pls help overcome challenges before passing EMC

[bene-tam]

first of all, thx all for help

pls help verify my design, it's a poor one since it's not designed for
EMC from first day and my supervisor didnt alert too!!

i am having a serious headache about this.

brief description of my hardware : 110/230VAC to 12VDC output adaptor
is used to power my hardware(a switching isolated power supplies
giving "no earth"), at the input of my hardware, i have fuse, relay,
diode and MVO to make sure polarity, current are "safe", the 12VDC
input is converted to +5V&0V&-5V using another onboard
switch-mode-power-supply for my circuits(including ADC, MCU, OpAmp,
LED drivers), there are integrated ethernet chips on the MCU already,
while i am using Maxim ICs for making RS232.

past pain(solved) :
no common mode filter cap for the both 0V&12V input(to chassis)
no filtering cap at the DB-9 connector for RS232, once failed for EMI,
now okay

currently, 2 big challenges are faced and after that i really hope
that i can learn from expert here, i still have tons of question/doubt

***the first challenge is EN61000-4-4(Electrical fast transient/Burst)

my device can work with RJ45 and standalone under the fast transient
test, but once the +-1kV is exerted, the connection with a 56kmodem is
lost.

the questions are:
should i connect the circuit ground to chassis??
how could i "shield" the cable connection with 56kmodem, given there
is no earth on both devices?(both not using 3-pin plug, no true earth)

***the second challenge is EN61000-4-3(Radiated Immunity)

again, how should i increase my device's radiated immunity?
the connection with the modem is lost in such a test condition.
currently, the metal chassis is connected to the circuit ground, and a
cap is used to "bridge" the input 0V and the circuit ground(violating
the isolation of the 12V to +-5V output converter as i think)

thx so much

 

[John Woodgate]

I read in sci.electronics.design that bene-tam <bene-
[email protected]> wrote (in <fe335d4f.0401150235.63ae78fc@posting.
google.com>) about 'pls help overcome challenges before passing EMC', on
Thu, 15 Jan 2004:

currently, 2 big challenges are faced and after that i really hope
that i can learn from expert here, i still have tons of question/doubt

There are not many EMC experts on this NG. You would do better to join
the IEEE EMC-PSTC mailgroup and ask there.
Go to http://www.ieee-pses.org/

***the first challenge is EN61000-4-4(Electrical fast transient/Burst)

my device can work with RJ45 and standalone under the fast transient
test, but once the +-1kV is exerted, the connection with a 56kmodem is
lost.

Which product standards are you applying? EN 55022 and 55024?

Where are you applying the EFT? Check that the standard says that you
*need* to apply it to that port.

the questions are:
should i connect the circuit ground to chassis??
how could i "shield" the cable connection with 56kmodem, given there
is no earth on both devices?(both not using 3-pin plug, no true earth)

'Earth' in the sense of 'protective conductor in the building wiring' is
almost irrelevant for EMC. The wiring has far too much inductance to
present a low impedance at radio frequencies.

Many modern products don't have anything much you could call a
'chassis', but if you have a metal enclosure, it's usually NOT a good
idea not to connect the circuit common to it. if it's left floating, it
can act as an antenna to inject energy unpredictably into circuits.

***the second challenge is EN61000-4-3(Radiated Immunity)

again, how should i increase my device's radiated immunity?
the connection with the modem is lost in such a test condition.
currently, the metal chassis is connected to the circuit ground, and a
cap is used to "bridge" the input 0V and the circuit ground(violating
the isolation of the 12V to +-5V output converter as i think)

Why do you need isolation between the 12 V and +/- 5 V? Your 12 V is
already isolated from the mains, isn't it? Which safety standard are you
applying? EN 60950-1?

 

[Rene Tschaggelar]

first of all, thx all for help

pls help verify my design, it's a poor one since it's not designed for
EMC from first day and my supervisor didnt alert too!!

i am having a serious headache about this.

brief description of my hardware : 110/230VAC to 12VDC output adaptor
is used to power my hardware(a switching isolated power supplies
giving "no earth"), at the input of my hardware, i have fuse, relay,
diode and MVO to make sure polarity, current are "safe", the 12VDC
input is converted to +5V&0V&-5V using another onboard
switch-mode-power-supply for my circuits(including ADC, MCU, OpAmp,
LED drivers), there are integrated ethernet chips on the MCU already,
while i am using Maxim ICs for making RS232.

past pain(solved) :
no common mode filter cap for the both 0V&12V input(to chassis)
no filtering cap at the DB-9 connector for RS232, once failed for EMI,
now okay

currently, 2 big challenges are faced and after that i really hope
that i can learn from expert here, i still have tons of question/doubt

***the first challenge is EN61000-4-4(Electrical fast transient/Burst)

my device can work with RJ45 and standalone under the fast transient
test, but once the +-1kV is exerted, the connection with a 56kmodem is
lost.

Couln't the software just reconnect by software ?
The hardware doesn't latch up ?
There are overvoltage protectors, Eg ZincOxide, Gaz tubes, Z diodes
to cut the pulses.
Then there are series resistor, ferrites and caps to further soften
the pulses.

the questions are:
should i connect the circuit ground to chassis??
how could i "shield" the cable connection with 56kmodem, given there
is no earth on both devices?(both not using 3-pin plug, no true earth)

That depends on the application. Give it a try.
Not having a faraday shield around the electronics means in most times
you have lost already. You can have the faraday DC connected or AC
connected with the electronics. AC connected means basically floating
but has 10nF caps on all sides to the case. Check whether the 50V
compliance of the SMT caps is sufficient.

***the second challenge is EN61000-4-3(Radiated Immunity)

again, how should i increase my device's radiated immunity?
the connection with the modem is lost in such a test condition.
currently, the metal chassis is connected to the circuit ground, and a
cap is used to "bridge" the input 0V and the circuit ground(violating
the isolation of the 12V to +-5V output converter as i think)

You need fast caps, meaning 100nF or so, to GND. All slow lines, eg
Power must have a 100nF against GND. The fast lines, eg communication
lines are to be connected with 1-10nF to GND.

Then, there are common mode chokes for power and signal. But that
all depends on the application.

You should be able to do some measurements yourselves.

Rene

 

[Rene Tschaggelar]

I read in sci.electronics.design that bene-tam <bene-
[email protected]> wrote (in <fe335d4f.0401150235.63ae78fc@posting.
google.com>) about 'pls help overcome challenges before passing EMC', on
Thu, 15 Jan 2004:




There are not many EMC experts on this NG. You would do better to join
the IEEE EMC-PSTC mailgroup and ask there.
Go to http://www.ieee-pses.org/

Not experts, but we're minimally seasoned here too.

Rene

 

[bene-tam]

I read in sci.electronics.design that bene-tam <bene-
[email protected]> wrote (in <fe335d4f.0401150235.63ae78fc@posting.
google.com>) about 'pls help overcome challenges before passing EMC', on
Thu, 15 Jan 2004:


There are not many EMC experts on this NG. You would do better to join
the IEEE EMC-PSTC mailgroup and ask there.
Go to http://www.ieee-pses.org/

Which product standards are you applying? EN 55022 and 55024?

Thx first
it's EN55024
the +-1kV applies to the 110/230VAC power lines, but the supplier of
the 110/230VAC to 12VDC adaptor says its design wont do the job very
well, it has asked us to soften the input again

Where are you applying the EFT? Check that the standard says that you
*need* to apply it to that port.

before applying the EFT to my device, my device has made a connection
with a remote computer thru a pair of modem(all their power supplies
are "clean")
once the EFT is applied, the connection is broken

'Earth' in the sense of 'protective conductor in the building wiring' is
almost irrelevant for EMC. The wiring has far too much inductance to
present a low impedance at radio frequencies.

Many modern products don't have anything much you could call a
'chassis', but if you have a metal enclosure, it's usually NOT a good
idea not to connect the circuit common to it. if it's left floating, it
can act as an antenna to inject energy unpredictably into circuits.

so i should connect the 0V of the output of the 12VDC adaptor to the
metal enclose instead of using the circuit ground!?

Why do you need isolation between the 12 V and +/- 5 V? Your 12 V is
already isolated from the mains, isn't it? Which safety standard are you
applying? EN 60950-1?

i dont intend to do isolate myself
that 12VDC to +/-5VDC converter is a "component" supplied from a
vendor, i try to open the cover and discussed with one of my teacher,
he said it has isolation design, u can see the design try to separate
the primary and secondary so that the noise from either side doesnt
bother the other one.

yes, it's EN60950, already passed

thx so much

 

[John Woodgate]

I read in sci.electronics.design that bene-tam <bene-
[email protected]> wrote (in <fe335d4f.0401151817.21480269@posting.
google.com>) about 'pls help overcome challenges before passing EMC', on
Thu, 15 Jan 2004:

Thx first
it's EN55024
the +-1kV applies to the 110/230VAC power lines, but the supplier of
the 110/230VAC to 12VDC adaptor says its design wont do the job very
well, it has asked us to soften the input again

Then you need to get an adaptor that DOES pass the EFT test. The way the
EMC Directive works, it is effectively the responsibility of the adaptor
manufacturer to meet the requirement.

[snip]


so i should connect the 0V of the output of the 12VDC adaptor to the
metal enclose instead of using the circuit ground!?

You should connect the 0 V to the enclosure and then the circuit ground
to the enclosure.

[snip]

Why do you need isolation between the 12 V and +/- 5 V? Your 12 V is
already isolated from the mains, isn't it? Which safety standard are you
applying? EN 60950-1?

i dont intend to do isolate myself
that 12VDC to +/-5VDC converter is a "component" supplied from a
vendor, i try to open the cover and discussed with one of my teacher,
he said it has isolation design, u can see the design try to separate
the primary and secondary so that the noise from either side doesnt
bother the other one.

Isolation is a safety matter, not an EMC matter. The *adaptor* has to
provide the isolation, otherwise you have serious safety problems.