Searching \ for '[EE] Big holes in PCB causing emission problems?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/pcbs.htm?key=pcb
Search entire site for: 'Big holes in PCB causing emission problems?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Big holes in PCB causing emission problems?'
2007\10\05@233641 by Xiaofan Chen

face picon face
To solve the heat dissipation problems, we put two rectangular holes in the
the PCBs (the shape is now like two flat E-Core put together). It is quite
effective for this purpose. But one US colleague (a power supply guru)
commented that the holes are bad for emission as it acts as a dipole antenna.
The emission is indeed not so good but I am not so sure it is because of
this antenna.

I do not know much about RF. What is your opinion?

Xiaofan

2007\10\06@081718 by Funny NYPD

picon face
Instead two big holes, make two big zones with some small via/holes. Then copper pour these zones to enhance the shielding performance. This will improve heat exchange performance without losing the EMI performance, especially for SMPS (switch mode power supply).

The disadvantage for both designs is that both will create stress to deform the board if locations of holes are not well-balanced.

Funny



{Original Message removed}

2007\10\06@110718 by Vasile Surducan

face picon face
I always liked PIC folklor, PCB folklor and electronic philosoply.
An incomplete question will gent an incomplete answer.
A dipole antenna is emiting a specific frequency. As long you have
ground plane around the hole it can't be an antenna.
There are so many tehniques for dissipating heat from a PCB (starting
with aluminium layers and finishing with heat spreading vias and
copper surfaces without being solder masked) than a rectangular hole
is the the last one should be manufactured just because the price of
manufacturing (if indeed rectangular means recatangular and no other
round edge shape)


Vasile


On 10/5/07, Xiaofan Chen <spam_OUTxiaofancTakeThisOuTspamgmail.com> wrote:
> To solve the heat dissipation problems, we put two rectangular holes in the
> the PCBs (the shape is now like two flat E-Core put together). It is quite
> effective for this purpose. But one US colleague (a power supply guru)
> commented that the holes are bad for emission as it acts as a dipole antenna.
> The emission is indeed not so good but I am not so sure it is because of
> this antenna.
>
> I do not know much about RF. What is your opinion?
>
> Xiaofan
> -

2007\10\06@111952 by Marcel Duchamp

picon face
Xiaofan Chen wrote:
> To solve the heat dissipation problems, we put two rectangular holes in the
> the PCBs (the shape is now like two flat E-Core put together). It is quite
> effective for this purpose. But one US colleague (a power supply guru)
> commented that the holes are bad for emission as it acts as a dipole antenna.
> The emission is indeed not so good but I am not so sure it is because of
> this antenna.
>
> I do not know much about RF. What is your opinion?
>
> Xiaofan

Before testing solutions, try verifying the problem.  Perform emission
tests as is.  Note readings.  Then cover holes with copper tape or other
suitable conductor and re-test.  If the holes are a problem, that should
show it up.  If so, then try several smaller holes.  This could be done
with more copper tape into which you punch some holes.  Or heatsink or ...

2007\10\06@114441 by Kenneth Lumia

picon face
There is not really enough information to accurately respond,
however in general:

1. Slots (rectangular holes) are generally a bad idea.  Not just
for dipoles, but for fringing effects.
2. If indeed a dipole, remember that it can radiate almost any
frequency, it is just more efficient at 1/4 wavelength, 1/2, etc.
3. Remove all doubt.  Take a field probe and spectrum analyzer,
set it to the frequency of interest and run it over the board
toward the slot.  If the output jumps at the edge/over the slot,
you have a problem.

Ken

{Original Message removed}

2007\10\06@124425 by Peter P.

picon face
Vasile Surducan <piclist9 <at> gmail.com> writes:

> I always liked PIC folklor, PCB folklor and electronic philosoply.
> An incomplete question will gent an incomplete answer.
> A dipole antenna is emiting a specific frequency. As long you have
> ground plane around the hole it can't be an antenna.

Of course not. It can be a closed dipole or a slot antenna. The *area* of the
disturbance, its characteristic impedance and its relationship to the
characteristic impedance of free space (and convenient cables^H^H^H^H^Hantennas
connected to the item), and to the wavelength(s) to be radiated determine its
efficiency as a radiator.

Nowadays conduction heat dissipation is the preferred way to get rid of heat
from equipment. That includes conductive foils and conductive (heat) silicone
pads and so on.

Peter P.


2007\10\06@130742 by David VanHorn

picon face
Makes sense to me, especially if high frequency currents that should
be localized with explicit return tracks are being dumped into the
plane instead.

One good diagnostic, is wether attaching a shielded cable, or even a
length of wire, to the ground plane, makes the radiation worse. This
tells you that you have RF currnents flowing in the plane.

In general, I try hard to handle all high frequency returns with
discrete tracks, and only use the plane for low frequency, low
amplitude currents.

I use a lot of planeing, on both sides of my boards, but I just don't
let the fast and loud stuff find it's own way home.. If you think
about the infinite grid of resistors problem, you see that current
dumped into a plane will take every possible path home. Restricting it
to a track prevents it from spreading out and increasing loop area.
I try to keep the return tracks as close as possible to the sourcing
tracks, running them over each other if possible, though that does
tend to chew up the routing possibilities.

With large holes in the plane, and if you have significant RF currents
flowing in the plane, then you're even more likely to get radiation
problems.   If you don't have much going on in the ground plane
though, then it probably won't matter.

2007\10\06@163307 by Dave Tweed

face
flavicon
face
Vasile Surducan <.....piclist9KILLspamspam@spam@gmail.com> wrote:
> As long you have ground plane around the hole it can't be an antenna.

That's just wrong.

A slot makes a *very* effective antenna under certain circumstances.
Just ask anyone who's had to seal a metal box that has a movable lid
at UHF or microwave freqeuncies!

Try googling "slot antenna".

To address the OP's problem, the high-pass cutoff frequency of a hole
in conductive material is related to its largest dimension. Replacing
a large hole with lots of small holes, even if the total area is the
same, will have the desired effect.

-- Dave Tweed

2007\10\06@223703 by David VanHorn

picon face
On 10/6/07, Vasile Surducan <piclist9spamKILLspamgmail.com> wrote:
> I always liked PIC folklor, PCB folklor and electronic philosoply.
> An incomplete question will gent an incomplete answer.
> A dipole antenna is emiting a specific frequency. As long you have
> ground plane around the hole it can't be an antenna.

That's not true.  If current is flowing around the hole, and it's not
evenly distribuited, then there's a net magnetic field, and you're
radiating.

2007\10\07@213455 by Xiaofan Chen

face picon face
On 10/7/07, David VanHorn <.....microbrixKILLspamspam.....gmail.com> wrote:
> On 10/6/07, Vasile Surducan <EraseMEpiclist9spam_OUTspamTakeThisOuTgmail.com> wrote:
> > I always liked PIC folklor, PCB folklor and electronic philosoply.
> > An incomplete question will gent an incomplete answer.
> > A dipole antenna is emiting a specific frequency. As long you have
> > ground plane around the hole it can't be an antenna.
>
> That's not true.  If current is flowing around the hole, and it's not
> evenly distribuited, then there's a net magnetic field, and you're
> radiating.

A bit more detail here. There are indeed DC/DC converters in the
center leg. And indeed the total current flowing around the holes
are not evenly distributed. The top leg has higher current then the
lower leg. So there are radiated emissions from the holes from
your assertion.

The spetrum analyser is not availble right now (I was doing the
CE tests in Cleveland, Ohio for the first prototype). So I can not
do much right now. Once it is available, I will do some tests.
But the RF emission/immunity chambers will not be ready anytime
soon in the Singapore labs (too expensive to set up rigth now).
So the idea is to do evaluation in Singapore and do the qualification
tests in 3rd party labs or in US. It is always easy to do the tests
and troubleshooting in house so I'd prefer to fly to US for testing
even though the long flight (20 hours+) is not that enjoyable.

Xiaofan

2007\10\08@044252 by Alan B. Pearce

face picon face
>But one US colleague (a power supply guru) commented that
>the holes are bad for emission as it acts as a dipole antenna.
>The emission is indeed not so good but I am not so sure
>it is because of this antenna.
>
>I do not know much about RF. What is your opinion?

He could be correct, especially if the holes are surrounded by ground plane.
I would certainly look at putting breaks in the ground plane so you did not
get complete loops around the holes.

Think in terms of a typical VHF or UHF TV aerial. The dipole is a complete
loop, with two wires feeding a break in the loop on one side, and the
opposite side of the loop typically grounded to the support pole. Now you do
not need a loop of wire to act as the dipole. It could be a slot in an
infinite sheet of metal, and this is effectively what you would have with
the PCB. You could calculate the effective frequency of maximum radiation
from the dimensions of the hole you have made in the PCB, and go looking
with a spectrum analyzer with a sniffer loop, and I would almost bet that
you would find a peak at that frequency and harmonics of it, due to any
ground plane currents that are circulating.

It may require further thinking about grounding after breaking the loop, to
stop the resultant ends acting as tuned stubs at a different frequency. This
is all part of the fun of getting such beasts to be 'RF quiet'.

2007\10\08@051724 by Xiaofan Chen

face picon face
On 10/8/07, Alan B. Pearce <A.B.Pearcespamspam_OUTrl.ac.uk> wrote:
> >But one US colleague (a power supply guru) commented that
> >the holes are bad for emission as it acts as a dipole antenna.
> >The emission is indeed not so good but I am not so sure
> >it is because of this antenna.
> >
> >I do not know much about RF. What is your opinion?
>
> He could be correct, especially if the holes are surrounded by ground plane.
> I would certainly look at putting breaks in the ground plane so you did not
> get complete loops around the holes.
>

The ground planes are not continuous as I have isolated I/O channels.
So there are 5 ground planes on two daughter boards and 10
ground plane on the main boards ( 8 isolated channel, one back
plane and one customer power). That is not so good
for immunity but the customer wants isolated I/O channels.

By the way, the PCBs are all 8-layer boards. One of them with
blind and buried vias because of the density and use of
a rather complicated planar transforemer. The planar
transformer was designed by the same US engineer.

Xiaofan

2007\10\08@083353 by David VanHorn

picon face
>
> The spetrum analyser is not availble right now (I was doing the
> CE tests in Cleveland, Ohio for the first prototype).

Cleveland??  I'm in Pittsburgh, just a bit to the east!  :)

2007\10\08@093525 by Xiaofan Chen

face picon face
On 10/8/07, David VanHorn <@spam@microbrixKILLspamspamgmail.com> wrote:
> >
> > The spetrum analyser is not availble right now (I was doing the
> > CE tests in Cleveland, Ohio for the first prototype).
>
> Cleveland??  I'm in Pittsburgh, just a bit to the east!  :)

We went for a half-day visit to Pittsburgh last year
(we were under 3.5 months training in Mayfield Heights,
Cleveland campus) and it seems to be a nice city.

Xiaofan

2007\10\08@121240 by Carl Denk

flavicon
face
Have been happy in Cleveland for 69 years, except 4 years in Ann Arbor,
Mich. Lots of quiet high tech stuff around, near to us. In Elyria,
Ridgid tools has their "Micro SEE SNAKE" to look up to 30 feet into
small diameter pipes and other inaccessible places, and other high tech
tools developed here. Picker Xray, Cleveland Clinic are a few other big
names. We have 4 seasons here with snow skiing (not like Rockies),
boating, fishing, and other water sports. Lots of nice hiking and bike
riding close to home.

Son-in-law with friends were Pittsburgh yesterday for the football game.

Xiaofan Chen wrote:
{Quote hidden}

2007\10\08@121924 by David VanHorn

picon face
> We went for a half-day visit to Pittsburgh last year
> (we were under 3.5 months training in Mayfield Heights,
> Cleveland campus) and it seems to be a nice city.

If you get over this way, or even if you're in cleveland in the near
future, let me know.  We can get together, if you like.

2007\10\08@223455 by David VanHorn

picon face
> Son-in-law with friends were Pittsburgh yesterday for the football game.

I was driving home "the back way" to avoid the traffic.
I live within a mile of the stadium.


It's nice. I haven't been here that long, only 6 months as of the 14th.
Home is Muncie Indiana where my wife is, and/or Dayton OH where my OSO lives.
I don't get to see either very often.  :(

But, it's a good job, and has a lot of upside opportunity.
My boss is running for Mayor as well, which is getting interesting.

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...