Re: X10 signals can be TOO strong!

["Alan Vogel" <avogel2@xxxxxxxxxxxxx>]

"Marc_F_Hult" <MFHult@xxxxxxxxxxxxxxxxxxxxx> wrote in message
news:oq74o2t9ac60teh956n96lah30ej7l165d@xxxxxxxxxx
> On Fri, 15 Dec 2006 03:18:50 GMT, "Alan Vogel" <avogel2@xxxxxxxxxxxxx>
> wrote in
> message  <uuogh.5921$Gr2.922@xxxxxxxxxxxxxxxxxxxxxxxxxx>:
>
>>The inductor in series with the load is there to limit di/dt.  It was
>>probably required by the FCC to reduce EMI.
>
> CE has EMI standards but US seem non-existent. Consumer tolerance for
> dimmer
> and filament hum and buzz may be the controlling factor. Whatever the
> reason,
> the snubber circuit is an LRC  network even if the inductance is entirely
> owing
> to the load and not part of the device as is the case with a built-in
> choke.

Because the Insteon dimmer uses a uP, and its frequency is greater than
9kHz, FCC part 15 regulations will have to be met.

>>I think this is all starting to make sense.  The highest di/dt transition
>>is
>>when there is a heavy load and the conduction angle for the triac is ~4.16
>>ms in from the zero crossing (50% dim).
>
> 4.16 ms is indeed 50% into the half-cycle, and so is peak voltage, and,
> assuming the filament is as at constant temperature during any given cycle
> and
> thus R ~ constant, is also peak current. But at the peak,  di/dt = 0 , not
> "highest".  Or by "di/dt transition" did you mean the second derivative?

For dimming, the triac is abrubtly turned on mid-cycle, so di/dt is high.
The triac turns off at the next zero crossing.  The uP repeats this process
for a set dim level.

> (And FWIW, 4.16 ms is about 71% of RMS voltage and 33% of luminous
> intensity
> for typical incandescent lamp and not "50% dim" by those measures of
> dimming.)

Yes, this is the RMS value.  But 33% dim as measured by RMS may appear to be
brighter to the viewer.  Doesn't the eye work more as a peak detector than
an RMS detector?  Isn't this why multiplexed LED displays appear brighter
than they otherwise should?

>>This high current pulse must have
>>coupled back into a high impedance node (either the zero crossing
>>detector,
>>or the XTAL input pin on the uP).

Okay, maybe my wording was too strong here.

> Not to dismiss the possibility, but why "must have" ?
>
>>Either way, the altered timing caused
>>abnormal conduction angles which showed up as flicker.
>
>>Increasing the value of the inductor attenuates the signal enough to
>>alleviate the symptom.
>
> This the fix I mentioned that SmartLabs put in place for INSTEON .

Yes.  This is the subject that I was talking about.

> The tendency to flicker with increased load could conceivably  be caused
> by
> decrease in series inductance (and consequent decreased attenuation) with
> increased current (load). The inductance of non-gapped inductors like the
> rods
> and toroid used in dimmers does decrease with current, but I don't know
> the
> magnitude of the decrease or whether it is significant with a negligible
> DC
> component.

A powdered toroid core can have a distributed air-gap.

>
> ... Marc
> Marc_F_Hult
> www.ECOntrol.org