Re: Motion Sensor Light for Front Entrance

[nobody@xxxxxxxxxxxx (Dave Houston)]

Rather than further spaghetti-ize things by trying to respond inline to your
inline comments, I'll just p*ss off all the anti-top-posters and try to
respond to your major points here.

Inline ammeters only work for purely resistive loads where voltage and
current always remain in phase with each other. With reactive loads
(inductive or capacitive), voltage and current are 90° out of phase,
requiring multiple, simultaneous measurements. Power=Voltage*Current so
averaging each doesn't work. You have to average the products of the many
readings.

Non-linear loads are usually those related to power suplies with diodes that
tend to draw current only at certain points in the voltage sine wave. These
also require multiple, simultaneous measurements.

I'll leave it to Jeff Volt to explain why you had the problem with the XTB.

CFLs have non-linear power supplies which cause (small) surges, triggering
the X10 flashes.

You really don't need the closed box. Just put your hand on the 7W
nightlight. It should be hot enough to be quite uncomfortable. IIRC, I used
3W which was still uncomfortable. X10 modules (with no load) have to
dissipate the heat associated with the idle load they represent. They never
get too warm to touch (unless defective) so this is a quick and dirty sanity
check whenever someone suggests they waste significant power.

Robert Green" <robert_green1963@xxxxxxxxx> wrote:

>"Dave Houston" <nobody@xxxxxxxxxxxx> wrote in message
>news:4c9f8b76.31830734@xxxxxxxxxxxxxxxxxxxx
>> >The Kill-A-Watt is notorious for not measuring very low wattage devices
>> >accurately.  My electronic wizard friend theorized that it's because
>they've
>> >optimized their reading range for items people would most likely measure.
>>
>> Yes and no. It doesn't indicate fractional watts so you need to use the
>kWh
>> mode and measure over a long time frame. I measured several X10 modules
>this
>> way for ~100 hour periods. See...
>>
>>      http://davehouston.org/x10-power.htm
>>
>> NOTE: The readings were under no load conditions but I doubt that the
>> trickle current will be that significant.
>
>Yes, senility strikes again.  But I demand half credit - I remembered that
>it happened and that you were involved and believe me, like the Langoliers,
>things out beyond a certain distance are getting all dark and hazy and a
>little bit scary.
>
>I did recall the most important result of your research to me, at least
>concerning the Kill-a-Watt.  For low end reading to be accurate, low power
>loads needed to be measured over time and there was a reason that you
>couldn't just take the Kill-a-Watt's low current, instant readings as
>gospel.  Thanks for correcting my "holey" memory.  One hole filled means
>something else just leaked out, though.
>
>The amount of trickle current - do you think it varies in some way with the
>nature of the load connected to it?  Does the current that causes CFL's to
>flash also heat a tungsten filament slightly or does the differing nature of
>the load affect how much current is passing?
>
>I get some pretty serious flashing of two hi-eff 34W fluorescent tubes in
>the workshop - enough to see by if you move very slowly (-;  It would be
>easy enough to first to compare an appliance module's power consumption
>without such a load and then with.  In about 200 hours or about 8 days from
>now.  (-:
>
>So, 100 hours reading an appliance module, no load v. 100 hours reading
>appliance module with the two bulb shoplite that flashes more brightly than
>any other fluorescent in the house.  I can even supplement those with
>readings of an appliance module with a small CFL and another with a 100W
>tungsten bulb.  I also want to see what it costs to run a 7W nitelight with
>a CFL load since it's now X-10's tech support recommendation to cure the
>flashing problem.  No wires monsters, no lights in boxes - it should get
>safety approval.
>
>> >I've been fooled by it on a number of occasions.  An in-line amperage
>> >measurement is likely to show you something different.  Also, the X-10
>power
>> >supplies have a rather unusual design which contributes, I think, to the
>> >Kill-A-Watt's erroneous readings.
>>
>> In-line amperage measurements are worse than useless with non-linear loads
>> like these. The Kill-A-Watt takes thousands of instantaneous readings of
>> both voltage and curreny each second and then averages them to get a very
>> accurate reading. The kWh mode has 30ppm accuracy.
>
>The issue I ran into was the limit imposed on the plug in outlet on Jeff's
>XTB.  It's got a wattage limit for plug in devices, I believe it's either 10
>or 15W, but I could be wrong. I measured the devices with a Kill-a-Watt (2
>Maxi's, 1 ControlLinc Maxi, 1 Mini controller and a Mini Timer that uses a
>wall wart transformer and not the standard X-10 type power supply).  The
>Kill-a-Watt gave me a very low number that was below the limit imposed by
>the XTB, and so I plugged them all into one unit and it worked. I did not
>measure it using the accumulation kWh mode because waiting around for 100
>hours to get a reading is not very practical, but alas, seems necessary when
>dealing with low wattage items using the Kill-a-Watt.  FWIW, the
>ControlincMaxi's nameplate lists 100mA, the X-10 lamp and appliances modules
>list no wattage information, and the X-10 Mini lists 2 watts.
>
>The XTB operated quite well with 5 units plugged into the amplification
>outlet UNTIL I removed the Mini-timer with the wall wart.  Then, it fried.
>When it happened, I got into a discussion with Jeff Volp who immediately
>went over my admittedly low altitude head with discussions of mixed reactive
>and inductive loads on both his circuit and the metering methods I was using
>to make sure I was within the wattage range specified on the XTB label.  It
>was, as they say in the Army, above my pay grade.
>
>> >IIRC, a while back Dave Houston (has anyone heard from him?) did some
>> >measurements that more accurately pegged them at five watts.
>>
>> I've neen dealing with major health issues (and have additional surgery
>> scheduled).
>
>Welcome back!  Been a while since we last heard form you.  Sorry that you're
>still having health issues.  I hope all goes well with your medical
>procedure.
>
>> As for my measurements, I did exactly the opposite, refuting
>> numerous people who were claiming X10 modules used 5-10W based on ammeter
>> readings.
>
><Sigh>  It seems I lost the decimal place in the fog of senility.  Your
>figures were .4 and .5 and not 4 and 5.   I hopefully still have my notes
>around, but I distinctly recall a metric you used was that they were much
>cooler to the touch than a 7W nightlight, giving a "rough" reading that said
>common sense dictated it had to be below 7W.  That's when I decided that
>under careful enough controls, the relative amount of heat output between a
>module and a 7W night lite could be used to verify readings with meters,
>that for whatever reason, gave erroneous readings at the low end as when
>used in a typical way.  (Like me!)
>
>IIRC, and it's clear I don't "Recall Correctly" anymore, they did make the
>heat in a closed, insulated box rise respectfully.  They were clearly
>drawing power that could be measured as heat output.  I recall they differed
>from your readings, but I also recall the experiment being shut down by the
>resident safety engineer, Ms. Swmbo.  The idea of enclosing electronic gear
>(especially lamps) in wooden boxes lined with Styrofoam did not pass muster,
>even though it was "unlikely but not impossible" to start a fire.  I might e
>ven have the readings saved in a draft message.
>
>I believe the modules I was running were running under load because I wanted
>to know the effect and size of the trickle current flow.  That was when
>CFL's had just gone mainstream and  I first noticed the flashing problem.  I
>began delving deeper into how to measure the trickle current flow, but
>someone, I think it might have been Dan L. advised against it, charitably
>citing safety concerns but more likely having to do with the primitive
>equipment and skills I possess. (-:
>
>> >If you have an in-line ammeter (the tong meters aren't really suitable,
>> >either, you might want to revisit the study with loads plugged into a
>batch
>> >of modules.
>>
>> Don't waste your time - this method is, as noted above, worse than
>useless -
>> worse because it is extremely misleading.
>
>Certainly when using X-10 modules and their linear power supplies.  Am I
>right to assume an in-line ammeter would be much more accurate with
>inductive and purely resistive type loads?   I would assume by the short
>time between the posts that Art took an instantaneous measure, and not an
>averaged one.  But I'm Often Wrong, so only he say for sure.
>
>At least all this is what I recall when I insisted to an incredulous Jeff
>that the XTB had been running with no incident with all five X-10 items
>listed above plugged in via power strip.  I am sure he'll be around to
>correct me shortly as well.  As I said, this is really outside my bubble.
>The XTB problem occurred AFTER I REMOVED the sole inductive load (a
>wall-wart powered Mini-timer) from the powerstrip with the other four
>components.  It seems quite counterintuitive that reducing the overall load
>caused the XTB to burn up, but apparently the combination of inductive and
>linear loads on the XTB was preventing the load from burning up the XTBs
>input circuitry.  I won't paraphrase Jeff's explanation and embarrass myself
>further.  Maybe he can reiterate for us.
>
>I should note that Jeff both offered to repair the unit free and changed the
>labels and instructions to accommodate the unusual results I discovered by
>flagrantly disregarding the label warning.  Now, in addition to the wattage
>limit, he indicates that no more than two X-10 devices be plugged into the
>XTB's amplification outlet.  I think a lot of non-engineers and techies have
>real problems understanding the intricacies of the different types of
>devices as well as other concepts, like the PF (power factor), phase angles
>and so on.  I know I do!
>
>If reactive and inductive loads can interact as they apparently did with the
>XTB, is it possible that measuring multiple modules and dividing that
>outcome as both you and Art did is not actually equivalent to reading a
>single unit?  I assume you did that because reading single units that draw
>under a watt is problematical for the KaW, even in the kWh "accumulation"
>mode.
>
>Good to here from you again, Dave, even if you're still correcting the
>living hell out of me and shaming my Mad Cow brain - that has to be it - Mad
>Cow.  Anyway, that's how we learn, even if the ego gets a little burned
>around the edges.
>
>And again, here's to a good surgical outcome.
>
>P.S. to Art.  Glad you asked this question, nothing at all's changed but I
>feel alot better that X-10's not eating 400 watts just "being there."  I
>feel a lot worse though, about where all those extra watts are going.  A
>while back someone suggested that the older the house wiring, the more
>likely substantial amount of juice are going up in heat in the wires.  That
>would not be good. )-: