years before you can start claiming that current weather conditions
falsify it.
Bottom line is that you have set up a straw man and you don't know
enough about what you are talking about to realise it.
--
Bill Sloman, Nijmegen
Bill Sloman
> John Larkin wrote:
> > On Fri, 20 May 2011 11:38:16 -0400, Ecnerwal
> > <MyNameForw...@ReplaceWithMyVices.Com.invalid> wrote:
>
> [...]
>
>
>
> >> All the rain/snow which has made the Mississippi exceedingly full also
> >> didn't fall in California, so it didn't happen on your "globe", either.
> >> Got it. Or your warped view is that global warming somehow means
> >> automatically drier, despite the fact that warm air carries more
> >> moisture from the ocean and dumps it in the mountains around you as more
> >> snow when it rises and cools. Or in Ohio to run down and flood
> >> Louisiana, as the case may be.
>
> > Google "climate change california drought" for lots of hilarious
> > "emotive imagery", like...
>
>
> Quote "As global warming melts the mountain snow that supplies fresh
> water to the densely populated Californian region ..."
>
> ROFL! What are they smoking over there? But wait. East Anglia? Wasn't
> that where ...
But it is likely that what he was actually talking about was this
http://web.mit.edu/newsoffice/2011/drought-risk-climate-change.html
which comes from MIT and involves predictions for the period 2036 to
2065, which is still some way off.
It seems to be close enough to keep John Larkin happy, but he's tends
to stop thinking when he sees an answer he likes.
--
Bill Sloman, Nijmegen
k...@att.bizzzzzzzzzzzz
Oh, that. I thought you were talking about the board-edge stitching. I ran
their numbers on input-to-output. Musta been put together by the same people
who dreamt up AGW. MY board isn't that big and the isolated areas are less
than 10% of what it would take. I'll have stitching caps at the ends of the
packages.
>>> Then one also has to worry about the reverse path. What if some large EM
>>> spike gets in?
>>
>> Not *too* concerned about that. I'm more worried about the huge current
>> switching at 180MHz and its effect in the test chamber. ;-)
>>
>
>Wait until them thar flashin' shows up in them clouds and the computer
>gets all hairbrained :-)
Computer? No computer in this box. ;-)
>But it take an incredible field-strength to upset these couplers.
It's a gazinta-to-twelve-gazouta at 201Hz. ;-)
>>> No way to use optocouplers instead? It'll be cheaper, too.
>>
>> No. What powers the optocoupler? I really need an isolated 422 interface.
>> Lots of them.
>
>
>I have mostly powered them via little transformers driven with a few
>100kHz tapped off somewhere. Additional motivators were the requirements
>of 2nd source and low cost.
I need around 50mA for each 422 driver. I could use individual isolated
supplies but then I have 100KHz, and up, that I need to filter out. Second
source, as long as #1 is ADI, isn't important. Cost, for this box, isn't a
biggie, either. If we make a hundred of them, I'll be surprised.
k...@att.bizzzzzzzzzzzz
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
Haven't built it yet (just started in layout this afternoon), so no failing
EMI testing is not my problem. Yet. ;-) I am worried about it, though.
>We use tons of these couplers, but we don't officially EMI test our
>VME modules, since they are just boards that go into racks.
>
>We sometimes wave a spectrum analyzer around a board and see what's
>radiating. Things are getting scary. We buy some "programmable"
>crystal oscillators that have an internal PLL that runs at a couple of
>GHz, and they radiate a lot. The ADI couplers are indeed noisy. A 270
>MHz ARM, or an FPGA with a 1 GHz serdes PLL, are probably bad; we'll
>be doing that combo soon, and I'll snoop it for fun.
>
>We sometimes worry that such stuff will get into opamps and get
>rectified into noise and weird offsets, but haven't seen that as a
>problem so far. The RF seems very local to the chips, usually.
>
>If you're failing EMI, you might try using the Silabs isolators. Or
>the TI parts. I think both have drop-ins for the ADUM parts.
I was more interested in your take on the theories and conclusions about EMI
in app note. Is the board edge fence useful? As I mentioned to Joerg in
another post, the numbers for the buried input-to-output cap just don't make
sense. I figure a 390pF cap on both ends should work (360MHz). A 0603 won't,
though.
Which brings up a related question... Does anyone have a formula for board
spacing vs. isolation voltage? There are 0805 500V caps.
Bill Sloman
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Fri, 20 May 2011 14:07:59 -0700, Joerg <inva...@invalid.invalid>
At which point one starts thinking about wrapping the coax around a
toroid a couple of times, creating a transmission line transformer and
a common mode choke. If you want to block 50/60Hz, that's a biggish
iron or mu-metal toroid ...
Transmission line transformers work fine with soft-iron cores.
--
Bill Sloman, Nijmegen
Joerg
A clients once measured tens of amps. I told them to hold a current
clamp around it and that resulted in some "Oh s..t!" exclamations.
All it takes is a minor difference in chassis ground potential and a
very low resistance shield.
Jim Thompson
wrote:
I have measured ~50VAC difference between the "grounds" of two
adjacent high-rise buildings :-(
...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Joerg
> On Fri, 20 May 2011 07:11:41 -0700, Joerg <inv...@invalid.invalid> wrote:
>
>> k...@att.bizzzzzzzzzzzz wrote:
>>> On Thu, 19 May 2011 17:22:26 -0700, Joerg <inv...@invalid.invalid> wrote:
[...]
>>>> No way to use optocouplers instead? It'll be cheaper, too.
>>> No. What powers the optocoupler? I really need an isolated 422 interface.
>>> Lots of them.
>>
>> I have mostly powered them via little transformers driven with a few
>> 100kHz tapped off somewhere. Additional motivators were the requirements
>> of 2nd source and low cost.
>
> I need around 50mA for each 422 driver. I could use individual isolated
> supplies but then I have 100KHz, and up, that I need to filter out. Second
> source, as long as #1 is ADI, isn't important. Cost, for this box, isn't a
> biggie, either. If we make a hundred of them, I'll be surprised.
50mA is not a problem. I never had to filter much there, sailed through
EMC no sweat. Good transformers don't radiate and then there's a
smoothing choke right after the rectifiers.
But if you make only 100/year it won't matter.
Joerg
[...]
> I was more interested in your take on the theories and conclusions about EMI
> in app note. Is the board edge fence useful? As I mentioned to Joerg in
> another post, the numbers for the buried input-to-output cap just don't make
> sense. I figure a 390pF cap on both ends should work (360MHz). A 0603 won't,
> though.
>
> Which brings up a related question... Does anyone have a formula for board
> spacing vs. isolation voltage? There are 0805 500V caps.
A formula usually isn't accepted by notified bodies. You'll have to get
the standard that applies to your market and then look up creepage and
clearance. I only have the standards for medical and aerospace.
For example, in the med standard basic insulation between parts of
opposite polarity for 500V operation would be 4mm for creepage,
basic/supplemental is 8mm and double or reinforced is 16mm.
You'll have to get special parts that are longer. And don't forget the
hi-ohms bleeder resistor which is usually required so no static can
accumulate up. Vishay has fairly decent prices for those.
k...@att.bizzzzzzzzzzzz
We had some LEFs (Light Emitting Ferrites) on one of our first boards. The
only way to get that much current in the circuit is if the grounds were
significantly different. We never did find the problem.
Rich Grise
> On Fri, 20 May 2011 13:48:28 -0700, Rich Grise
>>> On Fri, 20 May 2011 10:15:15 -0700, Rich Grise
>>>>
>>>>But I have no idea how that relates to shaking sticks out of a garden
>>>>hose.
>>>
>>> When you shake the hose, the twigs shake too.
>>>
>>> When a shield is ungrounded, and there's noise induced onto the
>>> shield, the wires inside will have the same noise. Shake the shield
>>> and the wires inside shake too.
>>>
>>Ah, there's the rub. Since only one end of the shield is connected,
>>there's no current path, so how can any signal be induced, except at, say,
>>1/4 wave? I think I remember hearing (reading) that that's the rationale
>>for
>>leaving one end flapping in the breeze in the first place.
>
> If the shield is not very short, it has serious inductance back to
> ground. And there are a zillion frequencies where the impedance spikes
> up. Not just 1/4 wave, but 5/4, 9/4, etc. And nearby frequencies where
> Z is high but not infinite. Every motor controller, light dimmer,
> relay contact, radio station, cell phone, ESD zap, and lightning bolt
> will shake the ungrounded end. And the wires inside.
>
> Take a long cable. Ground the shield a distance away, in another room
> or another building.
OK, I'll bow out now - I was thinking of interconnects inside one box,
like no longer than a few inches or a foot or two.
Thanks,
Rich
k...@att.bizzzzzzzzzzzz
>k...@att.bizzzzzzzzzzzz wrote:
>> On Fri, 20 May 2011 07:11:41 -0700, Joerg <inv...@invalid.invalid> wrote:
>>
>>> k...@att.bizzzzzzzzzzzz wrote:
>>>> On Thu, 19 May 2011 17:22:26 -0700, Joerg <inv...@invalid.invalid> wrote:
>
>[...]
>
>>>>> No way to use optocouplers instead? It'll be cheaper, too.
>>>> No. What powers the optocoupler? I really need an isolated 422 interface.
>>>> Lots of them.
>>>
>>> I have mostly powered them via little transformers driven with a few
>>> 100kHz tapped off somewhere. Additional motivators were the requirements
>>> of 2nd source and low cost.
>>
>> I need around 50mA for each 422 driver. I could use individual isolated
>> supplies but then I have 100KHz, and up, that I need to filter out. Second
>> source, as long as #1 is ADI, isn't important. Cost, for this box, isn't a
>> biggie, either. If we make a hundred of them, I'll be surprised.
>
>
>50mA is not a problem. I never had to filter much there, sailed through
>EMC no sweat. Good transformers don't radiate and then there's a
>smoothing choke right after the rectifiers.
>
>But if you make only 100/year it won't matter.
Well, it's 100 * 13 per board. ...and likely not "per year". ;-)
k...@att.bizzzzzzzzzzzz
>k...@att.bizzzzzzzzzzzz wrote:
>
>[...]
>
>> I was more interested in your take on the theories and conclusions about EMI
>> in app note. Is the board edge fence useful? As I mentioned to Joerg in
>> another post, the numbers for the buried input-to-output cap just don't make
>> sense. I figure a 390pF cap on both ends should work (360MHz). A 0603 won't,
>> though.
>>
>> Which brings up a related question... Does anyone have a formula for board
>> spacing vs. isolation voltage? There are 0805 500V caps.
>
>
>A formula usually isn't accepted by notified bodies. You'll have to get
>the standard that applies to your market and then look up creepage and
>clearance. I only have the standards for medical and aerospace.
I doubt there is any standard in our market.
>For example, in the med standard basic insulation between parts of
>opposite polarity for 500V operation would be 4mm for creepage,
>basic/supplemental is 8mm and double or reinforced is 16mm.
Opposite polarity? So that's double the V-G voltage? 16mm kinda lets out
0805 capacitors. ;-)
>You'll have to get special parts that are longer. And don't forget the
>hi-ohms bleeder resistor which is usually required so no static can
>accumulate up. Vishay has fairly decent prices for those.
I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
though I'd always wondered about that.
Joerg
1300 still falls under "miniscule quantity" :-)
However, if this is something that repeats itself as sort of an IP block
in other products then the more discrete solution could make sense.
k...@att.bizzzzzzzzzzzz
It does, sorta. This is the simplest version. This application is only
unidirectional (one in, twelve out). The others either require it to supply
or accept power sorta like POE. We're looking for a discrete solution for
that box but it'll be far too complicated, and large, for this box.
Joerg
> On Fri, 20 May 2011 18:26:58 -0700, Joerg <inv...@invalid.invalid> wrote:
>
>> k...@att.bizzzzzzzzzzzz wrote:
>>
>> [...]
>>
>>> I was more interested in your take on the theories and conclusions about EMI
>>> in app note. Is the board edge fence useful? As I mentioned to Joerg in
>>> another post, the numbers for the buried input-to-output cap just don't make
>>> sense. I figure a 390pF cap on both ends should work (360MHz). A 0603 won't,
>>> though.
>>>
>>> Which brings up a related question... Does anyone have a formula for board
>>> spacing vs. isolation voltage? There are 0805 500V caps.
>>
>> A formula usually isn't accepted by notified bodies. You'll have to get
>> the standard that applies to your market and then look up creepage and
>> clearance. I only have the standards for medical and aerospace.
>
> I doubt there is any standard in our market.
>
I don't know what your market is but there's a standard for pretty much
everything. Medical, industrial, residential, aircraft, spacecraft,
extraterrestrial restaurants, UFOs ...
>> For example, in the med standard basic insulation between parts of
>> opposite polarity for 500V operation would be 4mm for creepage,
>> basic/supplemental is 8mm and double or reinforced is 16mm.
>
> Opposite polarity? So that's double the V-G voltage? 16mm kinda lets out
> 0805 capacitors. ;-)
>
No, that is standards committee bureaucrat speak for voltage difference.
Reversing polarity would be ok, too.
>> You'll have to get special parts that are longer. And don't forget the
>> hi-ohms bleeder resistor which is usually required so no static can
>> accumulate up. Vishay has fairly decent prices for those.
>
> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
> though I'd always wondered about that.
If you find that your barrier needs to be wider than an 1812 or
20-something cap can bridge I'd use leaded parts. Larger ceramic caps
can suffer stress fractures over time.
John Larkin
<k...@att.bizzzzzzzzzzzz> wrote:
I'm skeptical of the fence thing. Looking at the ege of a pc board,
any dipole made from copper layers is tiny compared to the 1-meter
wavelength at 300 MHz.
They seem to be comparing a 2-layer board to a multilayer. Not fair.
I'd guess that the major emitter is antennas caused by traces, not
board-edge emission. What I'd do is make sure the power to the
isolators is well bypassed close to the chips. The dipole effect,
their fig 3, is probably serious. The stitching caps should help, if
you can tolerate the side effects.
Ferrite beads in the signal leads that leave the board can work very
well. If you're worried about EMI, I'd do that.
We have one board with 12 of the ADUM things and 12 DC/DC
converters... 12 isolated 4-20 mA channels. The DC/DC sips caused a
lot more trouble at 60 KHz than the isolators at 300 MHz.
This is all fuzzy stuff. Lots of opinions.
John
k...@att.bizzzzzzzzzzzz
>k...@att.bizzzzzzzzzzzz wrote:
>> On Fri, 20 May 2011 18:26:58 -0700, Joerg <inv...@invalid.invalid> wrote:
>>
>>> k...@att.bizzzzzzzzzzzz wrote:
>>>
>>> [...]
>>>
>>>> I was more interested in your take on the theories and conclusions about EMI
>>>> in app note. Is the board edge fence useful? As I mentioned to Joerg in
>>>> another post, the numbers for the buried input-to-output cap just don't make
>>>> sense. I figure a 390pF cap on both ends should work (360MHz). A 0603 won't,
>>>> though.
>>>>
>>>> Which brings up a related question... Does anyone have a formula for board
>>>> spacing vs. isolation voltage? There are 0805 500V caps.
>>>
>>> A formula usually isn't accepted by notified bodies. You'll have to get
>>> the standard that applies to your market and then look up creepage and
>>> clearance. I only have the standards for medical and aerospace.
>>
>> I doubt there is any standard in our market.
>>
>
>I don't know what your market is but there's a standard for pretty much
>everything. Medical, industrial, residential, aircraft, spacecraft,
>extraterrestrial restaurants, UFOs ...
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
How did you know? ;-)
I have never seen a real "standard" for anything.
>>> For example, in the med standard basic insulation between parts of
>>> opposite polarity for 500V operation would be 4mm for creepage,
>>> basic/supplemental is 8mm and double or reinforced is 16mm.
>>
>> Opposite polarity? So that's double the V-G voltage? 16mm kinda lets out
>> 0805 capacitors. ;-)
>>
>
>No, that is standards committee bureaucrat speak for voltage difference.
>Reversing polarity would be ok, too.
Gotcha. Clarity in standards-speak has never been a priority (e.g. "grounded
conductor" and "grounding conductor").
>>> You'll have to get special parts that are longer. And don't forget the
>>> hi-ohms bleeder resistor which is usually required so no static can
>>> accumulate up. Vishay has fairly decent prices for those.
>>
>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>> though I'd always wondered about that.
>
>
>If you find that your barrier needs to be wider than an 1812 or
>20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>can suffer stress fractures over time.
Leaded parts would seriously piss off the powers-that-be.
k...@att.bizzzzzzzzzzzz
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
I was skeptical, too, but the polarity of the emissions suggest that's where
it comes from. I should have (just) enough space to do it, so I'll try.
>They seem to be comparing a 2-layer board to a multilayer. Not fair.
They also compared a more "standard" stackup to a tightly coupled G-V pair.
>I'd guess that the major emitter is antennas caused by traces, not
>board-edge emission. What I'd do is make sure the power to the
>isolators is well bypassed close to the chips. The dipole effect,
>their fig 3, is probably serious. The stitching caps should help, if
>you can tolerate the side effects.
Side effects of the intrusion into the no-man's-land?
>Ferrite beads in the signal leads that leave the board can work very
>well. If you're worried about EMI, I'd do that.
Yes, I've already put some 2.5K ferrites in the signal (and ground lines).
There are a couple of other lines that already have 1K and 10K resistors in
them, so I figured a ferrite was wasted. I also have a 390pF cap on the
outbound side of the ferrite. That's something like .1ohm at 360MHz.
>We have one board with 12 of the ADUM things and 12 DC/DC
>converters... 12 isolated 4-20 mA channels. The DC/DC sips caused a
>lot more trouble at 60 KHz than the isolators at 300 MHz.
The problem is that these ADuM things have a power regulator built in. One,
though I've decided not to use it for other reasons, can supply 100mA. It
draws 290mA (34% efficiency) and has switching currents of .5A @ 180MHz!
>
>This is all fuzzy stuff. Lots of opinions.
Exactly. The boss has already acknowledged an extra spin in the board, though
not because of anticipated EMI difficulty. Any port in a storm, though. ;-)
John Larkin
<k...@att.bizzzzzzzzzzzz> wrote:
Yikes. We have done several boards where we made our own isolated
supplies, using cheap ISDN-type transformers, driven with 60 KHz
square waves with softened edges. That's pretty quiet, and certainly
more efficient than 34%.
I wonder how much of that 66% loss is heat, and how much is radiated
RF!
John
Joerg
> On Sat, 21 May 2011 07:36:26 -0700, Joerg <inv...@invalid.invalid> wrote:
>
>> k...@att.bizzzzzzzzzzzz wrote:
>>> On Fri, 20 May 2011 18:26:58 -0700, Joerg <inv...@invalid.invalid> wrote:
>>>
>>>> k...@att.bizzzzzzzzzzzz wrote:
>>>>
>>>> [...]
>>>>
>>>>> I was more interested in your take on the theories and conclusions about EMI
>>>>> in app note. Is the board edge fence useful? As I mentioned to Joerg in
>>>>> another post, the numbers for the buried input-to-output cap just don't make
>>>>> sense. I figure a 390pF cap on both ends should work (360MHz). A 0603 won't,
>>>>> though.
>>>>>
>>>>> Which brings up a related question... Does anyone have a formula for board
>>>>> spacing vs. isolation voltage? There are 0805 500V caps.
>>>> A formula usually isn't accepted by notified bodies. You'll have to get
>>>> the standard that applies to your market and then look up creepage and
>>>> clearance. I only have the standards for medical and aerospace.
>>> I doubt there is any standard in our market.
>>>
>> I don't know what your market is but there's a standard for pretty much
>> everything. Medical, industrial, residential, aircraft, spacecraft,
>> extraterrestrial restaurants, UFOs ...
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> How did you know? ;-)
>
> I have never seen a real "standard" for anything.
>
Actually the ones for aerospace are pretty good. Very down to earth,
easy to understand. Meaning they were obviously written by real engineers.
Believe me, there is a standard for just about anything that has
civilian, military, medical or commercial use.
[...]
>
>>>> You'll have to get special parts that are longer. And don't forget the
>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>> accumulate up. Vishay has fairly decent prices for those.
>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>> though I'd always wondered about that.
>>
>> If you find that your barrier needs to be wider than an 1812 or
>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>> can suffer stress fractures over time.
>
> Leaded parts would seriously piss off the powers-that-be.
Field failures down the raod probably even more so. Although a
micro-fractured cap across the barrier probably won't be noticed. For a
while ...
Joerg
> On Sat, 21 May 2011 10:21:22 -0500, "k...@att.bizzzzzzzzzzzz"
> <k...@att.bizzzzzzzzzzzz> wrote:
[...]
>> The problem is that these ADuM things have a power regulator built in. One,
>> though I've decided not to use it for other reasons, can supply 100mA. It
>> draws 290mA (34% efficiency) and has switching currents of .5A @ 180MHz!
>
> Yikes. We have done several boards where we made our own isolated
> supplies, using cheap ISDN-type transformers, driven with 60 KHz
> square waves with softened edges. That's pretty quiet, and certainly
> more efficient than 34%.
>
Do you have a favorite cheap XFMR type there that has 2nd source? In the
past I've used European ones at times. ISDN never caught on too much in
the US. The XFMRs are cheap alright but I've had purchasers complaining
that the mfgs can sometimes be "difficult to deal with".
> I wonder how much of that 66% loss is heat, and how much is radiated
> RF!
>
Probably mostly switching losses. Even if you got the transitions down
to under 2nsec that's not very good at 180MHz.
John Larkin
wrote:
>John Larkin wrote:
>> On Sat, 21 May 2011 10:21:22 -0500, "k...@att.bizzzzzzzzzzzz"
>> <k...@att.bizzzzzzzzzzzz> wrote:
>
>[...]
>
>>> The problem is that these ADuM things have a power regulator built in. One,
>>> though I've decided not to use it for other reasons, can supply 100mA. It
>>> draws 290mA (34% efficiency) and has switching currents of .5A @ 180MHz!
>>
>> Yikes. We have done several boards where we made our own isolated
>> supplies, using cheap ISDN-type transformers, driven with 60 KHz
>> square waves with softened edges. That's pretty quiet, and certainly
>> more efficient than 34%.
>>
>
>Do you have a favorite cheap XFMR type there that has 2nd source? In the
>past I've used European ones at times. ISDN never caught on too much in
>the US. The XFMRs are cheap alright but I've had purchasers complaining
>that the mfgs can sometimes be "difficult to deal with".
They are sometimes long-lead. I think ISDN was a mediocre idea,
lacking insight as to data rates to come.
The Talemas are 400 piece min, 16 weeks lately.
We buy custom ISDN-drop-in transformers from Minntronics. They wound
us some with less inter-winding capacitance then the usual ones, and
we can always get them.
Minntronics is great. Ask for Butch.
John
Joerg
> On Sat, 21 May 2011 11:42:05 -0700, Joerg <inv...@invalid.invalid>
> wrote:
>
>> John Larkin wrote:
>>> On Sat, 21 May 2011 10:21:22 -0500, "k...@att.bizzzzzzzzzzzz"
>>> <k...@att.bizzzzzzzzzzzz> wrote:
>> [...]
>>
>>>> The problem is that these ADuM things have a power regulator built in. One,
>>>> though I've decided not to use it for other reasons, can supply 100mA. It
>>>> draws 290mA (34% efficiency) and has switching currents of .5A @ 180MHz!
>>> Yikes. We have done several boards where we made our own isolated
>>> supplies, using cheap ISDN-type transformers, driven with 60 KHz
>>> square waves with softened edges. That's pretty quiet, and certainly
>>> more efficient than 34%.
>>>
>> Do you have a favorite cheap XFMR type there that has 2nd source? In the
>> past I've used European ones at times. ISDN never caught on too much in
>> the US. The XFMRs are cheap alright but I've had purchasers complaining
>> that the mfgs can sometimes be "difficult to deal with".
>
> They are sometimes long-lead. I think ISDN was a mediocre idea,
> lacking insight as to data rates to come.
>
> The Talemas are 400 piece min, 16 weeks lately.
>
400 would be ok, 16wks not so much. Tamura is usually not in stock and
Pulse wants a lot of money for theirs. Although, they do have 8-packs
which you can get for around $4 in qties:
http://ww2.pulseeng.com/products/datasheets/T622.pdf
> We buy custom ISDN-drop-in transformers from Minntronics. They wound
> us some with less inter-winding capacitance then the usual ones, and
> we can always get them.
>
> Minntronics is great. Ask for Butch.
>
Thanks. Aren't they the folks who packed homebaked cookies with some
shipments?
John Larkin
wrote:
Yeah, sometimes. They bake a batch of some sort of Swedish sugar
cookies now and then.
John
k...@att.bizzzzzzzzzzzz
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).
>Believe me, there is a standard for just about anything that has
>civilian, military, medical or commercial use.
That's the nice thing about standards; there are so many to choose from.
>
>>
>>>>> You'll have to get special parts that are longer. And don't forget the
>>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>>> accumulate up. Vishay has fairly decent prices for those.
>>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>>> though I'd always wondered about that.
>>>
>>> If you find that your barrier needs to be wider than an 1812 or
>>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>>> can suffer stress fractures over time.
>>
>> Leaded parts would seriously piss off the powers-that-be.
>
>
>Field failures down the raod probably even more so. Although a
>micro-fractured cap across the barrier probably won't be noticed. For a
>while ...
Likely never. They *do* know about TH parts, though. I'd think the lead
inductance of a TH part would be much worse, too.
k...@att.bizzzzzzzzzzzz
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
For this board I'm also in a big space crunch. I need to fit an isolated
driver channel into about .6"x1".
>I wonder how much of that 66% loss is heat, and how much is radiated
>RF!
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.
Joerg
> On Sat, 21 May 2011 11:15:52 -0700, Joerg <inv...@invalid.invalid> wrote:
>
>> k...@att.bizzzzzzzzzzzz wrote:
[...]
>>> I have never seen a real "standard" for anything.
>>>
>> Actually the ones for aerospace are pretty good. Very down to earth,
>> easy to understand. Meaning they were obviously written by real engineers.
>
> 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).
>
There you go, that's exactly what I meant. CE/UL are the usual standards
in the civilian world.
>> Believe me, there is a standard for just about anything that has
>> civilian, military, medical or commercial use.
>
> That's the nice thing about standards; there are so many to choose from.
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.
>>>>>> You'll have to get special parts that are longer. And don't forget the
>>>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>>>> accumulate up. Vishay has fairly decent prices for those.
>>>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>>>> though I'd always wondered about that.
>>>> If you find that your barrier needs to be wider than an 1812 or
>>>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>>>> can suffer stress fractures over time.
>>> Leaded parts would seriously piss off the powers-that-be.
>>
>> Field failures down the raod probably even more so. Although a
>> micro-fractured cap across the barrier probably won't be noticed. For a
>> while ...
>
> Likely never. They *do* know about TH parts, though. I'd think the lead
> inductance of a TH part would be much worse, too.
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).
Tom Del Rosso
> Tom Del Rosso wrote:
> > John Larkin wrote:
> > >
> > > I tried to give them an analogy they could understand:
> > >
> > > Imagine the end of a garden hose. Imagine a bunch of twigs
> > > sticking out the end. Grab the end of the hose and shake it. What
> > > happens to the twigs?
> >
> > I think they fall out, but how is that an analogy?
> >
> I'm wondering the same thing. If it were my call, I'd connect the
> shield(s) to ground at the point that's closest to the hub of the
> star. Or maybe closest to the signal source. Or toss a coin. ;-P (I
> wonder if there's
> a rule of thumb?)
>
> But I have no idea how that relates to shaking sticks out of a garden
> hose.
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.
k...@att.bizzzzzzzzzzzz
>k...@att.bizzzzzzzzzzzz wrote:
>> On Sat, 21 May 2011 11:15:52 -0700, Joerg <inv...@invalid.invalid> wrote:
>>
>>> k...@att.bizzzzzzzzzzzz wrote:
>
>
>[...]
>
>>>> I have never seen a real "standard" for anything.
>>>>
>>> Actually the ones for aerospace are pretty good. Very down to earth,
>>> easy to understand. Meaning they were obviously written by real engineers.
>>
>> 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).
>>
>
>There you go, that's exactly what I meant. CE/UL are the usual standards
>in the civilian world.
I thought you meant standards for the market. There are none and few of our
products go through UL or any like certification.
>>> Believe me, there is a standard for just about anything that has
>>> civilian, military, medical or commercial use.
>>
>> That's the nice thing about standards; there are so many to choose from.
>
>
>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.
We don't do medical or aviation. There are no such standards.
>>>>>>> You'll have to get special parts that are longer. And don't forget the
>>>>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>>>>> accumulate up. Vishay has fairly decent prices for those.
>>>>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>>>>> though I'd always wondered about that.
>>>>> If you find that your barrier needs to be wider than an 1812 or
>>>>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>>>>> can suffer stress fractures over time.
>>>> Leaded parts would seriously piss off the powers-that-be.
>>>
>>> Field failures down the raod probably even more so. Although a
>>> micro-fractured cap across the barrier probably won't be noticed. For a
>>> while ...
>>
>> Likely never. They *do* know about TH parts, though. I'd think the lead
>> inductance of a TH part would be much worse, too.
>
>
>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 ;-)
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).
>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).
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).
Joerg
> On Sun, 22 May 2011 08:00:44 -0700, Joerg <inv...@invalid.invalid> wrote:
>
>> k...@att.bizzzzzzzzzzzz wrote:
>>> On Sat, 21 May 2011 11:15:52 -0700, Joerg <inv...@invalid.invalid> wrote:
>>>
>>>> k...@att.bizzzzzzzzzzzz wrote:
>>
>> [...]
>>
>>>>> I have never seen a real "standard" for anything.
>>>>>
>>>> Actually the ones for aerospace are pretty good. Very down to earth,
>>>> easy to understand. Meaning they were obviously written by real engineers.
>>> 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).
>>>
>> There you go, that's exactly what I meant. CE/UL are the usual standards
>> in the civilian world.
>
> I thought you meant standards for the market. There are none and few of our
> products go through UL or any like certification.
>
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.
>>>> Believe me, there is a standard for just about anything that has
>>>> civilian, military, medical or commercial use.
>>> That's the nice thing about standards; there are so many to choose from.
>>
>> 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.
>
> We don't do medical or aviation. There are no such standards.
>
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.
>>>>>>>> You'll have to get special parts that are longer. And don't forget the
>>>>>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>>>>>> accumulate up. Vishay has fairly decent prices for those.
>>>>>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>>>>>> though I'd always wondered about that.
>>>>>> If you find that your barrier needs to be wider than an 1812 or
>>>>>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>>>>>> can suffer stress fractures over time.
>>>>> Leaded parts would seriously piss off the powers-that-be.
>>>> Field failures down the raod probably even more so. Although a
>>>> micro-fractured cap across the barrier probably won't be noticed. For a
>>>> while ...
>>> Likely never. They *do* know about TH parts, though. I'd think the lead
>>> inductance of a TH part would be much worse, too.
>>
>> 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 ;-)
>
> 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).
>
IME you almost have to migrate from board level to system level EMI
mitigation once your concerns are spectra above half a gigeehoitz.
>> 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).
>
> 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).
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.
k...@att.bizzzzzzzzzzzz
>k...@att.bizzzzzzzzzzzz wrote:
>> On Sun, 22 May 2011 08:00:44 -0700, Joerg <inv...@invalid.invalid> wrote:
>>
>>> k...@att.bizzzzzzzzzzzz wrote:
>>>> On Sat, 21 May 2011 11:15:52 -0700, Joerg <inv...@invalid.invalid> wrote:
>>>>
>>>>> k...@att.bizzzzzzzzzzzz wrote:
>>>
>>> [...]
>>>
>>>>>> I have never seen a real "standard" for anything.
>>>>>>
>>>>> Actually the ones for aerospace are pretty good. Very down to earth,
>>>>> easy to understand. Meaning they were obviously written by real engineers.
>>>> 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).
>>>>
>>> There you go, that's exactly what I meant. CE/UL are the usual standards
>>> in the civilian world.
>>
>> I thought you meant standards for the market. There are none and few of our
>> products go through UL or any like certification.
>>
>
>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 ;-)
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.
>Then there's the other standards such as signal levels. Or the RS422 you
>just mentioned, because that's also a standard.
It's RS-422 because that's the standard we chose to use. It wasn't selected
by someone else. We "own" both ends.
>>>>> Believe me, there is a standard for just about anything that has
>>>>> civilian, military, medical or commercial use.
>>>> That's the nice thing about standards; there are so many to choose from.
>>>
>>> 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.
>>
>> We don't do medical or aviation. There are no such standards.
>>
>
>But you just mentioned UL/CE which is where you'll find (or already
>found) the answer to your creepage distance question.
For the voltage protection *we* decide is useful. There is no standard for
such.
>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.
On the mains side UL says something about this, sure. On the other end of the
box, nope.
>>>>>>>>> You'll have to get special parts that are longer. And don't forget the
>>>>>>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>>>>>>> accumulate up. Vishay has fairly decent prices for those.
>>>>>>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>>>>>>> though I'd always wondered about that.
>>>>>>> If you find that your barrier needs to be wider than an 1812 or
>>>>>>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>>>>>>> can suffer stress fractures over time.
>>>>>> Leaded parts would seriously piss off the powers-that-be.
>>>>> Field failures down the raod probably even more so. Although a
>>>>> micro-fractured cap across the barrier probably won't be noticed. For a
>>>>> while ...
>>>> Likely never. They *do* know about TH parts, though. I'd think the lead
>>>> inductance of a TH part would be much worse, too.
>>>
>>> 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 ;-)
>>
>> 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).
>>
>
>IME you almost have to migrate from board level to system level EMI
>mitigation once your concerns are spectra above half a gigeehoitz.
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.
>>> 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).
>>
>> 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).
>
>
>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.
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.
Joerg
Then you are ok, as long as your product specs don't claim a certain
breakdown immunity.
[...]
>>>>>>>>>> You'll have to get special parts that are longer. And don't forget the
>>>>>>>>>> hi-ohms bleeder resistor which is usually required so no static can
>>>>>>>>>> accumulate up. Vishay has fairly decent prices for those.
>>>>>>>>> I'll have a look at Vishay. Thanks. None of our stuff has bleeders now,
>>>>>>>>> though I'd always wondered about that.
>>>>>>>> If you find that your barrier needs to be wider than an 1812 or
>>>>>>>> 20-something cap can bridge I'd use leaded parts. Larger ceramic caps
>>>>>>>> can suffer stress fractures over time.
>>>>>>> Leaded parts would seriously piss off the powers-that-be.
>>>>>> Field failures down the raod probably even more so. Although a
>>>>>> micro-fractured cap across the barrier probably won't be noticed. For a
>>>>>> while ...
>>>>> Likely never. They *do* know about TH parts, though. I'd think the lead
>>>>> inductance of a TH part would be much worse, too.
>>>> 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 ;-)
>>> 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).
>>>
>> IME you almost have to migrate from board level to system level EMI
>> mitigation once your concerns are spectra above half a gigeehoitz.
>
> 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.
>
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.
>>>> 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).
>>> 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).
>>
>> 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.
>
> 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.
>
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.
Jeroen Belleman
Anyway, this means that *someone else* is dumping large amounts
of current into ground, where he should have used a proper
return conductor. It's definitely unneighbourly.
Jeroen Belleman
Joerg
> k...@att.bizzzzzzzzzzzz wrote:
>> On Fri, 20 May 2011 18:11:25 -0700, Joerg <inv...@invalid.invalid>
>> wrote:
>>
>>> John Larkin wrote:
>>>> On Fri, 20 May 2011 14:07:59 -0700, Joerg <inv...@invalid.invalid>
>>>> wrote:
[...]
>>>> But sometimes grounding at both ends is not
>>>>> advised for other reasons, such as erroneous large currents. "Dude,
>>>>> don't that smell like pee vee cee around here?"
>>>> That would be unusual, two grounds with enough voltage diference, at a
>>>> low enough impedance, to heat up a cable shield.
>>>>
>>> A clients once measured tens of amps. I told them to hold a current
>>> clamp around it and that resulted in some "Oh s..t!" exclamations.
>>>
>>> All it takes is a minor difference in chassis ground potential and a
>>> very low resistance shield.
>>
>> We had some LEFs (Light Emitting Ferrites) on one of our first
>> boards. The
>> only way to get that much current in the circuit is if the grounds were
>> significantly different. We never did find the problem.
>
> Anyway, this means that *someone else* is dumping large amounts
> of current into ground, where he should have used a proper
> return conductor. It's definitely unneighbourly.
>
An engineer friend summed it up this way when he came back from a long
problem-hunting trip: "Once I saw numerous crushed beer cans and
cigarette butts in the bottom of the electrical cabinet I new the
reason" :-)
John Larkin
wrote:
>Jeroen Belleman wrote:
We installed one system in Nigeria and found much worse stuff in the
bottoms of rack cabinets.
John
k...@att.bizzzzzzzzzzzz
wrote:
That's easy to say, but when the distribution system is in another building,
or worse, a truck (generator) parked out in the parking lot, things can get
quite ugly.
> It's definitely unneighbourly.
10-4 on that.