Carrier current communication on Low voltage/high current AC

[Jim Thompson]

On Fri, 10 Jun 2005 19:35:39 +1200, Terry Given <my_name@ieee.org>
wrote:

Jim Thompson wrote:
On Thu, 09 Jun 2005 19:03:45 -0700, Jim Thompson
thegreatone@example.com> wrote:


On Thu, 09 Jun 2005 18:49:52 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:


On Thu, 09 Jun 2005 13:01:48 -0400, Spehro Pefhany
speffSNIP@interlogDOTyou.knowwhat> wrote:


On Thu, 09 Jun 2005 07:58:10 -0700, the renowned John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

"Is also rejected"? Some sort of Royal decree?

John

I like it better than that execrable term "deemed".



The one I hate is "good engineering practice" which usually means "I
always do it that way but can't actually explain why."


John

I use the term "good engineering practice" to indicate to the client
that, if they'd done it my way, they wouldn't be in ka-ka-land right
now ;-)

...Jim Thompson


I should have added... I told a client more than two weeks ago, "your
design failed because....".

They vehemently denied it could happen, that I was totally wrong, how
could I be so wrong, all their simulations showed it was stable, etc.

Guess who's kissing my ass today ?

...Jim Thompson

Ya gotta love it when that happens. I love that Gore Vidal quote - "It
is not enough to succeed, others must fail"

Cheers
Terry

The only problem is that you can't gloat and rub it in, it drives the
clients away.

So all I can do is laugh with the wife over a glass of wine ;-)

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Ken Smith]

In article <4d78d01dfctonyw@ledelec.demon.co.uk>,
Tony Williams <tonyw@ledelec.demon.co.uk> wrote:
[....]
Notes added to drawing:

Hot +-------------------- Black ------------------+

60Hz supply Load
L1 L2
+------///----+------ White -----+------///---+
YES |? | | C3 |? NOT
GND +-------------|---+-- Green -+---|------------+ REALLY
| | | |
C1=== | | ===C2
| | | |
Tx Rx

The white and green are hooked together by being grounded at the power
entry point. There should not be any other connection between the two
anywhere in the wiring.

The EMI reduction filters at the input of electronic equipment usually
have a lowish AC impedance at high frequencies because they are commonly
pi filters.

L1 and L2 are still handy additions because they isolate
the comms-driven section of line, away from any Grounding
that may happen at either end.

Watch that you don't run up against the NEC. Anything that introduces an
impedance in the white wire is frowned on. They flat hate anything that
could cause an open white wire. Since the impedance at the RF frequencies
is fairly low, L1 and L2 can be darn near zero at 60Hz.

L2 faces some unknown capacitance. It would be a real bummer if it went
into series resonance at the working frequency. C3 can prevent this by
making the series resonance alway appear below the working frequency.

Also: You may want to make L1 and L2 go into parallel resonance somewhat
above the working frequency. This makes their impedance higher in the
working band.

--
--
kensmith@rahul.net forging knowledge

 

[Winfield Hill]

Tony Williams wrote...

An inductor at each end of the line might be used to raise
the impedance at the carrier frequency. This would allow
capacitively-coupled differential signalling.

* L1 50A,60Hz * L2
+-----/////---+--------->--------+--/////----+
60Hz supply ----- | | ----- Load
+-----/////-------+----------+------/////----+
* | | | | *
C1=== | | ===C2
| | | |
Tx Rx

Perhaps look for inductors from switchers, in the range
20-50uH at 50Adc polarisation. Use Speff's 455KHz
carrier frequency.

.. * L1 50A,60Hz * L2
.. +-----/////---+----+---->--------+--/////----+
.. 60Hz supply ===== | '--||--, | ===== Load
.. +-----/////-------+-------+--+------/////----+
.. * | | oops! | | *
.. C1=== | | ===C2
.. | | | |
.. Tx Rx
..

What about high-frequency capacitive loading by other stuff
that's plugged into the AC mains? Another thing, L1 would
have to be a big beast, handling the entire AC main branch.


--
Thanks,
- Win

 

[John Larkin]

On Fri, 10 Jun 2005 06:40:21 -0700, Jim Thompson
<thegreatone@example.com> wrote:

On Fri, 10 Jun 2005 19:35:39 +1200, Terry Given <my_name@ieee.org
wrote:

Jim Thompson wrote:
On Thu, 09 Jun 2005 19:03:45 -0700, Jim Thompson
thegreatone@example.com> wrote:


On Thu, 09 Jun 2005 18:49:52 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:


On Thu, 09 Jun 2005 13:01:48 -0400, Spehro Pefhany
speffSNIP@interlogDOTyou.knowwhat> wrote:


On Thu, 09 Jun 2005 07:58:10 -0700, the renowned John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

"Is also rejected"? Some sort of Royal decree?

John

I like it better than that execrable term "deemed".



The one I hate is "good engineering practice" which usually means "I
always do it that way but can't actually explain why."


John

I use the term "good engineering practice" to indicate to the client
that, if they'd done it my way, they wouldn't be in ka-ka-land right
now ;-)

...Jim Thompson


I should have added... I told a client more than two weeks ago, "your
design failed because....".

They vehemently denied it could happen, that I was totally wrong, how
could I be so wrong, all their simulations showed it was stable, etc.

Guess who's kissing my ass today ?

...Jim Thompson

Ya gotta love it when that happens. I love that Gore Vidal quote - "It
is not enough to succeed, others must fail"

Cheers
Terry

The only problem is that you can't gloat and rub it in, it drives the
clients away.

So all I can do is laugh with the wife over a glass of wine ;-)

Right. You can prove that you're smarter than them, or you can get
repeat business, but not both.

John

 

[John Larkin]

On Fri, 10 Jun 2005 11:01:37 GMT, Fred Bloggs <nospam@nospam.com>
wrote:

John Larkin wrote:
On Thu, 09 Jun 2005 12:20:54 GMT, Fred Bloggs <nospam@nospam.com
wrote:



Tony Williams wrote:

In article <g5iea11r8tge2cpf7j0c3ri79h0ls0osip@4ax.com>,
Spehro Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote:



I'm interested in doing carrier current communication over 12VAC
power. Total AC current will probably be in the 10-50A range, and
I'd like to get 4800 or 9600 baud, but less would be acceptable.
There could be a couple hundred feet of wire, and I don't control
the source impedance.


Maybe think about transmitting the comms as a common-mode
voltage. The 24Vac system can be connected to Ground at
60Hz (or have stray-C to ground), just devise something
that makes that connection a high impedance at (and above)
the comms carrier frequency.


That's one of three possibilities of selecting two of three terminals
for the signal transmission , but for some reason it is forbidden for
safety reasons. Also, the current carrier method of inducing a signal
via transformer with secondary in series with line is also rejected-
probably too much bulk required there.



"Is also rejected"? Some sort of Royal decree?

John


Your series inductive coupling requires you to break the high power
utility circuit,

Speff said it was 'low voltage' which doesn't sound 'utility' to me; I
got the impression it might be some 'private' low-voltage system. And
you can slip a split-core ferrite over even an insulated conductor.

you do not need to break the circuit to install a shunt
source and the components do not have to be rated to withstand 10KA
surges and whatnots.

A low-mu core with a high-ratio winding, resonated on the primary,
clamped with a couple of zeners, won't have any spike problems. Making
a voltage connection into a utility line takes about as many
resonating and transient protection components.

ASK is the last kind of modulation anyone would
want to use.

Agree on that one. FSK or PSK makes more sense.


John

 

[John Larkin]

On Fri, 10 Jun 2005 10:44:01 GMT, Fred Bloggs <nospam@nospam.com>
wrote:

Tony Williams wrote:
In article <42A83423.1040507@nospam.com>,
Fred Bloggs <nospam@nospam.com> wrote:


Tony Williams wrote:

Maybe think about transmitting the comms as a common-mode
voltage....etc.


That's one of three possibilities of selecting two of three
terminals for the signal transmission , but for some reason it
is forbidden for safety reasons............


Pity. Common-mode signalling would probably be the
easier to implement.


........Also, the current carrier method of inducing a signal via
transformer with secondary in series with line is also rejected-
probably too much bulk required there.


An inductor at each end of the line might be used to raise
the impedance at the carrier frequency. This would allow
capacitively-coupled differential signalling.

* L1 50A,60Hz * L2
+-----/////---+--------->--------+--/////----+
60Hz supply ----- | | ----- Load
+-----/////-------+----------+------/////----+
* | | | | *
C1=== | | ===C2
| | | |
Tx Rx

Perhaps look for inductors from switchers, in the range
20-50uH at 50Adc polarisation. Use Speff's 455KHz
carrier frequency.


Now that looks darn near perfect....

Except for needing four windings in series with the 50-amp circuit.
Those will be huge transformers and nasty connections.

Why not

| tx | | rx |
| | | |
+---/////---+ +---/////---+

+-----/////---+--------->--------+--/////----+
60Hz supply Load
+-----------------+----------+---------------+

where the trannies are 1-turn primaries, just cores slipped over the
insulated power conductors?

John

 

[Spehro Pefhany]

On 10 Jun 2005 07:22:49 -0700, the renowned Winfield Hill
<hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote:

carrier frequency.

. * L1 50A,60Hz * L2
. +-----/////---+----+---->--------+--/////----+
. 60Hz supply ===== | '--||--, | ===== Load
. +-----/////-------+-------+--+------/////----+
. * | | oops! | | *
. C1=== | | ===C2
. | | | |
. Tx Rx
.

What about high-frequency capacitive loading by other stuff
that's plugged into the AC mains? Another thing, L1 would
have to be a big beast, handling the entire AC main branch.

The input is 12V and is dedicated. Probably the secondary of a fat
60Hz 120:12VAC transformer (up to 600VA). I have control over the
noise on the branch circuit(s), but there may be substantial
transients.



Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[John Larkin]

On Fri, 10 Jun 2005 12:09:42 -0400, Spehro Pefhany
<speffSNIP@interlogDOTyou.knowwhat> wrote:

On 10 Jun 2005 07:22:49 -0700, the renowned Winfield Hill
hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote:

carrier frequency.

. * L1 50A,60Hz * L2
. +-----/////---+----+---->--------+--/////----+
. 60Hz supply ===== | '--||--, | ===== Load
. +-----/////-------+-------+--+------/////----+
. * | | oops! | | *
. C1=== | | ===C2
. | | | |
. Tx Rx
.

What about high-frequency capacitive loading by other stuff
that's plugged into the AC mains? Another thing, L1 would
have to be a big beast, handling the entire AC main branch.

The input is 12V and is dedicated. Probably the secondary of a fat
60Hz 120:12VAC transformer (up to 600VA). I have control over the
noise on the branch circuit(s), but there may be substantial
transients.



Best regards,
Spehro Pefhany

How about the PoE trick, center-tapping both ends and communicating
via the center-taps, presumably against ground. For a 455 KHz signal,
the ct's could be just two caps on each end.

You can make a nice stable FM discriminator out of a ceramic resonator
and an xor gate.

John

 

[Spehro Pefhany]

On Fri, 10 Jun 2005 09:46:25 -0700, the renowned John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

How about the PoE trick, center-tapping both ends and communicating
via the center-taps, presumably against ground. For a 455 KHz signal,
the ct's could be just two caps on each end.

I don't think I've got access to ground. Otherwise that would be very
nice.

You can make a nice stable FM discriminator out of a ceramic resonator
and an xor gate.

John

That sounds like an interesting trick. How do you do that?


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[John Larkin]

On Fri, 10 Jun 2005 13:03:20 -0400, Spehro Pefhany
<speffSNIP@interlogDOTyou.knowwhat> wrote:

On Fri, 10 Jun 2005 09:46:25 -0700, the renowned John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:


How about the PoE trick, center-tapping both ends and communicating
via the center-taps, presumably against ground. For a 455 KHz signal,
the ct's could be just two caps on each end.

I don't think I've got access to ground. Otherwise that would be very
nice.

You can make a nice stable FM discriminator out of a ceramic resonator
and an xor gate.

John

That sounds like an interesting trick. How do you do that?

Once you get a bandlimited, saturated signal, apply it to one input of
an xor. Also run it through a resonator into the other xor input, and
lowpass filter the xor output. The resonator must produce a 90 degree
phase shift at the cf, and should have a 3 dB bw about twice the fm
deviation. You can get the phase shift from an r-l-c (technically a
lowpass filter, but with a big amplitude peak at cf) or better yet a
ceramic res circuit, which won't/might not require tuning.

John

 

[Mike Monett]

Tony Williams wrote:

[...]

I think that is a variation of the common-mode
signalling that has already been mentioned Mike.

Hot +-------------------- Black ------------------+

60Hz supply Load
L1 L2
+------///----+------ White -----+------///---+
|? | | |?
GND +-------------|---+-- Green -+---|------------+
| | | |
C1=== | | ===C2
| | | |
Tx Rx

L1 and L2 are still handy additions because they isolate
the comms-driven section of line, away from any Grounding
that may happen at either end.

Tony Williams.

Thanks, Tony. I didn't see it was already mentioned.

As far as isolation inductors, many power supplies have bypass caps from
hot and neutral to case gnd, which is connected to the green wire. So
there is a possible short for rf, which would negate the effect of the
isolation inductors.

If the driver can supply sufficient current, some voltage will be
developed across the power cable connecting the device to the line and
allow the signal to propagate. It's a bit like common-mode noise on a
scope ground.

In this case, perhaps the series inductors may not be needed. There are
many examples of carrier current devices on the market, such as
intercoms, baby-sitting units, and so on. I'm sure they don't use series
inductors, which would be expensive and difficult to install.

Mike Monett

 

[Mike Monett]

Spehro Pefhany wrote:

On Fri, 10 Jun 2005 09:43:49 +0100, the renowned Tony Williams
tonyw@ledelec.demon.co.uk> wrote:

[...]

Perhaps look for inductors from switchers, in the range
20-50uH at 50Adc polarisation. Use Speff's 455KHz
carrier frequency.

That should work nicely, of course one of the advantages of getting
the frequency up is to use a smaller inductor. It's almost into
reasonable air-core range, which could be helpful at one end.

I keep thinking I'm missing some obvious possibility with those
chokes, but maybe not.

Best regards,
Spehro Pefhany

You probably want to stay away from 455KHz - that could block every AM
radio on your street

Perhaps also try to avoid frequencies used by local low-power
non-directional beacons (NDB) for aircraft navigation. These are often
used to establish holding patterns for aircraft during storms.

I'd look at some commercial products using carrier current, such as
intercoms and baby-sitting devices, and figure out how they manage to
work without series inductors.

Here's one that simply injects the signal into the neutral, like my
proposal in another post that was quickly shot down

http://home.att.net/~weatheradio/carrier.htm

Mike Monett

 

[Mike Monett]

Mike Monett wrote:

[...]

and here's a complete transceiver at 135KHz

http://www.redcircuits.com/Page56.htm

Some signal processing may be needed to reduce interference from line
noise, but it seems you can simply inject the signal wherever you feel
like.

Mike Monett

 

[Joerg]

Hello Mike,

In this case, perhaps the series inductors may not be needed. There are
many examples of carrier current devices on the market, such as
intercoms, baby-sitting units, and so on. I'm sure they don't use series
inductors, which would be expensive and difficult to install.

Many PLC schemes simply blast RF onto the wires, with the driving
impedance as low as can be. Or as low as the parts budget allows. A
current coupling core would be way too expensive for most.

In Spehro's case this might work well if the driving transformer is high
enough in RF impedance. Most likely it is. A few ohms can be enough. So
in a low voltage app it would boil down to a stiff driver that is
capacitively coupled to the lines and protected against spikes by diodes
or something.

PLC such as X10 tend to fail when there is too much EMI filtering on the
line, like all the X-caps and Y-caps. In LV gear that often isn't the
case. IOW, know thy loads.

Regards, Joerg

http://www.analogconsultants.com

 

[Spehro Pefhany]

On Fri, 10 Jun 2005 10:39:56 -0700, the renowned John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

Once you get a bandlimited, saturated signal, apply it to one input of
an xor. Also run it through a resonator into the other xor input, and
lowpass filter the xor output. The resonator must produce a 90 degree
phase shift at the cf, and should have a 3 dB bw about twice the fm
deviation. You can get the phase shift from an r-l-c (technically a
lowpass filter, but with a big amplitude peak at cf) or better yet a
ceramic res circuit, which won't/might not require tuning.

John

Cute. A bit like the type I phase comparator in a 4046.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Mike Monett]

Mike Monett wrote:

[...]

Here's a 5 watt AM carrier current broadcast transmitter for college campuses
that shows another way of injecting the signal:

http://www.geocities.com/raiu_harrison/mwa/tech/circuits/xmitter3.html

OK, that's enough examples. Fire up your soldering iron and get one working

Mike Monett

 

[Mike Monett]

Joerg wrote:

Hello Mike,

In this case, perhaps the series inductors may not be needed. There are
many examples of carrier current devices on the market, such as
intercoms, baby-sitting units, and so on. I'm sure they don't use series
inductors, which would be expensive and difficult to install.

Many PLC schemes simply blast RF onto the wires, with the driving
impedance as low as can be. Or as low as the parts budget allows. A
current coupling core would be way too expensive for most.

Yes, that seems most reasonable. I listed urls for examples of three different
injection methods in other posts.

In Spehro's case this might work well if the driving transformer is high
enough in RF impedance. Most likely it is. A few ohms can be enough. So
in a low voltage app it would boil down to a stiff driver that is
capacitively coupled to the lines and protected against spikes by diodes
or something.

Perhaps the small inductance of a few feet of line cord is enough to allow the
signal to propagate. Protection against spikes is definitely good. Also may be a
good idea to think about transients due to nearby lightning stikes.

PLC such as X10 tend to fail when there is too much EMI filtering on the
line, like all the X-caps and Y-caps. In LV gear that often isn't the
case. IOW, know thy loads.

Some carrier current transmitters may go up to 50 Watts. That might overcome the
attenuation from a few line filters

Regards, Joerg

Mike Monett

 

[Spehro Pefhany]

On Fri, 10 Jun 2005 14:17:05 -0400, the renowned Mike Monett
<no@spam.com> wrote:

Mike Monett wrote:

[...]

Here's a 5 watt AM carrier current broadcast transmitter for college campuses
that shows another way of injecting the signal:

http://www.geocities.com/raiu_harrison/mwa/tech/circuits/xmitter3.html

OK, that's enough examples. Fire up your soldering iron and get one working

Mike Monett

Yes, probably a good idea!


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Spehro Pefhany]

On Fri, 10 Jun 2005 14:24:25 -0400, the renowned Mike Monett
<no@spam.com> wrote:

Some carrier current transmitters may go up to 50 Watts. That might overcome the
attenuation from a few line filters

From cost and EMI considerations, I was thinking more of <100mA at
100mV or so. ;-)


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Mike Monett]

Spehro Pefhany wrote:

From cost and EMI considerations, I was thinking more of <100mA at
100mV or so. ;-)

That might work if the transmitter and receiver share the same plug

Best regards,
Spehro Pefhany

Mike Monett