UPB, etc. WAS: Need some antenna advice...

[nobody@xxxxxxxxxxxx (Dave Houston)]

This has drifted rather far afield from the original thread.

I have one of the Leviton HCPRF transceivers but don't recall that it has a
built-in lamp module. At least not a dimmer - mine has a solid state relay.
It's RF range seems comparable to standard X-10 transceivers. Unfortunately,
there's no room within the case to add a connector for an external antenna
so I did not try to trace the circuit to see whether they are using the hot
side as electronic ground. Without knowing that I cannot make any antenna
recommendation.

There should be no collisions as the Leviton will back off and wait for a
clean line before it sends PLC. You will get duplicates of all signals. If
you have marginal coupling between phases you might see different results at
different times. If the Leviton can't "hear" the other transceiver, there
will be collisions.

Don't confuse the UPB PLC pulses with RF. They're different species.

IMO, AGC doesn't do much for PLC. AGC was invented for AM radio as a way to
maintain a steady volume. It does this by reducing gain as signal strenth
increases.

Automatic Threshold Control is slightly different. It raises the threshold
so that only stronger pulses are effective. If the signal is stronger than
the noise, the output is clean. The page I referenced illustrates this. The
comparator circuit is called a "data-slicer" because, in effect, it slices
off the bottom half of the pulses, omitting most of the noise and passing
only the cleaner top half along.

I'm not at all sure what UPB pulses are composed of. PCS specs say they are
capicitor discharges or, in other words, noise. The page you cited implies
they are something else entirely. The page you cited also spews a lot of
male bovine excrement about PLC noise. A bigger problem is the line filters
added to AV equipment, to meet FCC requirements, that attenuate PLC signals.

The data on the UPB tests by PCS would seem to indicate that they get most
of their reliability from having a much stronger pulse and from having a
two-way protocol so that any mangled signals are repeated. However, I
question the validity of a test that did not use actual dimmers, etc. Triac
noise looks a lot like their pulses.

Your ESM1 tests sound like collisions.

I think the Monterey is vastly overpriced. It falsely reports every 1110
sequence not followed by a valid X-10 sequence as a "bad start code" when
they are usually the results of collisions between signals offset by 1/2
cycle. The TesterLinc is cheaper. I think Bruce Robin posted comparisons of
all of the available testers.

This is a good choice for a "scope"...

     http://www.usb-instruments.com/documents/small_stingray.pdf

You should also get ACT's Scope-Test2 if you want to look at X-10 PLC
signals.

I doubt I'll do anything with UPB. The prices are too high. I get good
reliability with X-10.

"Robert Green" <ROBERT_GREEN1963@xxxxxxxxx> wrote:

>"Dave Houston" <ydobon@xxxxxxxxxxx> wrote in message
>
>>>The RSSI measurement of the receiver linear output made with the
>>>BX24-AHT wasn't really useful for making definitive comparisons
>>>between antennas because the RF receiver module had AGC which
>>>reduces the gain as the signal gets stronger. It was useful for fine
>>>tuning the receiver. Lacking expensive lab style equipment, the
>>>most reliable measurement is to compare range.
>
>I see.  Sort of like a VU meter on an AGC tape recorder always showing a
>strong signal as long as the source is within a certain range.
>
>As long as the distance tests are rigorous enough, they're what really
>counts to the end user.  It's just that in so many posts I see less than
>optimal testing procedures.  I would assume you'd have to take a least four
>different readings (one from each compass point) to make even a marginally
>useful range map.
>
>> I forgot that one of the pages on my web site has some screenshots that
>> might be helpful in understanding this.
>>
>>       http://www.mbx-usa.com/rf-noise.htm
>>
>
>Yes, thanks.  That's very informative and leads me to wonder how UPB handles
>noise. At the site above you wrote:
>
>"Closely spaced pulses affect the AGC and the capacitor on the comparator
>input differently than widely spaced pulses."
>
>What do UPB pulses, which (IIRC) are variably spaced, look like to an X-10
>transceiver?  It's alleged that UPB works with everything except  the Lutron
>Homeworx product lines.  I wonder why?  There's some interesting reading at:
>
>http://www.hometoys.com/htinews/jun04/articles/upb/upb.htm
>
>where they say: "Case Study: Web Mountain has installed UPB devices in
>approximately 15 homes. With the exception of one outlet in one large house,
>we installed the devices, powered them up and they worked. The one outlet
>that didn?t work was fixed by adding a phase coupler. No noise filters were
>needed and no troubleshooting was required. Only one of the 15 homes (>6500
>sq feet) required the UPB phase coupler."
>
>> I did make comparisons between antennas by running tests where the
>> antennas were placed in the "same" location and I sent signals from
>> certain locations about the building, always using the same Palmpad and
>> using the "same" locations. The locations were far enough away from the
>> antenna to guarantee a moderate signal level. From these it seemed clear
>> that the eggbeater was better but there are too many uncontrolled
>> factors to supply a definitive percentage of "betterness".
>
>Your test setup seems quite rigorous so I'm assuming yours are real-world
>estimates, unlike the overly optimistic figures given by X-10 and some X-10
>enthusiasts.
>
>Have you ever looked at Leviton's HCPRF All-housecode transceiver with the
>built-in lamp control module?  It's got a little wire pigtail antenna that's
>not very effective.  It's a unit that can certainly stand a "hearing aid" of
>some sort but I believe it's based on the older, AC hot = ground principle
>because the insulated pigtail has no shield - just a single wire connected
>to the PCB.
>
>I took it off line a month ago in an effort to stop a collision problem I
>was having.  Unplugging the transceiver stopped the problem, but it
>reappeared a month later.  More detective work required.  I am frustrated
>because I have reached the limit of what my trusty ESM1 can tell me about
>the powerline signal.  Today, I saw a 1.4V good signal followed by 1.4v
>noise followed by a .40 good signal after using a  credit card RF X-20
>transmitter.  Very repeatable.  Very weird.
>
>I'm wondering if I should buy a used o'scope or pony up for one of the
>signal analyzers.  Any suggestions?  Both Lynx and Monterey look like
>they'll give me at lot more useful information about the boogie men in my
>X-10 setup.  I bought a dozen new filters and am determined to clear the
>lines of all the noisemakers and signal suckers.  It just seems that it *is*
>possible to have a fairly reliable X-10 system because enough people manage
>to do it.  If a better meter or o'scope doesn't work, it will be time for
>Zigbee!
>
>Thanks for all the work in putting those pages together with screen shots,
>etc.  It really helps explain what's been a pretty amorphous concept.  Do
>you think you might ever do a similar analysis for UPB?  I'd really like to
>know whether their "Pulse Position Modulation" scheme is as robust as they
>claim.