Re: Need help with PLC noise problems in a Manhattan (New York City) apartment
AlanTinNYC
the post in groups.google.com in its fixed-font mode and google
apparently wraps text at around 58 characters, not 60.
I'm reposting it below, reformatted for a smaller line.
--------------------------------------------------------------------
Hi there. I've been having ongoing noise issues with my
new apartment and our lights go on and off at random.
After testing for three months and trying several different
noise filters, I don't yet have the results I need. Time
to ask around for help.
Also, all the installers I can find in my "area" are
actually not in Manhattan but in other boros or Northern
New Jersey, where single-family homes are more prevalent
than apartment buildings.
Here's what I know:
1) My apartment building isn't huge - only about 50
apartments (8 or 9 on each of six floors)
2) My apartment is actually TWO apartments combined
(let's call them 2B and 2C) with the wall between them
knocked down. But we have two separate breaker panels,
each on a separate electric meter, each servicing half
the apartment.
3) I have learned, unfortunately, that I am far more
knowledgeable about X10/DHC than my general
contractor's licensed electrician is (or cares to be).
4) Using a Smarthome Testerlinc meter, I can determine
that I have 121kHz noise coming in from outside my
apartments because I can meter 120kHz activity, and bad
codes, while all our circuits breakers in both panels
are switched off (except for the breaker feeding this
meter).
5) The master shut-off breakers in the basement (two
per apartment; we have single split-phase wiring) are
grouped together in sets of 5 apartments, as follows
(you may need to switch to a fixed-width font):
______________________ ______________________
| | | |
| | | |
| (MASTER BREAKERS) | | (MASTER BREAKERS) |
| | | |
| |----| |----| | | |----| |----| |
| 6A | | | | | | 6A | | 5B | | | | | | 5B |
| |----| |----| | | |----| |----| |
| | | |
| |----| |----| | | |----| |----| |
| 1B | | | | | | 1B | | 6B | | | | | | 6B |
| |----| |----| | | |----| |----| |
| | | |
| |----| |----| | | |----| |----| |
| 2B | | | | | | 2B | | 1C | | | | | | 1C |
| |----| |----| | | |----| |----| |
| | | |
| |----| |----| | | |----| |----| |
| 3B | | | | | | 3B | | 2C | | | | | | 2C |
| |----| |----| | | |----| |----| |
| | | |
| |----| |----| | | |----| |----| |
| 4B | | | | | | 4B | | 3C | | | | | | 3C |
| |----| |----| | | |----| |----| |
| | | |
| | | |
------------------------ ------------------------
So my 2B apartment shares a common neutral with
apartments 6A, 1B, 3B, and 4B, while my 2C apartment
shares a common neutral with 5B, 6B, 1C and 3C.
6) This make some sense given our results: although I
have noise coming in from other apartments or elsewhere
in the neighborhood, the noise levels differ on each
phase of each service panel. My assumption is that 2B
is seeing noise generated mostly by 6A through 4B and
that 2C is seeing noise generated mostly by 5B through
3C.
7) For what it's worth: the 2B master shutoff breakers are
50 amps and the 2C master shutoff breakers are 40 amps.
That describes the environment. Now for our filtering attempts:
At first we tried installing PZZ01 filters in the 2B and 2C
service panels, with some success but not enough: noise
levels and random lighting malfunctions were reduced but
still present. Also, controller signals generated on a 2C
line were still able to operate devices on the 2B panel,
suggesting that these filters aren't powerful enough for my
application. (Not that I want to use two separate
controllers in one home, but that's the nature of the
beast. I don't want a downstairs neighbor controlling my
lights, either.)
I now have Leviton 6285 signal attenuators installed at
each service panel, with greater success. We've had some
lights flicker on in the 2 days since installation, but
none flicker off. There seems to be good isolation between
the two service panels. I have not yet installed Leviton
HCA02-10E signal amplifiers, but I'm not sure we'll need
them: my ActiveHome Pro controller usually works well,
while noise levels are usually quiet. But I'm metering
1 - 2 hours each night with high noise levels and the
ActiveHome Pro signals can't get through (like the old
paradigm of whispering in a crowded restaurant).
Here's the weird part: during these high noise periods, the
Leviton 6285 seems to be AMPLIFYING the noise, not
attenuating it. When I switch its breakers off, the noise
level drops.
During the other 14 or so hours of the waking day, the 6285
is indeed reducing noise levels: when we switch the
breakers off, the noise level jumps, as we would expect.
Also add to the mix: last night I discovered that one item
in the apartment -- an old PC monitor -- is indeed adding
noise at random times to the system, whether it's on, off
or in standby. I pulled its plug.
My questions for the group:
1) Are there known situations or environments when a
Leviton 6285 can boost noise rather than attenuate it?
Is it ever known to be flaky like this (suggesting I
have a bad unit), or is this an indication that I
didn't install it properly?
2) Is the Leviton HCA02-10E likely to fix this? I would
think not: while an amplified controller code could
blast through some noise, the noise itself could
trigger device malfunctions. Or does the HCA02-10E
also attenuate noise?
3) Should the Leviton 6285 attenuate the noise generated
by the old computer monitor along with the noise coming
in from outside the apartment, or will I need a
separate plug-in filter in addition?
4) Is there a better way of addressing my noise problems?
Best wishes to all for a happy Thanksgiving weekend!
-- Alan
Robert Green
<stuff snipped>
> At first we tried installing PZZ01 filters in the 2B and 2C
> service panels, with some success but not enough: noise
> levels and random lighting malfunctions were reduced but
> still present. Also, controller signals generated on a 2C
> line were still able to operate devices on the 2B panel,
> suggesting that these filters aren't powerful enough for my
> application.
It could also suggest that there's a major screwup in the apartment wiring.
I've found that the older a building and the more remodeling it's seen, the
more likely this is the case. If the cross-linking involves you in the
electrical "networks" of several other tenants, you'll likely NEVER be able
to control a normal PLC installation to your satisfaction. You could be
plunged into a powerline nightmare whenever another tenant gets a new
laptop, UPS, shaver, battery charger, CFL, dimmer, etc.
If I were in your shoes, I would look to an RF solution that didn't care a
whit how dirty the powerline was. We've recently seen a "2 apartments
combined" scenario stump our resident X-10 expert, who was armed with both
an oscilloscope and a million man-ours of X-10 experience. That doesn't
bode well for you.
> (Not that I want to use two separate
> controllers in one home, but that's the nature of the
> beast. I don't want a downstairs neighbor controlling my
> lights, either.)
X-10 wise, you may have no other choice in an environment like yours, and
dual controllers induce all sorts of other problems that are difficult to
track down. I love X-10. In a stand-alone house with low tech neighbors,
it works just fine for me. In an environment where the basic house wiring
has "morphed" the way yours has (apartments merged) and where one is likely
to have neighbors into technical things that can "dirty" the powerline, I
would look into one of the new wireless mesh networks like Zigbee. X-10 is
decades old. The designers just couldn't foretell how much things would
change in 20+ years.
I don't know how much you spent on the basics and how much you've spent
trying to get it to work, but for me at about $1,000 I would throw in the
X-10 towel completely. You can't possibly keep your environment clean
unless you generate your own power and disconnect from the grid. Unless
you're thrilled at the possibility of your X-10 system being knocked out
every time a neighbor buys a new gizmo, you may wish to consider other
topologies.
> I now have Leviton 6285 signal attenuators installed at
> each service panel, with greater success. We've had some
> lights flicker on in the 2 days since installation, but
> none flicker off.
What kind of lights? Does the "flickering on" come shortly after they've
just been turned off?
> There seems to be good isolation between
> the two service panels.
Don't bet your life on it. Experienced sparkies treat comboed apartments
the way experienced cops approach domestic disputes. With great care. I'd
bet $20 that within an hour, armed only with a 3-wire plug in tester, I
could find places where two circuits are crossed. Sometimes it's only so
that the hall light switch can control lamps on both sides of the new,
single large apartment. Sometimes someone needs a neutral wire at a switch
leg and grabs one from another circuit. The older the building, the more
likely someone's done something non-standard.
> I have not yet installed Leviton
> HCA02-10E signal amplifiers,
I am not sure you can install two of those in a single installation without
serious problems. Better check with Leviton.
> but I'm not sure we'll need
> them: my ActiveHome Pro controller usually works well,
> while noise levels are usually quiet. But I'm metering
> 1 - 2 hours each night with high noise levels and the
> ActiveHome Pro signals can't get through (like the old
> paradigm of whispering in a crowded restaurant).
As I said, there's no guarantee you're truly isolated from your neighbors,
even with filters installed. If you see these noise levels during prime
time, it's likely someone's got an expensive toy that's souring the
powerline. It would be pretty unfriendly to demand he filter or remove it.
Consider how many new high tech toys are likely to be added to your building
power grid this Christmas alone and you may want to seriously rethink X-10.
> Here's the weird part: during these high noise periods, the
> Leviton 6285 seems to be AMPLIFYING the noise, not
> attenuating it. When I switch its breakers off, the noise
> level drops.
Huh? Drops where? If you turn it off, all circuits it was filtering should
be dead. Where are you plugging your meter in?
Without extensive signal tracing, it's very hard to locate a miswired
appliance that's been using that neutral instead of the correct one finds
another path to neutral through some incredible cluster of miswired
switches. Items that effect X-10 also interact with each other. My gut
tells me there's something very weird with your wiring.
> During the other 14 or so hours of the waking day, the 6285
> is indeed reducing noise levels: when we switch the
> breakers off, the noise level jumps, as we would expect.
How are you measuring the noise level? When you switch the breakers off,
the meter should die with them. Where is the noise level jumping when you
turn the breakers off?
> Also add to the mix: last night I discovered that one item
> in the apartment -- an old PC monitor -- is indeed adding
> noise at random times to the system, whether it's on, off
> or in standby. I pulled its plug.
Given that almost everyone has an X-10 killer somewhere and that your wiring
seems less than perfect, I'd go RF since pulling CAT5 to each load (the
preferred solution in noisy environments) is probably very hard in your
case. A good RF mesh network wouldn't care if you had 1000 old PCs nor
would it care if each outlet in your apartment came from a different
breaker.
> My questions for the group:
>
>
> 1) Are there known situations or environments when a
> Leviton 6285 can boost noise rather than attenuate it?
> Is it ever known to be flaky like this (suggesting I
> have a bad unit), or is this an indication that I
> didn't install it properly?
Tell us how you know it's boosting noise and we might be able to explain. A
panel filter should filter all outlets and switched connected downstream.
When you shut it off, you kill those outlets unless you've wired a shunt
that pulls the 6285 completely out of the circuit.
> 2) Is the Leviton HCA02-10E likely to fix this? I would
> think not: while an amplified controller code could
> blast through some noise, the noise itself could
> trigger device malfunctions. Or does the HCA02-10E
> also attenuate noise?
I'd say it would only make your problems worse.
> 3) Should the Leviton 6285 attenuate the noise generated
> by the old computer monitor along with the noise coming
> in from outside the apartment, or will I need a
> separate plug-in filter in addition?
The old monitor probably isn't noisy - it probably "sucks" the X-10 signals
up. They are two different sorts of problems although a filter can often
cure either. The answer to your question is "yes, you will need a separate
plug-in filter for every X-10 unfriendly device in your house." Think of
the breaker mounted filters as firewalls. They stop outside noise from
getting into your apartment's power grid. They stop your noisy (or signal
sucking) old monitors from affecting your *neighbors* X-10 system. They do
NOT stop any locally generated problems like your PC monitor from affecting
the rest of your system.
> 4) Is there a better way of addressing my noise problems?
ZigBee. It's a wireless mesh network in which each device hears and repeats
commands. It has an enormously higher data rate than X-10 and is nearly
immune to powerline quality issues.
Realistically, you could be just a few plug in filters away from fairly
reliable operation. You need to go to the breakers and take measurements of
the propagated X-10 signal at various points in the apartment and then do it
again while disabling each breaker in turn. Eventually, you'll turn a
breaker off and the X-10 signal will rise by almost a volt. Somewhere on
that circuit is a signal sucker or noise generator.
I'd also be checking each circuit for leak current when its breaker was
killed. In a case like yours I would also make a table of every outlet and
light switch and appliance and make sure I knew which breaker controlled
each and every one of them.
--
Bobby G.
Dave Houston
transmitters such as your CM15A, any two-way devices and any other
"tranceivers (e.g. RR501, TM751) are of particular interest.
In a brief search I was unable to find out much about the 6285 signal
attenuator but I doubt that it can amplify. None of the literature I found
on it gives a figure for how much it attenuates.
You use the phrase "noise level" but I don't see any reference to anything
that can actually measure signal or noise levels. The TesterLinc manual
indicates that it does not measure a level but only counts cycles within
some (unspecified) window. What quality counts do you see if you use the
CM15A to send PLC during relatively noise-free times? What quality counts do
you see during high noise times (without sending anything yourself)?
The TesterLinc also has a much narrower bandwidth (110-140kHz) than the
typical X-10 receiver although that doesn't appear to be a factor here.
What housecodes does the CM15A transceive?
You will need a plug-in filter for the noisy monitor - the 6285 only blocks
signal or noise from passing in or out of your apartment's wiring.
Has the building always contained apartments or has it been converted from
commercial use?
I'd love to know how Insteon would behave in this situation.
"AlanTinNYC" <AlanT...@aol.com> wrote:
>At first we tried installing PZZ01 filters in the 2B and 2C
>service panels, with some success but not enough: noise
>levels and random lighting malfunctions were reduced but
>still present. Also, controller signals generated on a 2C
>line were still able to operate devices on the 2B panel,
>suggesting that these filters aren't powerful enough for my
>application. (Not that I want to use two separate
>controllers in one home, but that's the nature of the beast.
>I don't want a downstairs neighbor controlling my lights,
>either.)
>
>I now have Leviton 6285 signal attenuators installed at each
>service panel, with greater success. We've had some lights
>flicker on in the 2 days since installation, but none
>flicker off. There seems to be good isolation between the
>two service panels. I have not yet installed Leviton
>HCA02-10E signal amplifiers, but I'm not sure we'll need
>them: my ActiveHome Pro controller usually works well, while
>noise levels are usually quiet. But I'm metering 1 - 2 hours
>each night with high noise levels and the ActiveHome Pro
>signals can't get through (like the old paradigm of
>whispering in a crowded restaurant).
>
>Here's the weird part: during these high noise periods, the
>Leviton 6285 seems to be AMPLIFYING the noise, not
>attenuating it. When I switch its breakers off, the noise
>level drops.
>
>During the other 14 or so hours of the waking day, the 6285
>is indeed reducing noise levels: when we switch the breakers
>off, the noise level jumps, as we would expect.
>
>Also add to the mix: last night I discovered that one item
>in the apartment -- an old PC monitor -- is indeed adding
>noise at random times to the system, whether it's on, off
>or in standby. I pulled its plug.
>
>
>
>My questions for the group:
>
>1) Are there known situations or environments when a
> Leviton 6285 can boost noise rather than attenuate it?
> Is it ever known to be flaky like this (suggesting I
> have a bad unit), or is this an indication that I didn't
> install it properly?
>
>2) Is the Leviton HCA02-10E likely to fix this? I would
> think not: while an amplified controller code could
> blast through some noise, the noise itself could trigger
> device malfunctions. Or does the HCA02-10E also
> attenuate noise?
>
>3) Should the Leviton 6285 attenuate the noise generated by
> the old computer monitor along with the noise coming in
> from outside the apartment, or will I need a separate
> plug-in filter in addition?
>
>4) Is there a better way of addressing my noise problems?
>
>
>
Dave Houston
environment.
My mystery signal disappeared after being here 24/7 for about a week.
tparent
(neightbour RF or PLC), I discovered in my case that most of my noise
problem are coming from fluorescent bulb.
Like already explained, I'm talking about interference meaning that one
specific X10 command become an unpredictable result (issuing D10 ON,
result is D10 ON and D12 OFF at the same time). Sometime X10 device are
triggered without any issued X10 command (often lamp module).
These kind of strange behaviour is not detectable by computer interface
or dumpable through PC software.
Most of the time, "uncontrollable device" are basic X10 type (LM465,
AM..), but it's not systematic (less often Smarthome or ADS could be
hit).
After hours trying to guess what really happens with a digital
oscilloscope, I have the following feeling (I'm not a pro, just an
enthusiast) :
- Fluorescent bulb generate at zero crossing a current spike around
0.3V.
- This spike doesn't seems to be 120khz but are not filtered by LM465
electronic (as a consequence, spike is 'readable" on chip pin 1. In
specific condition (were the signal parasite match with X10 command,
the original command is tampered by fluorescent bulb spike".
-by the fact that spike are "low current", they are unable to be spread
on whole installation (attenuation will kill il before reaching
distribution panel).
On the basis of my test, I believe that basic X10 module doesn't
"stricto sensu" follow X10 protocol for the following reasons :
- they are not waiting the expected delay with 120khz on the line to
consider the signal valid (spike is enough to be considered valid).
-they are not matching the first and second copy of the X10 command
(only one is analysed and interpreted). In one of my room, two module
on the same address are triggered with a delay between the two. My
explanation is, the first one is on "good line" the second on a line
which require "signal repeater" to be listenable. First is directly
triggered on the first copy of X10, second is triggered by the signal
repeater amplified command with small delay.
It's very difficult to provide solution (excluding removing all
Fluorescent bulb).
Just few good practice :
- Plug your "X10 sending" device on a dedicate line directly wired to
the distribution panel.
- If it happens with one fluorescent bulb brand, try another one. On
basis of scope analysis, they are all noisy but some are producing
twice time bigger interference than others.
- replace "basic X10" by more reliable.
Hope it helps
(happy to be back "on the X10 battlefield",discovered that my new ISP
is not forwarding my replies on usenet since 6 month).
Thierry
Dave Houston
http://www.x10.com/support/x10trou.htm#on
Steven Bloom has a 'fix' for certain switches.
http://groups.google.com/group/comp.home.automation/msg/62d41d8a36fef5ea?oe=UTF-8&output=gplain
Some modules are less susceptible than others. I had problems with a
screw-in module (the older dimmable one) turning off when I turned a
flurescent light off. I replaced the screw-in module with an LM14A and npo
longer have the problem.
However, it's unlikely that such switching transients would show up as
"120kHz Activity" on his TesterLinc.
AlanTinNYC
1) I'm not an electrician or electrical engineer, but
I am an audio visual engineer working in Manhattan
for 20+ years now. My gut feeling when I saw
Robert Green's RF recommendation is that in this
city I don't trust wireless *anything* -- too many
glitches in both old, tested technologies (wireless
microphones) to new ones (Nextel GPRS) . We just
have too much traffic to keep a local RF system
reliable.
I thank Dave Houston for concurring on that, but
he asks about Insteon. Isn't Insteon half RF-
based? That's why I opted for a more "tried and
true" X10 solution; we even tested an X10 dimmer
in our old apartment four blocks away in the same
neighborhood and had flawless results for two
months. I've since learned the hard way that the
two buildings are not analogous. :-(
The other point about Insteon, I think, is that
it continues sending a command until the receiving
device acknowledges it. That might mean that
during a long noisy period we could see the
command execute a few hours after it was given.
But I'm speculating here; I have no hands-on
experience with Insteon.
2) I simplified the description of my apartment's
environment in my last post ("That was SIMPLE,
Alan???") It's not just two adjoining apartments
with a wall knocked down but in fact underwent a
gut renovation. Every old branch circuit was
severed, new lines were run through new walls,
with new switches, outlets, etc., and one of the
service panels was upgraded. I think the mains
wiring from the basement cutoff breakers may date
from 1941 but I doubt it; the feeder cable seems
to be PVC insulated.
So if we are cross-linked with any neighbors, this
would happen before service reaches our apartment,
as all our branch circuit wiring is new.
But if we are indeed cross-linked with our
neighbors before the service enters our panels,
wouldn't a whole-house block or attenuator stop
or greatly reduce that? Otherwise, what purpose
do those devices serve???
3) When I speak of our lights randomly flickering on
or off (I chose the wrong word; they don't
"flicker" but do switch on or off); I did mean
RANDOM: there is no recognizable time or
pattern, i.e., they don't turn off right after
going on, or vice versa. It just happens when it
happens; most of the time I can run over to the
TesterLinc and see high 120kHz activity but no
valid X10 codes (lots of bad blocks, though).
4) When I wrote of "good isolation" between our two
service panels, I used the signal generated by our
Active Home Pro (CM15A) and the TesterLinc
meter: without the attenuators now in place,
signals generated on one phase of one service
panel could be detected by devices on both phases
of both service panels. This was the case even
with a PZZ01 installed in each panel.
But with the 6285 installed, I could not operate
devices or meter anything on the other panel from
the one the CM15A was plugged into. We tested
both panels that way, and they do appear to be
isolated from each other.
5) My method of measuring noise levels: using
Mode 4 of the TesterLinc ("120kHz Activity"), I
can see Quality Counts of around 50, or 80, or 120
at various noisy times. This usually does not
change as I switch branch breakers off. The one
time I saw the Quality Count drop when I switched
off a branch circuit, I checked everything
plugged into the branch. That's how I found the
noisy PC monitor the other night.
I plug the meter into an outlet with no other
resistors (appliances, x10 devices, etc) on that
branch. Of course, when I shut that breaker, the
meter goes dead. But there is nothing else on
that branch that would generate any signal or
noise.
As per my electrician's and my understanding of
Leviton's written instructions, the filter is
installed on a separate pair of 15 amp circuit
breakers. So when I switch the filter "off", the
other branches stay on, including the TesterLinc's
meter, but the filter's branches (one for each
phase) are switched off. During relatively quiet
times, the Quality Count may jump from 000 or 001
to the 20s, 30s or higher when I switch the filter
off. That's how it's supposed to work. But
during peak noisy times, switching off the
breakers feeding the 6285 cuts the Quality Count
from around 120 down to around 60. That's
opposite of how it's supposed to work.
But based on feedback I'm getting from Robert
Green and Dave Houston, I suspect I don't have my
6285s installed properly. (See below, very bottom
of this post.)
6) My only controller device at this time is the
CM15A. When it transmits codes, I see Quality
Counts of around 50 - 60. That's at the low end
of the Quality Count during our noisy evenings, as
I mentioned above (count can be from the 50s to
the 120s).
All of my dimmers are X10 RSW17 or companion RSW19
slave switches. No other modules at this time,
but if this goes well I will add 2 or 3 plug-in
dimmers -- all for incandescent lights (no
fluorescents, CFLs, halogen low voltage
transformers, etc.)
We are at this time transceiving only House Code
L. We switched from House Code G within the first
week of our problems, before we bought the tester
and started serious troubleshooting.
7) Regarding using multiple HCA02-10E amplifiers in a
single home: Leviton in fact specifies that if
there are sub-panels, an amplifier is required at
each one. So I don't think that separate
HCA02-10Es at each of two main panels would be a
problem, but I'm not spending anything on them
until we deem them necessary.
8) This whole building has been solely residential
since it was built. Not even doctors' offices
on the ground floor or commercial storefronts.
I think that addresses the comments & questions posed so far. But both
Robert Green and Dave Houston mention stuff that bring me to doubt my
filter installation:
Following Leviton's instructions, the filter is installed on two poles
of a 15 amp breakers:
----------------------------------------------------
| |
| | | | |
| L1| L2| |N |
| | | | |
| ------------- ------------- | |
| | | | | | |
| | BREAKER A | | BREAKER B | | |
| | (1) | | (2) | | |
| ------------- ------------- | |
| | | | |-----| |----| |
| | BREAKER B | | BREAKER A | -------- | |
| | (3) | | (4) | | | | |
| ------------- ------------- | 6285 | | |
| | | | | | | | |
| | BREAKER A | | BREAKER B | -------- | |
| | (5) | | (6) |-----| |----| |
| ------------- ------------- | |
| | | | | | |
| | BREAKER B | | BREAKER A | | |
| | (7) | | (8) | | |
| ------------- ------------- | |
| | (etc.) | | (etc.) | | |
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
This is how mine are installed. At first I couldn't see how this would
filter *all* incoming noise, i.e., the branch circuits on breakers 4
and 6 are filtered, but wouldn't the noise bypass the filter on the
other branch circuits? (I mentioned earlier I'm an audio visual
engineer. If we apply a bandwidth filter or echo canceler to a channel
of a mixing board, the other channels remain unfiltered. To filter the
whole board we would need to filter its line input or mixed output.)
But I'm assuming the instructions are correct and the 6285 functions as
a big "signal sucker" in close proximity to the other branches. In
addition, the instructions specify that the filter be wired to a 15 amp
double-pole breaker (or two single pole breakers) to meet NEC spec.
Doesn't that mean it must be on its own branch circuits?
And wouldn't this mean that if there is noise generated on one of my
branches internally (like my old PC monitor), the 6285 would suck it
up?
But Robert points out that if I shut off the filter, "all circuits it
was filtering should be dead." And Dave points out that "the 6285 only
blocks signal or noise from passing in or out of your apartment's
wiring."
This suggests that the 6285 should be installed prior to the branch
circuits, not in parallel with them. But that would require something
far greater than a 15 amp breaker.
Does anyone have experience with this filter and can advise me on its
correct installation? Or do I have it installed correctly? If it's
installed correctly, why would my "noise level", defined as a Quality
Count on a TesterLinc, sometimes double when breakers 4 & 6 (in the
above illustration) are switched off, and at all other times be reduced
to almost zero when the same breakers are switched off?
Thanks again, and regards to all.
-- Alan T. in NYC
Dave Houston
detailed description. SmartHome's and Leviton's descriptions imply that it
blocks signals from passing but the connection scheme doesn't appear to
support that. If it merely attenuates 120kHz there is no way it can
distinguish between your valid signals and external "noise" (i.e. a rose is
a rose is a rose).
http://www.leviton.com/pdfs/dhctechman.pdf
Insteon only uses RF to couple the two phases. That might cause problems in
your environment (especially since they use a fixed frequency) but if it
does not, it's other features would likely be highly beneficial in your
situation. It will ignore signals from devices which you have not "enrolled"
so external Insteon signals would not be a problem. It has good noise
immunity. It will not resend forever but only for a fixed number of tries.
The signal packet duration is brief. If you have time, read the Insteon
whitepaper on the Insteon site.
If you haven't already invested heavily in X-10 it would be an ideal time to
investigate Insteon. (more below)
You are misinterpreting the TesterLinc output. It does not measure the
amplitude (i.e. level) of a signal but merely counts cycles within an
unspecified window just after zero crossing (ZC). X-10 says the receive
window extends from ZC+250猶-ZC+900猶 and that 48 cycles in that window
constitutes a logical Manchester 1. For transmitters they recommend the
120kHz burst begin as soon as possible after ZC but no later than ZC+100猶
and continue to ZC+1ms. That means there should be anywhere from 108-120
cycles in the 1ms following ZC with up to 78 cycles within the 650猶 receive
window. SmartHome does not specify where the TesterLinc window starts/ends.
Nor do they specify the sensitivity (i.e. minimum amplitude). I don't think
there is anyway to correctly interpret its output without knowing those
details.
I would suggest you get a loaner ESM1 from Martin Custer at AutomatedOutlet.
It will tell you the amplitude of the signals and noise. While its bandwidth
is not quite as wide as older X-10 receivers it is much wider than that of
the TesterLinc.
http://www.automatedoutlet.com/product.php?productid=463&cat=0&page=1
I also suggest that you email or phone Martin and ask if you can return an
Insteon starter kit if it turns out that it will not work in your
environment. If Insteon can make it there...
Dave Houston
You are also misunderstanding some other key things. (You have lots of
company.)
Aside from your signal attenuator/blocker the only thing that blocks the
passage of 120kHz is the utility company's transformer. This means that
using X-10 is a bit like having unprotected sex.
The wiring and breaker panels are bidirectional so any 120kHz that
originates within your apartment can reach all other apartments that share
the transformer and any 120kHz that originates in the other apartments can
reach your apartment. This is true whether the 120kHz is from others using
X-10 or is noise from their appliances and devices.
Add to this that should a neighbor be using X-10 RF on a housecode that your
CM15A is set to transceive, the CM15A will neither know nor care that the RF
is of alien origin and will blithely send the PLC codes. (If you are not
using RF, set the CM15A to transceive no housecodes.)
Essentially, you need two whole house blocking couplers (one at each panel).
Then you will need a coupler between the two panels. Then you probably will
still need to avoid X-10 RF.
How close are the two panels?
Robert Green
news:438abdb8...@nntp.fuse.net...
> Any RF system is likely to be problematic in a densely inhabited urban
> environment.
>
> My mystery signal disappeared after being here 24/7 for about a week.
It would be interesting to see how ZigBee fares in this near worst-case
scenario. I can't see it being any *more* problematic than X-10 will likely
be. I can't imagine something as robust as an RF mesh network working worse
than primitive and slow X-10 in a dense urban multi-unit dwelling. The
repeater function of mesh networks *should* eliminate many of the problems
that X-10 faces in problem environments.
How would you rate ZigBee's RF system v. X-10's in terms of robustness? By
that I mean reliability of communication and rejection of interference.
When you press J5 on, it actually GOES on and never changes state without
direct command?
The OP's apartment building more than likely was built many, many years
before the average household electrical consumption tripled or even
quadrupled. IIRC, household energy use tripled from 1949 to 2000. In old,
large buildings, there have probably been numerous upgrades to building
wiring. I'll bet that some were done more professionally than others, even
with NYC's rigorous code and inspection requirements. As you pointed out,
some buildings have gone from low per occupant electrical consumption
"office" to high per occupant "upper class residential."
I've found the further you get from "new" the harder X-10 seems to be to
implement because of the oddball home-brew fixes we see here where someone
pulled a neutral from the wrong circuit to power a switch. Of course, YMMV.
I guess I'd take my chances on a robust, low power RF protocol over an
undocumented local power grid that's connected to God knows what. It would,
indeed, be interesting to see how both ZigBee AND Insteon do in such a
situation. Surely with all your contacts, Dave, you could get someone to
treat the OP to a head to head test of protocols using loaner equipment?
If ZigBee and mesh networks are all they are touted to be, they should
trounce X-10 in the OP's environment. What an ideal "test case."
As for your mystery signal disappearing I am not sure whether that's
comforting or alarming. Where did it come from? Where did it go?
--
Bobby G.
Dave Houston
Crestron has some ZigBee based devices but they are primarily a hardwired
system and, in the past, have only used RF as an input (i.e. RF gateways). I
suspect they intend the ZigBee devices to augment the hardwired system.
Control 4 has some ZigBee switches but they also appear to be limited.
If you haven't done so you should read the RF whitepaper I cited about a
month ago from Echelon. While they have a vested interest in making ZigBee
and Z-Wave seem questionable, the paper is excellent and the points made
have been made elsewhere by academics and others with no apparent vested
interest.
Also, while the underlying ZigBee RF protocol is an open standard, it is
likely that the higher level communications protocols will be proprietary
and xenophobic.
Z-Wave doesn't appear to be making much market penetration.
I don't have many contacts but were I the Insteon marketing manager at
SmartHome I'd have an engineer on his way to Manhattan before lunchtime.
Alan's apartment would be the ideal test bed for all of Insteon's features.
(It might also convince them they need multiple RF channels to avoid
interference.)
Robert Green
> Thank you all for your guidance. To answer some questions/comments:
>
> 1) I'm not an electrician or electrical engineer, but
> I am an audio visual engineer working in Manhattan
> for 20+ years now. My gut feeling when I saw
> Robert Green's RF recommendation is that in this
> city I don't trust wireless *anything* -- too many
> glitches in both old, tested technologies (wireless
> microphones) to new ones (Nextel GPRS) . We just
> have too much traffic to keep a local RF system
> reliable.
The mesh network protocol is remarkably advanced and hasn't very much in
common with real-time audio transmission protocols. Each device is a
repeater and the data rates are high enough so that the time lag for
repeated commands isn't really an issue. I asked Dave to comment on the
robustness of the current X-10 RF scheme compared to a ZigBee-type mesh
network because he's far more competent to explain the difference.
> I thank Dave Houston for concurring on that, but
> he asks about Insteon. Isn't Insteon half RF-
> based? That's why I opted for a more "tried and
> true" X10 solution; we even tested an X10 dimmer
> in our old apartment four blocks away in the same
> neighborhood and had flawless results for two
> months. I've since learned the hard way that the
> two buildings are not analogous. :-(
> The other point about Insteon, I think, is that
> it continues sending a command until the receiving
> device acknowledges it. That might mean that
> during a long noisy period we could see the
> command execute a few hours after it was given.
> But I'm speculating here; I have no hands-on
> experience with Insteon.
Someone else would have to comment on that. I've not been interested enough
in Insteon to look very far into it.
>
> 2) I simplified the description of my apartment's
> environment in my last post ("That was SIMPLE,
> Alan???") It's not just two adjoining apartments
> with a wall knocked down but in fact underwent a
> gut renovation. Every old branch circuit was
> severed, new lines were run through new walls,
> with new switches, outlets, etc., and one of the
> service panels was upgraded. I think the mains
> wiring from the basement cutoff breakers may date
> from 1941 but I doubt it; the feeder cable seems
> to be PVC insulated.
That encompasses several generations of sparkies. You can bet there are
some interesting details to the electrical plan.
> So if we are cross-linked with any neighbors, this
> would happen before service reaches our apartment,
> as all our branch circuit wiring is new.
I believe Dave already pointed out, X-10 is a little like unprotected sex.
You *are* linked to every other apartment in the building UNLESS you install
a whole house blocker at the service panel that sits between your apt.
wiring and the rest of the building's wiring. Gotta be rated at the max
rating of your panel and probably installed in its own box. I've never
needed one but some people here use them and will hopefully chime in.
> But if we are indeed cross-linked with our
> neighbors before the service enters our panels,
> wouldn't a whole-house block or attenuator stop
> or greatly reduce that? Otherwise, what purpose
> do those devices serve???
They are supposed to insulate you from the rest of the world in both
directions. Some vendors call it an X-10 firewall.
> 3) When I speak of our lights randomly flickering on
> or off (I chose the wrong word; they don't
> "flicker" but do switch on or off); I did mean
> RANDOM: there is no recognizable time or
> pattern, i.e., they don't turn off right after
> going on, or vice versa. It just happens when it
> happens; most of the time I can run over to the
> TesterLinc and see high 120kHz activity but no
> valid X10 codes (lots of bad blocks, though).
Not good. Indicates your lines are noisy enough to randomly trip the
modules. Forgive me if I plug RF, but the newer technologies have an
incredible amount of redundancy and wouldn't respond to a noise-generated
signal. Worst case, a stronger transmitter might block legit transmissions
but should rarely, if ever, response to noise as a legit command. That's
because they operate so much faster that there is much more room for data
checking and security in each command interval.
Dave's covered most of this. Get an ESM-1 meter. It will give you more
useful information than quality counts.
> 6) My only controller device at this time is the
> CM15A. When it transmits codes, I see Quality
> Counts of around 50 - 60. That's at the low end
> of the Quality Count during our noisy evenings, as
> I mentioned above (count can be from the 50s to
> the 120s).
>
> All of my dimmers are X10 RSW17 or companion RSW19
> slave switches. No other modules at this time,
> but if this goes well I will add 2 or 3 plug-in
> dimmers -- all for incandescent lights (no
> fluorescents, CFLs, halogen low voltage
> transformers, etc.)
X-10 unfriendly devices are insidious. Almost anything from a shaver
charger to an aquarium heater can mess up X-10. From what you've been
describing, you still have some more signal suckers in your house - you just
haven't found them yet. A good meter is essential. Signal suckers act
cumulatively. A lot of small ones can really impede the signal but they are
hard to find unless you can see their effect on a good meter or a scope.
>
> We are at this time transceiving only House Code
> L. We switched from House Code G within the first
> week of our problems, before we bought the tester
> and started serious troubleshooting.
>
> 7) Regarding using multiple HCA02-10E amplifiers in a
> single home: Leviton in fact specifies that if
> there are sub-panels, an amplifier is required at
> each one. So I don't think that separate
> HCA02-10Es at each of two main panels would be a
> problem, but I'm not spending anything on them
> until we deem them necessary.
I thought you didn't have subpanels but two equivalent panels feeding one
combined apartment. Dave will correct me if I am wrong but I think it's
possible that twin panels will have phase issues where subpanels won't. I
disconnected my HCA02-10E shortly after installing it. Too many "broadcast
storms" of unknown origin.
I would start with a good meter (I use the Monterey because it's got a
digital readout but it's incredibly pricey). Then I would set up a standard
reference - hopefully with a maxicontroller and a lamp module. I would
start at the ends of two branch circuits and read the signal strength.
Then, turn off all other breakers and read the signal again. If it's about
a volt higher, you then turn on the other breakers one by one until you find
the one that's sucking the signal.
> 8) This whole building has been solely residential
> since it was built. Not even doctors' offices
> on the ground floor or commercial storefronts.
But it sounds like it's nearly 70 years old. Think of how few electrical
appliances there were back then.
Yes. It would filter only what on that branch circuit. It's not what you
want.
> But I'm assuming the instructions are correct and the 6285 functions as
> a big "signal sucker" in close proximity to the other branches. In
> addition, the instructions specify that the filter be wired to a 15 amp
> double-pole breaker (or two single pole breakers) to meet NEC spec.
> Doesn't that mean it must be on its own branch circuits?
>
> And wouldn't this mean that if there is noise generated on one of my
> branches internally (like my old PC monitor), the 6285 would suck it
> up?
>
> But Robert points out that if I shut off the filter, "all circuits it
> was filtering should be dead." And Dave points out that "the 6285 only
> blocks signal or noise from passing in or out of your apartment's
> wiring."
>
> This suggests that the 6285 should be installed prior to the branch
> circuits, not in parallel with them. But that would require something
> far greater than a 15 amp breaker.
Ubetcha. Smarthome says:
"Designed for installations where the Whole-House Blocking Coupler or X10
Pro Whole-House Blocking Coupler can't be used, the Signal Attenuator
reduces the signal strength of external X10 signals. Other outside noise
sources, such as baby monitors, that are on a similar frequency to X10 will
also be reduced."
Why aren't you using:
http://www.smarthome.com/4851.html
# Blocks X10 signals from entering the home
# Couples X10 signals across 2 and 3 phase systems
(Image says good for 225A)
--
Bobby G.
Dave Houston
What has happened with your problem? Did we confuse you too much?
carlford
I too am interested in your two panel setup as I have a similar
discussion going on in another post. Did you give up or did you try
the 2 HCA02-10E's ???
Good luck
Carl Ford
AlanTinNYC
down with a minor illness, so I couldn't follow up for a while.
After considering everything, and after both my electrician and I tried
to discuss this with Leviton on the phone (their expert was a bit
rude), I think the next step is to write to Leviton (I have a
friendlier tech's e-mail address) including some of the detail I wrote
to this group and ask for a better explanation or description of their
product.
Our lights are no longer going "crazy." They do occasionally shut down,
about once an evening, not three times an hour, and they no longer
spontaneously switch on. But they are not yet steadily controlled by
the CM15A, probably because we do have the attenuators installed
(improperly, I fear) and don't yet have the HCA02-10E's installed.
Much of the time, the controller works fine, but when I'm metering
sustained, stray 120kHz activity it doesn't work.
Considering everything so far...
1) Yes, I do want/need a "firewall" so to speak, although I understand
such a device would not digitally block all unwanted stuff (as a
computer network firewall would) but attenuate it by absorbing it (as
does the 6825) or phase-cancel it (as does the 6824 or PZZ01) down to
tolerable levels.
2) Acknowledging that "X-10 is decades old. The designers just
couldn't foretell how much things would change in 20+ years...," then
unless the incoming noise levels are EXTREMELY high (greater than the
30:1 attenuation provided by the Leviton filters), wouldn't the proper
application of hard-wired filters on my panels combined with plug-in
filters on my own noisy appliances protect the network against this,
without a need to "generate your own power and disconnect from the
grid"?
3) I'm just too wary of RF solutions in the Big City. And what is
likely to happen to any RF network's integrity in a few years? Didn't
we all love our 2.4GHz cordless phones until wireless home network
routers came out? I don't use the RF function in the CM15A; I just use
it as a whole house Sabbath timer (well, technically, half a house,
since I'm isolating my two service panels and will ultimately use two
controllers). I have the RF function on my CM15A shut off except
during testing.
Or should I trash my PLC system, go with RF and plan on trashing *any*
home control network every 10 years or so because we "just can't
foretell how much things will change"?
I propose to this group that "new" isn't necessarily better, "old"
isn't necessarily archaic or obsolete so long as we have the network
integrity protected. That could mean building an effective PLC
firewall, or it could mean encasing my home in concrete and lead to
shield it against anything stray flying through the air. I prefer to
filter just the wires.
4) I just don't "get" how the 6285 attenuator is supposed to work. If
it is supposed to knock down everything flowing through my house at
around 120kHz, then why does my unamplified controller work during
non-noisy times? And if it installs on two 15 Amp branch breakers (as
per the written instructions), then how does it block my neighbor's
noise from flowing from the main feeds onto the other branches in my
panel? (Please see my earlier illustration.) Is it an effective
"signal sucker," sucking everything at the panel? If not, what
protects my WS12A dimmers on the other branches the 6285 isn't wired
to? If it is an effective signal sucker, why wouldn't it take care of
my noisy PC monitor or other devices (every "laptop, UPS, shaver,
battery charger, CFL, dimmer, etc.") as well?
But I'm speaking hypothetically now. I'm not getting the 6285 to
perform consistently, so I think it's not properly -- or effectively --
installed.
When speaking with Leviton last week, the first tech said we should
connect the 6285 "as close to the incoming feed as possible." That
meant connecting it to two separate 15 amp breakers, each breaker the
first in line on its phase, and suggesting that this could block
incoming noise from trickling to other branch circuits. But then the
"specialist" who took over the call practically barked, "No, I don't
want this on two separate breakers. I want this on a single,
double-pole breaker!" So much for placing the filter as close to the
incoming feed as possible... Time to write to Leviton.
5) Before I sink in an additional $85 for each of two HCA02-10Es (and
possibly knock holes in my plaster & lath walls for the boxes to house
them), I want to know that they will be effective. Does it boost all
120kHz activity? Won't that put me back where I started? Does it
"recognize" clean X10 code separate from noise? How? What happens
when a clean X10 packet collides with noise? We already know that X10
won't work when two good packets collide. While I can get around
some packet collisions by programming the CM15A to repeat commands,
what do I do to get around 3 - 4 hours of sustained noise?
Can't one just amplify the output of a controller before it plugs
into the powerline? That would introduce X10 signal at several volts
onto a powerline network where all 120kHz activity is then attenuated
by a whole house filter. Does anyone make such a feed-thru amplifier?
That's where things stand. Given my workload and the upcoming
holidays, I don't think I'll be getting to this until 2006.
Merry/Happy whatever-you-celebrate, everybody!
-- Alan
Dave Houston
upstream of the 2B-2C split. Are there separate feeds from the basement to
the 2B & 2C panels?
I'm also a bit puzzled by how the 6825 works. It would really help to see a
schematic. The effects of a signal sink are somewhat localized because of
the impedance of the wiring itself but I don't quite understand how a 6825
can attenuate alien signals without also attenuating legitimate signals
trying to go from 2B to 2C.
My earlier recommendation of two HCA02-10Es was based on...
http://www.worthingtonsolutions.com/activekb/index.php?ToDo=view&questId=25&catId=5
Two ACT CR234s are probably a better choice because of the SCC feature.
With X-10, one other thing might help. The LM14A lamp modules and (I
suspect) the Leviton switches that respond to extended codes do a better job
in the presence of low level 120kHz noise than do LM465 lamp modules and
X-10 made switches. However, we still don't know the amplitude of the 120kHz
noise you are seeing.
The repeating couplers do not repeat all 120kHz and they don't really
amplify (at least those made by ACT don't) but merely repeat the X-10 signal
during the time frame for the second copy of the X-10 command.
I still think this might be a situation where Insteon would work far better
than X-10. It might be unaffected by the 120kHz noise. You only need RF
communication between the RF links which act to bridge the phases. You
should be able to locate these near each other, perhaps even putting them
inside a Faraday cage to prevent external RF from jamming them. You may need
four RF links.
Dan Lanciani
| I'm also a bit puzzled by how the 6825 works. It would really help to see a
| schematic.
I don't believe there is a 6825, but the 6285 (which I think is the one
being discussed) is very simple. It has two 1uF capacitors (each with
a small protective(?) series resistor), one from each line to neutral.
|The effects of a signal sink are somewhat localized because of
|the impedance of the wiring itself but I don't quite understand how a 6825
|can attenuate alien signals without also attenuating legitimate signals
|trying to go from 2B to 2C.
It can't.
Something I always thought would be neat (and I proposed this to ACT) is
a repeater that bridges a low-impedance onto the line during the transmission
of a "0" (rather than doing nothing). This would provide an active solution
to noise problems. When my last CR230 failed with the dead high-voltage power
supply it managed to block transmissions rather than repeat them, so it might
not even require significant changes to the design. (I.e., a good coupling
network for 120kHz probably also makes a good sink if shorted rather than
driven.)
Dan Lanciani
ddl@danlan.*com
Dave Houston
>I don't believe there is a 6825, but the 6285 (which I think is the one
>being discussed) is very simple. It has two 1uF capacitors (each with
>a small protective(?) series resistor), one from each line to neutral.
I realized I had transposed the 2 & 8 about a 猶 after clicking the send
button.
>|The effects of a signal sink are somewhat localized because of
>|the impedance of the wiring itself but I don't quite understand how a 6825
>|can attenuate alien signals without also attenuating legitimate signals
>|trying to go from 2B to 2C.
>
>It can't.
Then if the 6285 really will cause as much trouble for interpanel signals as
it will for alien signals, I don't see any advantage to using them here.
Marc F Hult
message <43a21592...@nntp.fuse.net>:
>The repeating couplers do not repeat all 120kHz and they don't really
>amplify (at least those made by ACT don't) but merely repeat the X-10 signal
>during the time frame for the second copy of the X-10 command.
An ACT 230 is "sensitive to 80 millivolts peak
to peak minimum" and outputs "6V peak to peak @ 5 ohms minimum".
http://www.act-solutions.com/pdfs/PCCSpecs/cr230_spec.pdf
So if it receives a signal at 80 millivolts and retransmits it at 6v it does in
fact amplify by 6000/80 = 75X = 37.5 db gain.
And an ACT CR234 which is "sensitive to 25 millivolts peak to
peak" and outputs "5V peak to peak @ 5 ohms" can amplify up to 5000/25 = 200X =
46db gain.
http://www.act-solutions.com/pdfs/PCCSpecs/cr234_spec.pdf
Same principle applies to Leviton's current and past units. For example, the
HCA02-10E is specifically identified as a "System Amplifier".
http://www.leviton-lin.com/catalog/BuildPage.aspx?BuildPageID=2120
...Marc
Marc_F_Hult
www.ECOntrol.org
Dan Lanciani
| ddl@danlan.*com (Dan Lanciani) wrote:
| >|the impedance of the wiring itself but I don't quite understand how a 6825
| >|can attenuate alien signals without also attenuating legitimate signals
| >|trying to go from 2B to 2C.
| >
| >It can't.
|
| Then if the 6285 really will cause as much trouble for interpanel signals as
| it will for alien signals, I don't see any advantage to using them here.
It's tricky to say without really knowing what's going on. If he is dealing
with fixed-threshold receivers and if the noise is above the threshold then
there is no choice but to reduce the noise(*), even if it is at the expense
of valid signals. It's one of those communications problems where the best
you can do is attenuate everything and increase the real signal to compensate.
Without some kind of AGC or automatic threshold adjustment in the receivers,
merely improving the signal to noise ratio doesn't help: the absolute noise
level is what matters.
(*) Ok, you do have the choice of modifying the receivers to change the
threshold. I did that with some generic X10 2-wire switches. There isn't
much headroom, so you have to enjoy fiddling...
Dan Lanciani
ddl@danlan.*com
Dave Houston
>It's tricky to say without really knowing what's going on. If he is dealing
>with fixed-threshold receivers and if the noise is above the threshold then
>there is no choice but to reduce the noise(*), even if it is at the expense
>of valid signals. It's one of those communications problems where the best
>you can do is attenuate everything and increase the real signal to compensate.
>Without some kind of AGC or automatic threshold adjustment in the receivers,
>merely improving the signal to noise ratio doesn't help: the absolute noise
>level is what matters.
Alan has told us that the 6285s block signals from either panel reaching the
other panel. Given that and your description of the 6285 internals I think
he's dealing with a noise source that's internal to the apartment.
Robert Green
<stuff snipped>
> I think the next step is to write to Leviton (I have a
> friendlier tech's e-mail address) including some of the detail I wrote
> to this group and ask for a better explanation or description of their
> product.
My experience with Leviton's tech support was SO awful I would buy *any*
competitor's product first. I got the impression that they would tell
customers anything just to get them off the line. As for answering tech
emails - well, it never happened. One of their employees appeared briefly
here a while back and confirmed a certain uniformity in their hiring
process. Rude certainly fits like a glove. Maybe you'll have better luck,
Bubee.
In your case, the situation is so complex that diagrams and explanations are
required. That's probably not going to happen with a Leviton phone tech. I
was never able to get an answer to my emails so I can't comment on their
online support other than to notice its total lack thereof.
> Our lights are no longer going "crazy." They do occasionally shut down,
> about once an evening, not three times an hour, and they no longer
> spontaneously switch on.
But you really don't know whether you have solved the problem or whether the
mystery noise source has shut down for some reason, correct? This is why I
think going with a repeater so soon is a bad idea. If there's noise so bad
that it makes X-10 devices respond unbidden, then it's likely the noise is
going to affect the repeater adversely, too. Besides, responding to a
second copy of a command is never as good as responding to the first one.
First off, it introduces a time delay - slight but noticeably. Secondly, it
defeats the idea of sending two copies for redundancy.
If you can possibly do it, clean up your own wiring of any signal suckers,
surge protectors, UPSs, switching power supplies for laptops, printers,
etc., or anything else likely to "soak up" or otherwise impede X-10 signals.
It means a lot of time at the breaker panel and running back and forth to
outlets, but the only way to find these problem devices is to switch off all
the breakers except the test transmitter and the meter and then turn the
breakers back on one by one and watching the meter.
> But they are not yet steadily controlled by
> the CM15A, probably because we do have the attenuators installed
> (improperly, I fear) and don't yet have the HCA02-10E's installed.
I'd been SO wary of adding more shmatteh to the system before I understood
the underlying issues. KISS. It works SO well.
> Much of the time, the controller works fine, but when I'm metering
> sustained, stray 120kHz activity it doesn't work.
You don't need two repeaters - at least not yet (especially hinky
HCA02-10E's), until your are SURE you have blocked your apartment wiring
from the rest of the world. Until you take that step, your powerlines are
at the mercy of whatever your neighbors are plugging in. No solution that
leaves your home "powerline grid" vulnerable to outside PLC signals is going
to keep the noise and perhaps other tenants X-10 signals from getting in.
That's about as plain as it gets.
> Considering everything so far...
>
> 1) Yes, I do want/need a "firewall" so to speak, although I understand
> such a device would not digitally block all unwanted stuff (as a
> computer network firewall would) but attenuate it by absorbing it (as
> does the 6825) or phase-cancel it (as does the 6824 or PZZ01) down to
> tolerable levels.
Yep.
> 2) Acknowledging that "X-10 is decades old. The designers just
> couldn't foretell how much things would change in 20+ years...," then
> unless the incoming noise levels are EXTREMELY high (greater than the
> 30:1 attenuation provided by the Leviton filters), wouldn't the proper
> application of hard-wired filters on my panels combined with plug-in
> filters on my own noisy appliances protect the network against this,
> without a need to "generate your own power and disconnect from the
> grid"?
It should, but that advice is mostly for single family homes. I still don't
know what your wiring looks like, and were it my apartment I would be
crawling all over it with an ESM-1, a Monterey PSLA and a fox and hound wire
tracer set to understand where EVERY outlet goes to and what the signal
attenuation looks like.
To do this I set up a TM751 in an outlet next to the circuit panel and then
send it continuous signals from a transmitter with a purposefully jammed
button. Then I go all over the house looking at the signal strength at
various outlets. From their, I get a pretty good idea of what branches are
problematic. Also, from there, I can shut off one leg of the house wiring
after another to find circuits that contain suckers or other X-10 signal
traps. A very long extension cord helps. You can plug the meter into almost
any outlet in the house with it and still read it back at the panel(s).
You'll probably be as surprised as I was to see how little voltage is lost
on a run of 100' of extension cord. It was a good reminder that the problem
is usually a few devices scattered throughout the home that are either
generating noise or otherwise interfering with X-10 signals.
> 3) I'm just too wary of RF solutions in the Big City. And what is
> likely to happen to any RF network's integrity in a few years? Didn't
> we all love our 2.4GHz cordless phones until wireless home network
> routers came out? I don't use the RF function in the CM15A; I just use
> it as a whole house Sabbath timer (well, technically, half a house,
> since I'm isolating my two service panels and will ultimately use two
> controllers). I have the RF function on my CM15A shut off except
> during testing.
Ever consider that God knows you're cheating by using an electronic shabbos
goy? Maybe that's why you are having so much tsores getting rid of your
electronic dybbuks! www.shabot6000.com
> Or should I trash my PLC system, go with RF and plan on trashing *any*
> home control network every 10 years or so because we "just can't
> foretell how much things will change"?
>
> I propose to this group that "new" isn't necessarily better, "old"
> isn't necessarily archaic or obsolete so long as we have the network
> integrity protected. That could mean building an effective PLC
> firewall, or it could mean encasing my home in concrete and lead to
> shield it against anything stray flying through the air. I prefer to
> filter just the wires.
Dave's Insteon suggestion sounded very good. You can phase couple via RF
inside Faraday cages, if necessary.
There's no certainty that you can eliminate the noise. Like Dave, I've had
periods where the ESM-1 shows high noise levels for days, and then nothing.
The only hope to straighten your mess out is to KISS and solve one problem
at a time.
Adding amplifiers and repeaters into a noisy environment isn't the right
solution from the start. If you are dedicated to using X-10, you'll have to
understand your apartment wiring completely. Are you near any radio or TV
broadcast towers. I recall reading about someone with a similar problem in
the archives. Apparently a nearby transmitter can put serious noise on the
line via induction.
> 4) I just don't "get" how the 6285 attenuator is supposed to work. If
> it is supposed to knock down everything flowing through my house at
> around 120kHz, then why does my unamplified controller work during
> non-noisy times? And if it installs on two 15 Amp branch breakers (as
> per the written instructions), then how does it block my neighbor's
> noise from flowing from the main feeds onto the other branches in my
> panel? (Please see my earlier illustration.) Is it an effective
> "signal sucker," sucking everything at the panel? If not, what
> protects my WS12A dimmers on the other branches the 6285 isn't wired
> to? If it is an effective signal sucker, why wouldn't it take care of
> my noisy PC monitor or other devices (every "laptop, UPS, shaver,
> battery charger, CFL, dimmer, etc.") as well?
>
> But I'm speaking hypothetically now. I'm not getting the 6285 to
> perform consistently, so I think it's not properly -- or effectively --
> installed.
>
> When speaking with Leviton last week, the first tech said we should
> connect the 6285 "as close to the incoming feed as possible." That
> meant connecting it to two separate 15 amp breakers, each breaker the
> first in line on its phase, and suggesting that this could block
> incoming noise from trickling to other branch circuits. But then the
> "specialist" who took over the call practically barked, "No, I don't
> want this on two separate breakers. I want this on a single,
> double-pole breaker!" So much for placing the filter as close to the
> incoming feed as possible... Time to write to Leviton.
That's only if you want your conflicting advice in writing. :-)
> 5) Before I sink in an additional $85 for each of two HCA02-10Es (and
> possibly knock holes in my plaster & lath walls for the boxes to house
> them), I want to know that they will be effective. Does it boost all
> 120kHz activity? Won't that put me back where I started? Does it
> "recognize" clean X10 code separate from noise? How? What happens
> when a clean X10 packet collides with noise? We already know that X10
> won't work when two good packets collide. While I can get around
> some packet collisions by programming the CM15A to repeat commands,
> what do I do to get around 3 - 4 hours of sustained noise?
You either have to filter it or find some other control topology.
> Can't one just amplify the output of a controller before it plugs
> into the powerline? That would introduce X10 signal at several volts
> onto a powerline network where all 120kHz activity is then attenuated
> by a whole house filter. Does anyone make such a feed-thru amplifier?
Dan and Dave have explained why that's not feasible.
>
>
> That's where things stand. Given my workload and the upcoming
> holidays, I don't think I'll be getting to this until 2006.
Optimist. Very few X-10 installations are *ever* finished. There's always
some bug or intermittent operation or failure to contend with. But it's the
cheapest thing on the market AND the most universal. We'll see if that
continues.
Good luck!
--
Bobby G.
Dan Lanciani
Alan did say in his initial posting that he had confirmed that the source
was external to his apartment in the usual way (i.e., by turning off all
his own breakers except for the one with the meter), but since he didn't
mention repeating the test with a different single breaker I suppose it's
possible that he was mistaken. As I said, it's tricky to say without really
knowing what is going on. I can only base my conclusions on the information
Alan has provided, taken at face value...
Dan Lanciani
ddl@danlan.*com