Standards sought for immunity of shielded cable links to power-frequency ground loops
First the background:
In some timing distribution applications, the primary source of
interference comes from different ground voltages in different parts of
the facility, such as a ship or a megawatt radar.
The effect of differing ground potentials on a shielded cable is to pull a
large current through the shield, so there is a significant voltage
between the ends of the cable. No matter how good the shield is at RF,
one consequence is that the same power-frequency offset voltage appears on
the conductors within that shield, because the skin depth at 60 Hz vastly
exceeds the thickness of any reasonable shield. Unshielded twisted pair
will suffer the same common-mode offset voltage, perhaps more. This
offset often contains significant harmonics of the power frequency,
nominally up to the seventh harmonic, not just the fundamental.
If the cable is shielded twisted pair, such as twinax, the offset appears
as a common-mode voltage on the two conductors, and (if not too large) is
eliminated by the CMRR of the receiver.
If the cable is coax, the offset voltage appears added to the timing
signal voltage, and if the offset isn't too large the signal receiver will
be sufficiently immune to this conducted EMI.
And now the question:
What standards exist governing required immunity of signal ports to these
ground-loop induced power-frequency (hum) voltages?
All the conducted suseptability standards I've found cover only
frequencies exceeding 10 KHz, not power frequencies and their harmonics.
Thanks,
Joe
During my experiences involving audio/phone, video and data
transmission, we were taught to ground the shield at one end only so we
would not cause a ground loop.
I ran into problems everywhere I went with this and as much as folks
disdain transformers, they are your friend in this type of problem.
Don White Consultants/Interference Control Technology published a whole
series on EMI, Grounding, and EMC for the military. They are located in
Gainesville, VA.
Brian
Joseph M Gwinn wrote:
First the background:
In some timing distribution applications, the primary source of
interference comes from different ground voltages in different parts of
the facility, such as a ship or a megawatt radar.
The effect of differing ground potentials on a shielded cable is to pull a
large current through the shield, so there is a significant voltage
between the ends of the cable. No matter how good the shield is at RF,
one consequence is that the same power-frequency offset voltage appears on
the conductors within that shield, because the skin depth at 60 Hz vastly
exceeds the thickness of any reasonable shield. Unshielded twisted pair
will suffer the same common-mode offset voltage, perhaps more. This
offset often contains significant harmonics of the power frequency,
nominally up to the seventh harmonic, not just the fundamental.
If the cable is shielded twisted pair, such as twinax, the offset appears
as a common-mode voltage on the two conductors, and (if not too large) is
eliminated by the CMRR of the receiver.
If the cable is coax, the offset voltage appears added to the timing
signal voltage, and if the offset isn't too large the signal receiver will
be sufficiently immune to this conducted EMI.
And now the question:
What standards exist governing required immunity of signal ports to these
ground-loop induced power-frequency (hum) voltages?
All the conducted suseptability standards I've found cover only
frequencies exceeding 10 KHz, not power frequencies and their harmonics.
Thanks,
Joe
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Joseph M Gwinn skrev:
First the background:
In some timing distribution applications, the primary source of
interference comes from different ground voltages in different parts of
the facility, such as a ship or a megawatt radar.
The effect of differing ground potentials on a shielded cable is to pull a
large current through the shield, so there is a significant voltage
between the ends of the cable. No matter how good the shield is at RF,
one consequence is that the same power-frequency offset voltage appears on
the conductors within that shield, because the skin depth at 60 Hz vastly
exceeds the thickness of any reasonable shield. Unshielded twisted pair
will suffer the same common-mode offset voltage, perhaps more. This
offset often contains significant harmonics of the power frequency,
nominally up to the seventh harmonic, not just the fundamental.
If the cable is shielded twisted pair, such as twinax, the offset appears
as a common-mode voltage on the two conductors, and (if not too large) is
eliminated by the CMRR of the receiver.
If the cable is coax, the offset voltage appears added to the timing
signal voltage, and if the offset isn't too large the signal receiver will
be sufficiently immune to this conducted EMI.
For most purposes an isolational transformer would solve this issue. The
unfortunate signal characteristics of a PPS pulse makes this a little
more cumbersome, but not unachievable, but it is no longer a simple
passive device. For higher frequencies will RF chokes be an aid of
course, but the RF choke needs "bolting down" in order to be effective,
so that there is a common mode current for the RF choke to object to.
However, the RF choke is not as effective with lower frequencies and
essentially useless for DC.
And now the question:
What standards exist governing required immunity of signal ports to these
ground-loop induced power-frequency (hum) voltages?
All the conducted suseptability standards I've found cover only
frequencies exceeding 10 KHz, not power frequencies and their harmonics.
You should look into the telecom set of standards. If you think of it,
they have been addressing this particular problem for ages. The words
which probably get you right on the target is "bonding network" since
you bond to the ground.
In short, there are two grounding strategies: all gear is floating
relative the safety ground or all gear is internally tied to the safety
ground. There is benefits and problems with both strategies. Regardless,
a hierarchial star ground strategy emerges.
One document to start with is the "Qwest Technical Publication
Grounding - Central Office and Remote Equipment Environment" at
http://www.qwest.com/techpub/77355/77355.pdf
Not to say that it is the standard of any sort, but I think it is a good
document to start from as it is a public source of telecom bonding
practices to be used in many facilities, implementing existing
international standards and involving transmitting towers (which is
within your field).
IEC 60950 should be a standard reference regardless.
You should also consult Bellcore GR-1089. There are additional Bellcore
specs, but starting with GR-63 and GR-1089 is not totally off the mark
at least. Bellcore specs costs money, but if you need to comply there is
no alternative.
ITU-T has a set of documents, such as the K-series of standards. You can
download these for free at:
http://www.itu.int/rec/T-REC-K/e
The European telecom world uses ETSI EN 300 253 as basis. They require a
login which you can get for free and then pull down all the documents
you like. There is also alot of specific EMC documents for various
contexts etc and they are all there. ETSI EMC is the TB handling them.
On the military side, MIL-HDBK-419 may be a guide:
http://tscm.com/MIL-HDBK-419A.PDF
Old standard MIL-STD-188-124B:
http://www.tscm.com/MIL-STD-188-124B.PDF
Newer stdandard MIL-STD-1310 for ships:
http://www.earth2.net/parts/basics/milstd1310g.pdf
In the end, all these documents forms a reference of standards and
practice in a varity of environments. I suspect that your environment
does has some bonding standard and practice and you need to figure out
what it is so that you know what you can expect, what you need to
fullfill (which is limiting freedom on what methods you may apply!) and
then it becomes easier to say what may help you. Also, you need to
figure out what is the type of problems you run into, how disturbances
actually induce into your lines. It could very well be that PSUs acts as
EMF due to bad conditioning for instance.
There are many anecdotes and horror stories to be told on the subject.
There are also sucsess stories to be told.
What makes the field a bit complex is that you need to think about
failures, EMC, bonding, interference, lightning strikes (on wire, in
tower, on building) which can cause a disparity of various indirect
effects. It's a bit like being a time-nut. We could probably have a
separate email list setup for that kind of discussions alone.
Cheers,
Magnus
Magnus,
time-nuts-bounces@febo.com wrote on 01/07/2009 01:27:52 AM:
Joseph M Gwinn skrev:
First the background:
In some timing distribution applications, the primary source of
interference comes from different ground voltages in different parts
of
the facility, such as a ship or a megawatt radar.
I left a useful detail out: The reference signal is a 10 MHz sinewave.
The effect of differing ground potentials on a shielded cable is to
pull a
large current through the shield, so there is a significant voltage
between the ends of the cable. No matter how good the shieldis at RF,
one consequence is that the same power-frequency offset voltage
appears on
the conductors within that shield, because the skin depth at 60 Hz
vastly
exceeds the thickness of any reasonable shield. Unshielded twisted
pair
will suffer the same common-mode offset voltage, perhaps more. This
offset often contains significant harmonics of the power frequency,
nominally up to the seventh harmonic, not just the fundamental.
If the cable is shielded twisted pair, such as twinax, the offset
appears
as a common-mode voltage on the two conductors, and (if not too large)
is
eliminated by the CMRR of the receiver.
If the cable is coax, the offset voltage appears added to the timing
signal voltage, and if the offset isn't too large the signal receiver
will
be sufficiently immune to this conducted EMI.
For most purposes an isolation transformer would solve this issue. The
unfortunate signal characteristics of a PPS pulse makes this a little
more cumbersome, but not unachievable, but it is no longer a simple
passive device. For higher frequencies will RF chokes be an aid of
course, but the RF choke needs "bolting down" in order to be effective,
so that there is a common mode current for the RF choke to object to.
However, the RF choke is not as effective with lower frequencies and
essentially useless for DC.
The receivers have built-in RF transformers. There is no 1PPS signal per
se, although the transformer would probably pass such a signal well
enough. What is being carried is 10 MHz.
The problem is to devise a test and spec that ensures that the actual
implemented circuit in the receivers suffice. There are many ways to
botch this circuit.
And now the question:
What standards exist governing required immunity of signal ports to
these
ground-loop induced power-frequency (hum) voltages?
All the conducted suseptability standards I've found cover only
frequencies exceeding 10 KHz, not power frequencies and
theirharmonics.
You should look into the telecom set of standards. If you think of it,
they have been addressing this particular problem for ages. The words
which probably get you right on the target is "bonding network" since
you bond to the ground.
This is just the sort of lead I was hoping to find.
In short, there are two grounding strategies: all gear is floating
relative the safety ground or all gear is internally tied to the safety
ground. There is benefits and problems with both strategies. Regardless,
a hierarchial star ground strategy emerges.
In our systems, everything is tied to ground for both safety and RF
reasons unrelated to timing signals. And we do have a star of sorts, but
the story always ends up more complex than that, so it always ends up
being a somewhat random grounding grid.
My problem is not safety, it is tolerance of conducted EMI.
One document to start with is the "Qwest Technical Publication
Grounding - Central Office and Remote Equipment Environment" at
http://www.qwest.com/techpub/77355/77355.pdf
Not to say that it is the standard of any sort, but I think it is a good
document to start from as it is a public source of telecom bonding
practices to be used in many facilities, implementing existing
international standards and involving transmitting towers (which is
within your field).
IEC 60950 should be a standard reference regardless.
You should also consult Bellcore GR-1089. There are additional Bellcore
specs, but starting with GR-63 and GR-1089 is not totally off the mark
at least. Bellcore specs costs money, but if you need to comply there is
no alternative.
ITU-T has a set of documents, such as the K-series of standards. You can
download these for free at:
http://www.itu.int/rec/T-REC-K/e
The European telecom world uses ETSI EN 300 253 as basis. They require a
login which you can get for free and then pull down all the documents
you like. There is also alot of specific EMC documents for various
contexts etc and they are all there. ETSI EMC is the TB handling them.
On the military side, MIL-HDBK-419 may be a guide:
http://tscm.com/MIL-HDBK-419A.PDF
Old standard MIL-STD-188-124B:
http://www.tscm.com/MIL-STD-188-124B.PDF
Newer standard MIL-STD-1310 for ships:
http://www.earth2.net/parts/basics/milstd1310g.pdf
I will be doing some homework. Some of these are tomes.
In the end, all these documents forms a reference of standards and
practice in a varity of environments. I suspect that your environment
does has some bonding standard and practice and you need to figure out
what it is so that you know what you can expect, what you need to
fullfill (which is limiting freedom on what methods you may apply!) and
then it becomes easier to say what may help you. Also, you need to
figure out what is the type of problems you run into, how disturbances
actually induce into your lines. It could very well be that PSUs acts as
EMF due to bad conditioning for instance.
There are many anecdotes and horror stories to be told on the subject.
There are also sucesses stories to be told.
We do have a bonding story, one that sort-of follows MIL-STD-1310, even
though the system is land based.
What makes the field a bit complex is that you need to think about
failures, EMC, bonding, interference, lightning strikes (on wire, in
tower, on building) which can cause a disparity of various indirect
effects. It's a bit like being a time-nut. We could probably have a
separate email list setup for that kind of discussions alone.
Fortunately for me, I do not have to worry about lightning. That's
handled elsewhere, as all these cables are within a steel-frame building
with a lightning protection system built in.
Joe
Joseph,
Joseph M Gwinn skrev:
Magnus,
time-nuts-bounces@febo.com wrote on 01/07/2009 01:27:52 AM:
Joseph M Gwinn skrev:
First the background:
In some timing distribution applications, the primary source of
interference comes from different ground voltages in different parts
of
the facility, such as a ship or a megawatt radar.
I left a useful detail out: The reference signal is a 10 MHz sinewave.
10 MHz into a transmitter. Should not be too hard to master. For some
reason I feel confident in that environment. :)
For most purposes an isolation transformer would solve this issue. The
unfortunate signal characteristics of a PPS pulse makes this a little
more cumbersome, but not unachievable, but it is no longer a simple
passive device. For higher frequencies will RF chokes be an aid of
course, but the RF choke needs "bolting down" in order to be effective,
so that there is a common mode current for the RF choke to object to.
However, the RF choke is not as effective with lower frequencies and
essentially useless for DC.
The receivers have built-in RF transformers. There is no 1PPS signal per
se, although the transformer would probably pass such a signal well
enough. What is being carried is 10 MHz.
The problem is to devise a test and spec that ensures that the actual
implemented circuit in the receivers suffice. There are many ways to
botch this circuit.
I see. It is fairly easy to induce common mode currents and DC voltages.
An isolational transformer from a source and then on the other side
simply DC offset or apply signal through a transformer if not directly
from an amplifier.
You should look into the telecom set of standards. If you think of it,
they have been addressing this particular problem for ages. The words
which probably get you right on the target is "bonding network" since
you bond to the ground.
This is just the sort of lead I was hoping to find.
Great.
In short, there are two grounding strategies: all gear is floating
relative the safety ground or all gear is internally tied to the safety
ground. There is benefits and problems with both strategies. Regardless,
a hierarchial star ground strategy emerges.
In our systems, everything is tied to ground for both safety and RF
reasons unrelated to timing signals. And we do have a star of sorts, but
the story always ends up more complex than that, so it always ends up
being a somewhat random grounding grid.
As always.
My problem is not safety, it is tolerance of conducted EMI.
The reason I mention safety is that some people suggest solutions which
does not fullfill the safety criteria in spirit or standard. It gets you
into the right category of solutions.
One document to start with is the "Qwest Technical Publication
Grounding - Central Office and Remote Equipment Environment" at
http://www.qwest.com/techpub/77355/77355.pdf
Not to say that it is the standard of any sort, but I think it is a good
document to start from as it is a public source of telecom bonding
practices to be used in many facilities, implementing existing
international standards and involving transmitting towers (which is
within your field).
IEC 60950 should be a standard reference regardless.
You should also consult Bellcore GR-1089. There are additional Bellcore
specs, but starting with GR-63 and GR-1089 is not totally off the mark
at least. Bellcore specs costs money, but if you need to comply there is
no alternative.
ITU-T has a set of documents, such as the K-series of standards. You can
download these for free at:
http://www.itu.int/rec/T-REC-K/e
The European telecom world uses ETSI EN 300 253 as basis. They require a
login which you can get for free and then pull down all the documents
you like. There is also alot of specific EMC documents for various
contexts etc and they are all there. ETSI EMC is the TB handling them.
On the military side, MIL-HDBK-419 may be a guide:
http://tscm.com/MIL-HDBK-419A.PDF
Old standard MIL-STD-188-124B:
http://www.tscm.com/MIL-STD-188-124B.PDF
Newer standard MIL-STD-1310 for ships:
http://www.earth2.net/parts/basics/milstd1310g.pdf
I will be doing some homework. Some of these are tomes.
You could also look up ETSI EN 300 132-* and EN 300 386 which is
relevant for telecom boxes. Further on is EN 300 199-* probably good to
have around, but maybe not so applicable to this particular problem.
What you want to transfer is similar to an E1 or E2 on an intra-office link.
EMF due to bad conditioning for instance.
There are many anecdotes and horror stories to be told on the subject.
There are also sucesses stories to be told.
We do have a bonding story, one that sort-of follows MIL-STD-1310, even
though the system is land based.
Sounds good. Will think about levels.
What makes the field a bit complex is that you need to think about
failures, EMC, bonding, interference, lightning strikes (on wire, in
tower, on building) which can cause a disparity of various indirect
effects. It's a bit like being a time-nut. We could probably have a
separate email list setup for that kind of discussions alone.
Fortunately for me, I do not have to worry about lightning. That's
handled elsewhere, as all these cables are within a steel-frame building
with a lightning protection system built in.
Actually, the most outer cabling links needs to be shielded or else they
would intduce into cables.
Cheers,
Magnus
In message <OFADE54B4F.D29DBA7A-ON85257537.00086866-85257537.000907CA@mck.us.ra
y.com>, Joseph M Gwinn writes:
The effect of differing ground potentials on a shielded cable is to pull a
large current through the shield, [...]
The correct enginering solution is to use twinax, ground the shield
in one end only and transformer-couple the signal at least in the
other end from the grounding.
Look at IBM's 5250 terminal hookup for an school book example of getting
it right.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Poul-Henning,
time-nuts-bounces@febo.com wrote on 01/07/2009 04:25:04 PM:
In message <OFADE54B4F.D29DBA7A-ON85257537.00086866-85257537.
000907CA@mck.us.ra
y.com>, Joseph M Gwinn writes:
The effect of differing ground potentials on a shielded cable is to
pull a
large current through the shield, [...]
The correct enginering solution is to use twinax, ground the shield
in one end only and transformer-couple the signal at least in the
other end from the grounding.
Yes, I know of this. Shielded twisted pair is also widely used in audio,
for the same reasons.
But my system is coax and was that way before I arrived.
I know of some similar but ship-board systems that use twinax for time
reference distribution, as you suggest. Don't know if they use
transformers though. Could be a differential TX and RX. I recall that
they send a RS422 signal. I imagine that the shield is grounded at both
ends, if only for safety reasons.
Fortunately, my system is not so noisy as a ship.
If I had it to do over, I might well use multimode fiber.
Look at IBM's 5250 terminal hookup for an school book example of getting
it right.
A blast from the past - shades of the 1970s!
A parallel story: Some years ago I was working on shipboard systems that
used 10BASE5 ethernet over thick coax (nominally RG8). The problem is
that there is no real ground on a ship, and there can be 7 volts
difference between bow and stern because the hull is used as the power
system neutral. Well, 10BASE5 ethernet uses 2-volt signals, so 7 volts
offset would prevent communications. The solution was to use triax. The
outer shield was grounded at both ends. The inner shield and center
conductor together formed the ethernet media. The inner shield was
connected to the outer shield in exactly one place. For safety, this
connection had to be able to handle 1,000 amps, to ensure that breakers
would pop before ground links opened. (One of our young engineers was
going to use a AWG #30 wire-wrap link.) The outer shield stopped at the
cabinet I/O panel, with only the inner shield and center conductor
continuing (as a bit of RG58) to the etherent transceivers. This worked
flawlessly.
Joe
In message <OF56303512.93B049A7-ON85257537.0079CDE3-85257537.007CC84F@mck.us.ra
y.com>, Joseph M Gwinn writes:
Could be a differential TX and RX. I recall that they send a RS422 signal.
Depending on the speed, RS422 works fine with transformers.
I imagine that the shield is grounded at both ends, if only for
safety reasons.
That is actually a very unsafe practice, unless there is another
much thicker and reliable ground connection between the two domains.
But you should never let the screen float in the far end, you should
terminate it with a 10M resistor and a sparkgap in parallel to the
local ground.
The resistor takes care of static electricity and the sparkgap will
do lightnings.
If I had it to do over, I might well use multimode fiber.
Yes, never roll copper more than 100m or between buildings if you
can get away with installing fiber.
The solution was to use triax. The
outer shield was grounded at both ends. The inner shield and center
conductor together formed the ethernet media. The inner shield was
connected to the outer shield in exactly one place.
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
But yes, double shielding works great, provided you don't have morons
with screwdrivers around.
Poul-Henning
(Who once lost all ethernet interfaces, the access control system
and a few minor computers when a moron first created and then cut
a 600+ A ground loop).
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Poul-Henning,
time-nuts-bounces@febo.com wrote on 01/07/2009 05:56:19 PM:
In message <OF56303512.93B049A7-ON85257537.0079CDE3-85257537.
007CC84F@mck.us.ra
y.com>, Joseph M Gwinn writes:
Could be a differential TX and RX. I recall that they send a RS422
signal.
Depending on the speed, RS422 works fine with transformers.
Yes. It would be 10 MHz or 20 MHz, depending on coding. Or 5 MHz, so the
transitions are at 10 MHz. I don't recall, or never knew.
I imagine that the shield is grounded at both ends, if only for
safety reasons.
That is actually a very unsafe practice, unless there is another
much thicker and reliable ground connection between the two domains.
There is a very heavy grounding grid, and such systems almost always
ground the (outer) shields at every connector.
But you should never let the screen float in the far end, you should
terminate it with a 10M resistor and a sparkgap in parallel to the
local ground.
The resistor takes care of static electricity and the sparkgap will
do lightnings.
I've done such things, but with a 100 ohm resistor (and a safety ground to
ensure that the voltage doesn't get too large. But this was a lab lashup.
If I had it to do over, I might well use multimode fiber.
Yes, never roll copper more than 100m or between buildings if you
can get away with installing fiber.
The solution was to use triax. The
outer shield was grounded at both ends. The inner shield and center
conductor together formed the ethernet media. The inner shield was
connected to the outer shield in exactly one place.
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
No, I think that's twinax: http://en.wikipedia.org/wiki/Twinax_cable.
Triax is a center plus two concentric shields:
http://en.wikipedia.org/wiki/Triaxial_cable.
The terms are very similar.
But yes, double shielding works great, provided you don't have morons
with screwdrivers around.
Poul-Henning
(Who once lost all ethernet interfaces, the access control system
and a few minor computers when a moron first created and then cut
a 600+ A ground loop).
Was there a big bang? What was the source of the 600 amps?
Joe
Poul-Henning Kamp skrev:
In message <OF56303512.93B049A7-ON85257537.0079CDE3-85257537.007CC84F@mck.us.ra
y.com>, Joseph M Gwinn writes:
Could be a differential TX and RX. I recall that they send a RS422 signal.
Depending on the speed, RS422 works fine with transformers.
You want a DC balanced encoding if you send data down the line,
otherwise you will saturate the transformer and saturation works very
well as method of damping. This was actually used for modulation of high
powers and then the transformer was called a transductor, essentially a
transformer with too small core. This is the core at the Grimeton long
wave transmitter (a world heritage site) in south of Sweden. Lovely
thing to visit. The 127 m high and 1,8 km long antenna is not easy to
miss. I beleive the output effect was 200 kW at 17,2 kHz. It has a
definitive steam-engine feel to it. Lovely.
Usually thought the signal is just dampend out. Only transitions
survive. A pure 10 MHz is not a problem at all, but generic RS422 may
not survive.
I imagine that the shield is grounded at both ends, if only for
safety reasons.
That is actually a very unsafe practice, unless there is another
much thicker and reliable ground connection between the two domains.
Which is what most bonding network standards will describe never the less.
But you should never let the screen float in the far end, you should
terminate it with a 10M resistor and a sparkgap in parallel to the
local ground.
The resistor takes care of static electricity and the sparkgap will
do lightnings.
You most probably want to use a capacitor from shield to chassi both for
providing a low impedance path for RF and static electricity blasts but
also helps in reducing the RF emission. Clamping an external RF choke on
the cable will be meaningfull when the cap is there as the RF choke is
being properly terminated.
If I had it to do over, I might well use multimode fiber.
Yes, never roll copper more than 100m or between buildings if you
can get away with installing fiber.
So true. Not that you can't get it to work, but it is tedious to make it
work under all conditions.
The solution was to use triax. The
outer shield was grounded at both ends. The inner shield and center
conductor together formed the ethernet media. The inner shield was
connected to the outer shield in exactly one place.
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
But yes, double shielding works great, provided you don't have morons
with screwdrivers around.
The ethernet habit of using vampire clamps provided a great opportunity
for less insightful installation practices.
Poul-Henning
(Who once lost all ethernet interfaces, the access control system
and a few minor computers when a moron first created and then cut
a 600+ A ground loop).
As I said... :)
Remember, we need to support our morons in their daily task. :)
Cheers,
Magnus
Joseph,
Could be a differential TX and RX. I recall that they send a RS422
signal.
Depending on the speed, RS422 works fine with transformers.
Yes. It would be 10 MHz or 20 MHz, depending on coding. Or 5 MHz, so the
transitions are at 10 MHz. I don't recall, or never knew.
RS422 does not imply any encoding as such so it would be 10 MHz but
naturally there is twice that many transitions, but it is the frequency
of the signal you are interested in for this case.
I imagine that the shield is grounded at both ends, if only for
safety reasons.
That is actually a very unsafe practice, unless there is another
much thicker and reliable ground connection between the two domains.
There is a very heavy grounding grid, and such systems almost always
ground the (outer) shields at every connector.
Which would imply that if the signal passes through a connector jack or
through a wall, much of the current would be sent back to its EMF source
locally in the room. This does have its merits.
But you should never let the screen float in the far end, you should
terminate it with a 10M resistor and a sparkgap in parallel to the
local ground.
The resistor takes care of static electricity and the sparkgap will
do lightnings.
I've done such things, but with a 100 ohm resistor (and a safety ground to
ensure that the voltage doesn't get too large. But this was a lab lashup.
The trouble with 100 ohm is that still can be a little low in relation
to ground loop impedances, it still allow some fair current to roll down
the cable. A capacitor in parallel would cut most of the transient
energy straight through and allow for a higher resistive path for the
low frequency energy.
If I had it to do over, I might well use multimode fiber.
Yes, never roll copper more than 100m or between buildings if you
can get away with installing fiber.
The solution was to use triax. The
outer shield was grounded at both ends. The inner shield and center
conductor together formed the ethernet media. The inner shield was
connected to the outer shield in exactly one place.
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
No, I think that's twinax: http://en.wikipedia.org/wiki/Twinax_cable.
Triax is a center plus two concentric shields:
http://en.wikipedia.org/wiki/Triaxial_cable.
The terms are very similar.
I have some triax cables and connectors, but not twinax...
But yes, double shielding works great, provided you don't have morons
with screwdrivers around.
Poul-Henning
(Who once lost all ethernet interfaces, the access control system
and a few minor computers when a moron first created and then cut
a 600+ A ground loop).
Was there a big bang? What was the source of the 600 amps?
I think there (with some delay) was some awfull scream of dispare.
The cost of Ethernet interfaces where much more significant back then.
Cheers,
Magnus
In message <OF3277AC5A.F5D1FAE8-ON85257537.008059CF-85257537.00817C56@mck.us.ra
y.com>, Joseph M Gwinn writes:
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
No, I think that's twinax: http://en.wikipedia.org/wiki/Twinax_cable.
Triax is a center plus two concentric shields:
http://en.wikipedia.org/wiki/Triaxial_cable.
Sorry, I fumbled what I wrote there. I would say wiki is wrong
here, the usage I am used to is:
coax: single conductor + shield
twinax: twisted pair + shield
triax: the wires + shield
(Who once lost all ethernet interfaces, the access control system
and a few minor computers when a moron first created and then cut
a 600+ A ground loop).
Was there a big bang? What was the source of the 600 amps?
They replaced the separation transformer with a UPS, and they
connected the two sides ground together at the UPS.
Unfortunately the grounding on our secondary side was much better
than the power companys grounding on the primary side, which was the
entire point of having the the transformer in the first place.
Yes, there were a significant bang and his two-hand wire-cutter was
recategorized from "tool" to "industrial art".
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
In message 49657762.5060504@rubidium.dyndns.org, Magnus Danielson writes:
Was there a big bang? What was the source of the 600 amps?
I think there (with some delay) was some awfull scream of dispare.
The cost of Ethernet interfaces where much more significant back then.
The most expensive one we lost was in a UNISYS 2200, where three
microprocessors worked together to limit bandwidth to 100 kB/s.
I belive the sticker prices as $15k.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Poul-Henning Kamp wrote:
In message <OF3277AC5A.F5D1FAE8-ON85257537.008059CF-85257537.00817C56@mck.us.ra
y.com>, Joseph M Gwinn writes:
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
No, I think that's twinax: http://en.wikipedia.org/wiki/Twinax_cable.
Triax is a center plus two concentric shields:
http://en.wikipedia.org/wiki/Triaxial_cable.
Sorry, I fumbled what I wrote there. I would say wiki is wrong
here, the usage I am used to is:
coax: single conductor + shield
twinax: twisted pair + shield
triax: the wires + shield
(Who once lost all ethernet interfaces, the access control system
and a few minor computers when a moron first created and then cut
a 600+ A ground loop).
Was there a big bang? What was the source of the 600 amps?
They replaced the separation transformer with a UPS, and they
connected the two sides ground together at the UPS.
Unfortunately the grounding on our secondary side was much better
than the power companys grounding on the primary side, which was the
entire point of having the the transformer in the first place.
Yes, there were a significant bang and his two-hand wire-cutter was
recategorized from "tool" to "industrial art".
Similarly for Quadraxial cable there are 2 interpretations:
-
an inner conductor surrounded by 3 coaxial tubular conductors all
insulated from each other. -
2 twisted pairs with an outer tubular shield used in some high speed
network cabling.
Both meanings are in common use.
Quintaxial cable seems only to be mentioned in texts on cable shielding.
In which it consists of a central conductor surrounded by 4 coaxial
tubular screens all of which are insulated from each other.
Bruce
Poul-Henning Kamp wrote:
In message <OF3277AC5A.F5D1FAE8-ON85257537.008059CF-85257537.00817C56@mck.us.ra
y.com>, Joseph M Gwinn writes:
That's technically speaking not triax, that's double shield. Triax
would have the conductors and one shield.
No, I think that's twinax: http://en.wikipedia.org/wiki/Twinax_cable.
Triax is a center plus two concentric shields:
http://en.wikipedia.org/wiki/Triaxial_cable.
Sorry, I fumbled what I wrote there. I would say wiki is wrong
here, the usage I am used to is:
coax: single conductor + shield
twinax: twisted pair + shield
triax: the wires + shield
Poul-Henning
I have been unable to find a reference to triax consisting of 3
conductors within a shield, however such confusion is understandable
given the confusion over quadrax:-
concentric conductors for triax, quadrax, quintax etc:
Cable Shielding for Electromagnetic Compatibility By Anatoly Tsaliovich
http://www.lemo.com/techlibrary/glossary.jsp?catID=003
Alternate definitions for quadrax (but not triax):
http://www.picwire.com/technical/Coax%20vs%20Triax.pdf
http://www.ecnmag.com/ethernet-cable-suits-aerospace.aspx?menuid=336
Triax consisting of a central conductor surrounded by 2 concentric
conductors is widely used to interconnect televsion cameras.
Bruce
In message 49665A6D.2030100@xtra.co.nz, Bruce Griffiths writes:
I have been unable to find a reference to triax consisting of 3
conductors within a shield, however such confusion is understandable
given the confusion over quadrax:-
I have only ever seen it used for very old 3-electrode condenser
microphones.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
At 01:59 PM 1/8/2009 , Poul-Henning Kamp wrote:
In message 49665A6D.2030100@xtra.co.nz, Bruce Griffiths writes:
I have been unable to find a reference to triax consisting of 3
conductors within a shield, however such confusion is understandable
given the confusion over quadrax:-
I have only ever seen it used for very old 3-electrode condenser
microphones.
Doesn't Keithley use triax connectors on some of their high-impedance
instrumentation?
--
newell N5TNL
Scott Newell wrote:
At 01:59 PM 1/8/2009 , Poul-Henning Kamp wrote:
In message 49665A6D.2030100@xtra.co.nz, Bruce Griffiths writes:
I have been unable to find a reference to triax consisting of 3
conductors within a shield, however such confusion is understandable
given the confusion over quadrax:-
I have only ever seen it used for very old 3-electrode condenser
microphones.
Doesn't Keithley use triax connectors on some of their high-impedance
instrumentation?
Scott
They use the concentic conductor definition/interpretation of the term
triax.
I was seeking an example of the 3 conductor within a shield interpretation.
See the links on quadrax posted earlier for an idea of the pervasive
rampant confusion.
Bruce
Magnus,
time-nuts-bounces@febo.com wrote on 01/07/2009 10:47:46 PM:
Joseph,
Could be a differential TX and RX. I recall that they send a RS422
signal.
Depending on the speed, RS422 works fine with transformers.
Yes. It would be 10 MHz or 20 MHz, depending on coding. Or 5 MHz, so
the
transitions are at 10 MHz. I don't recall, or never knew.
RS422 does not imply any encoding as such so it would be 10 MHz but
naturally there is twice that many transitions, but it is the frequency
of the signal you are interested in for this case.
I know that RS422 is not the encoding. I cheated, and talked to the
relevant engineer.
For digital signals (1PPS, various triggers), it's RS422 over 100 ohm
twinax (fancy shielded twisted pair).
The 10 MHz sinewave is sent over a pair of 50 ohm coax links, with the
signals 180 degrees out of phase. This is acheived with a pair of hybrid
transformers which convert from one-cable to two-cable and then back to
one-cable, where all cables are 50 ohm coax.
I imagine that the shield is grounded at both ends, if only for
safety reasons.
That is actually a very unsafe practice, unless there is another
much thicker and reliable ground connection between the two domains.
There is a very heavy grounding grid, and such systems almost always
ground the (outer) shields at every connector.
Which would imply that if the signal passes through a connector jack or
through a wall, much of the current would be sent back to its EMF source
locally in the room. This does have its merits.
Yes, but that isn't the reason. It's really a safety and EMC rationale.
But you should never let the screen float in the far end, you should
terminate it with a 10M resistor and a sparkgap in parallel to the
local ground.
The resistor takes care of static electricity and the sparkgap will
do lightnings.
I've done such things, but with a 100 ohm resistor (and a safety
ground to
ensure that the voltage doesn't get too large. But this was
a lab lashup.
The trouble with 100 ohm is that still can be a little low in relation
to ground loop impedances, it still allow some fair current to roll down
the cable. A capacitor in parallel would cut most of the transient
energy straight through and allow for a higher resistive path for the
low frequency energy.
The ground grid impedance between any two points is well less than one
ohm, so 100 ohms will pretty much abolish all ground loops. I've used 10
ohms in like labs, with success. I'll grant that this would not work with
long wires outside.
By the way, I also finally talked to one of our most experienced EMI/EMC
engineers. He suggested using MIL-STD-461 test CS109, even though CS109
was developed for enclosures. It turns out he was involved in developing
CS109 when he worked for the US Navy.
Joe
time-nuts-bounces@febo.com wrote on 01/08/2009 03:47:29 AM:
In message <OF3277AC5A.F5D1FAE8-ON85257537.008059CF-85257537.
00817C56@mck.us.ray.com>, Joseph M Gwinn writes:
Was there a big bang? What was the source of the 600 amps?
They replaced the separation transformer with a UPS, and they
connected the two sides ground together at the UPS.
Unfortunately the grounding on our secondary side was much better
than the power companys grounding on the primary side, which was the
entire point of having the the transformer in the first place.
One assumes that there were too many cooks.
Yes, there were a significant bang and his two-hand wire-cutter was
recategorized from "tool" to "industrial art".
He probably needed a stiff drink after that.
Joe
On Thu, Jan 08, 2009 at 08:51:45AM +0000, Poul-Henning Kamp wrote:
In message 49657762.5060504@rubidium.dyndns.org, Magnus Danielson writes:
Was there a big bang? What was the source of the 600 amps?
I think there (with some delay) was some awfull scream of dispare.
The cost of Ethernet interfaces where much more significant back then.
The most expensive one we lost was in a UNISYS 2200, where three
microprocessors worked together to limit bandwidth to 100 kB/s.
I belive the sticker prices as $15k.
I'm somewhat confused about how this took out Ethernet
transceivers or interfaces... from the beginning even vampire tap RG-8
yellow cable Ethernet transceivers were ground isolated from chassis
ground of the computer system just exactly to avoid ground loops and
back path ground currents. Both power and transmit/receive and control
signals are isolated... and usually transformer coupled... and as I
remember it rather a substantial voltage difference between shield on
the cable and computer system ground had to be tolerated (hundreds of
volts at least)...
I guess, however, if someone grounded the yellow cable at more
than one point enough current could flow on its outer conductor to
induce substantial voltage between the shield and the center conductor
which could trash the driver/receiver/carrier sense chips or protective
clamp diodes ...
One was never, of course, supposed to ground the yellow cable at
more than one point...
--
Dave Emery N1PRE/AE, die@dieconsulting.com DIE Consulting, Weston, Mass 02493
"An empty zombie mind with a forlorn barely readable weatherbeaten
'For Rent' sign still vainly flapping outside on the weed encrusted pole - in
celebration of what could have been, but wasn't and is not to be now either."
Joseph,
time-nuts-bounces@febo.com wrote on 01/07/2009 10:47:46 PM:
Joseph,
Could be a differential TX and RX. I recall that they send a RS422
signal.
Depending on the speed, RS422 works fine with transformers.
Yes. It would be 10 MHz or 20 MHz, depending on coding. Or 5 MHz, so
the
transitions are at 10 MHz. I don't recall, or never knew.
RS422 does not imply any encoding as such so it would be 10 MHz but
naturally there is twice that many transitions, but it is the frequency
of the signal you are interested in for this case.
I know that RS422 is not the encoding. I cheated, and talked to the
relevant engineer.
That is to cheat! :)
For digital signals (1PPS, various triggers), it's RS422 over 100 ohm
twinax (fancy shielded twisted pair).
The 10 MHz sinewave is sent over a pair of 50 ohm coax links, with the
signals 180 degrees out of phase. This is acheived with a pair of hybrid
transformers which convert from one-cable to two-cable and then back to
one-cable, where all cables are 50 ohm coax.
OUCH! The trouble with that arrangement is that the coax cables MUST be
twisted or else H-fields will induce differential mode current. It will
induce current into both directions which through the 180 degree will
not cancel but add up. The 0/180 degree arrangement will save you from
common mode problems. You would prefer a twisted cable over a twisted
cable pair, as the later allows for installation procedure errors to
have huge impact and the twisting properties will not be as good either
and thus compromising the quality. A single ended coax is not as
sensitive to H fields to induce diffrential currents, but can have some
other problems.
I imagine that the shield is grounded at both ends, if only for
safety reasons.
That is actually a very unsafe practice, unless there is another
much thicker and reliable ground connection between the two domains.
There is a very heavy grounding grid, and such systems almost always
ground the (outer) shields at every connector.
Which would imply that if the signal passes through a connector jack or
through a wall, much of the current would be sent back to its EMF source
locally in the room. This does have its merits.
Yes, but that isn't the reason. It's really a safety and EMC rationale.
As suspected, but this is really just another of these EMC rationales.
But you should never let the screen float in the far end, you should
terminate it with a 10M resistor and a sparkgap in parallel to the
local ground.
The resistor takes care of static electricity and the sparkgap will
do lightnings.
I've done such things, but with a 100 ohm resistor (and a safety
ground to
ensure that the voltage doesn't get too large. But this was
a lab lashup.
The trouble with 100 ohm is that still can be a little low in relation
to ground loop impedances, it still allow some fair current to roll down
the cable. A capacitor in parallel would cut most of the transient
energy straight through and allow for a higher resistive path for the
low frequency energy.
The ground grid impedance between any two points is well less than one
ohm, so 100 ohms will pretty much abolish all ground loops. I've used 10
ohms in like labs, with success. I'll grant that this would not work with
long wires outside.
Should be sufficient then. But remember that capacitive coupling helps
you in the RF area and impulse protection.
By the way, I also finally talked to one of our most experienced EMI/EMC
engineers. He suggested using MIL-STD-461 test CS109, even though CS109
was developed for enclosures. It turns out he was involved in developing
CS109 when he worked for the US Navy.
Need to look it up. Never had to do any of the MIL-STD-461 stuff.
Cheers,
Magnus