Re: I think they've done it again.

[Marc F Hult <MFHult@xxxxxxxxxxxxxxxxxxxxx>]

On Sun, 25 Jun 2006 17:30:15 -0400, Marc F Hult
<MFHult@xxxxxxxxxxxxxxxxxxxxx> wrote in message
<g2ut92do7d99l555jouk9fmttc15jefbv6@xxxxxxx>:

>
>I'll look at the reports some more so as to be able to respond more
>usefully. I think that i understand your timing question and add the
comment
>that when I looked at line noise created by generic and X-10 WS467 dimmers
>several years ago, I found that the highest peak voltages were measured at
>about the 1/4 the RMS output IIRC.
>
>Recall that an inductor 'stretches' the duration of a pulse, so the delay
>caused by the L1 choke  might be long enough to trigger gate triggering in
a
>second TRIAC in a second dimmer when the delayed/stretched INSTEON PLC
>signal is _added_ to the TRIAC noise from the INSTEON TRIAC at some time
>significantly after zero crossing and that signal is not sufficiently
>attenuated (there are two issues here; timing and noise/signal strength.)
>
>Later ... Marc
>Marc_F_Hult
>www.ECOntrol.org


I conclude that as a an actual performance problem, this is a tempest in a
teapot: a problem nobody here has actually experienced (not surprising since
it seems most folks participating in this thread may not even have an
INSTEON installation ;-)

An INSTEON rep wrote in the thread I cited previously that: "We saw this
problem during the beta testing of the SwitchLinc dimmer in two units out of
800 beta samples." and goes on to give contact info for a person with a
flickering dimmer to be able to contact Smarthome on a Sunday to arrange for
a replacement dimmer. The same rep reports that INSTEON spent $40,000
troubleshooting and replacing possible off-spec components before shipping
but apparently a few slipped through (or at least that is one explanation
offered).

According to INSTEON, the dimmers with revised/QA'd TRIAsC/chokes  will be
shipping 4th Qtr which will presumably reduce problem frequency closer even
to zero. I haven't experienced any problems myself nor apparently have the
vast majority of other INSTEON users. So the solution to this apparently
infrequent problem is at hand.

As to the challenge of trying to divine the cause, first step is problem
definition which is problematic because reports are not consistent. I'd be
more interested if there was a way of testing the various hypotheses posed.
Absent an actual case of flickering on which to test causes and remedies,
another way would be (eg Spice) modeling but that's more work than I'm
inclined to put into this particular academic exercise.

... Marc
Marc_F_Hult
www.ECOntrol.org