No announcement yet.

Does dimming lights really save energy?

  • Filter
  • Time
  • Show
Clear All
new posts

    The pdf includes 2 sets of timing measurements.
    Both show the 'dimming' is (fairly) linear in time, centred around the mid point of the half cycle.
    Page 1 is the results from a Lamplinc2000SLS (60Hz US spec), measured by viewing the control pulse sent from the PIC MCU to the triac.
    Page 2 is the LM12U module (50Hz UK spec)measured at the triac output to the Lamp.
    they are both different in the timing, even allowing for the change in ac supply frequency.
    So when you dim to X% you dont always get X% of the energy usage.
    I'm sure someone can work out the math if they really want to.

    PS I needed to measure these to correct the dimming for 50Hz operation...honest! [img]/infopop/emoticons/icon_wink.gif[/img]
    Attached Files


      <BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR> Like you said, this supports Johnh1's observation, but also shows that a 200W bulb never gets to 200 watts.

      Is it possible that a 200W bulb doesn't reach 200W because the voltage is less than "standard". I noticed in the first test zoikos did he mentioned that the voltage was 116VAC. If it were 120V, wouldn't the wattage have been a little higher (maybe even 200W)?

      Anyway, this is excellent information!


        DC: Yes, that would be my guess. I would also suspect variations in manufacturing, except that ALL of the bulbs are a little low at 100%.


          I tested a bunch of bulbs (116.8VAC/59.7Hz) just to see what the power dissipation was. All of them were under their rated wattage, but they all fell within 5-10%. I'm going to take a box of 'em into work tomorrow where I've got an AC variac and some better test equipment to check them out. I'll let you know what I find out.

          Rocco, nice graphs! Thanks for taking the time to put all that together for us. Johnh1, your data is interesting (and its prompting me to find parts to fix my old 'scope too)... Was your data centered around 'dim value' or 'triac delay' ?


            I generated the data by using HS to set the 'dim' value for the units.
            The Lamplinc steps are those stated in their manual(32 possible steps between 0 & 100), so I figured it best to measure at those settings. For each value set with HS the timing (delay relative to the zero crossing of the ac supply)was measured at the output from the MCU which then goes through a interface/isolator? before reaching the Triac gate.
            The LM12U data was simply done with the scope on the output (of the Triac) to the lamp. These 'dim' settings are a little rougher, as the basic X10 dimmer only has 22 possible steps for dim and so when I set the dim to say 20%, it was probably slightly off in reality. But only by a couple of %.



              Wow! In an effort to keep my tests 'simple' (using basic 10% dim increments) I totally overlooked the design characteristics of the X10 devices. My guess is that the dimming module or dimming switch would appear much more linear had I used the data points corresponding to the device. The SwitchLinc 2380 has 32 levels as well, and there's a handy chart in the instruction booklet.

              Thanks for pointing that out!

              A.J., regarding CF bulbs and 2-way switches: The second test I did was with a 2-way SwitchLinc 2380, and the last bulb I used was a 23W Philips Marathon dimmable CF. That bulb only responded to an X10 command from the full "off" state (it responded just like an incandescent using an LM465 though). It seems that the switch latched "on" to whatever dim level I sent it. Could not get it to respond to another dim level or an X10 "OFF" from the computer. It did on/off/dim locally though. This kinda kills my idea of replacing my front porch light with a CF & SwitchLinc, although an incandescent out there would work like a champ. You might want to experiment with one switch/bulb for your application before you go out and spend a bunch of money.


                <BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by zoikos: It seems that the switch latched "on" to whatever dim level I sent it. Could not get it to respond to another dim level or an X10 "OFF" from the computer. It did on/off/dim locally though.<HR></BLOCKQUOTE>
                One of the ways a Compact-Fluorescent gets it's 'Compact' is by using a high-frequency electronic ballast. I bet it's spewing enough noise in the 120Khz region as to render X10 undetectable. Try an X10 line filter between the lamp module and the CF bulb. (sure, it's easy for me to spend your money).


                  In the original test, the 23W CF performed just like an incandescent bulb using an LM465. It wasn't until I replaced the LM465 with a SwitchLinc that it had X10 issues. To help validate this, I popped a few CF's in places where I've got SwitchLincs installed (Master Bath, LR, Guest Bath) and came up with the same thing - the CF's would go "ON" initially, but would not respond to any other X10 command while they were on.
                  I do agree that CF bulbs are noisy, I just don't necessarily think that the noise generated caused the SwitchLincs not to respond in this case. Still pondering this though...


                    Don't know much about switchlinks, so the question in my mind is whether it uses a separate neutral wire or whether it completes the X10 circuit through the bulb (like the wall switches from X10). A lamp module always has a neutral wire available. If there is no neutral, maybe the X10 signal can't travel through the bulb when it's on.



                      The X10 modules are powered through a "series" connection with the lamp as the load. It needs to "keep alive", and so the lamp circuit is never off.

                      Switchlinc and other lamp modules require a neutral connection and are powered independently of the lamp load, and so can switch the loads completely off (or on).

                      At either end of the spectrum it should make no difference to the user as the light level variation is "invisible". This is also why the standard one way X10 modules are only one way. To be able to transmit back into the line, the current levels would need to be greater, and so the lamp would probably have a noticeable low level dim that might fluctuate whenever it was sending.




                        GDude: I understand what you are saying, but it is different then what I am thinking (not that I am thinking clearly).

                        I'm referring to the path an X10 signal needs to reach a module. In a plug-in lamp or appliance module (two prongs in, two prongs out) the X10 signal travels through the hot wire to the module, and back through the neutral wire. It never has to pass through the load. Wall switches that have a separate neutral wire work the same way.

                        With the two-wire wall switches, like the ones made by X10, the x10 signal has to travel through the load to make a complete circuit. These switches historically would not work with CF bulbs, because the ballast would block the X10 signal.

                        This is why I asked whether the SwitchLinc had a neutral wire. As it turns out, it does. So that can't be the problem. I withdraw the question.


                          I can tell you, I have about 20 dimmable screw in cf bulbs, some are on lamp modules and work fine. But the majority are on PCS switches about 14 I believe. Not one single problem with the CF bulbs ( they stay dimmed sometimes for a long period of time and they have been in for about 2 years. ) Or with the PCS / Lamp modules controlling them.

                          Why oh why didn't I just leave things alone, they had been working.


                            This thread has presented some excellent data that would be useful to everyone. This needs to be posted on someone's site.

                            (It would be really cool if stuff like this were available in a special section of this message board. Tests with real data. Tests of products compared against another.)



                              In response to Stevene's observation: So, like Zoikos said, it sounds like a SwitchLinc issue.

                              GenevaDude: I thought about what you said about why the the 'series' powered modules would not be two-way. I had never thought about it before. It makes a lot of sense. A light-bulb lit-up over my head (although not to full brightness).



                                I bought 1 switchlinc. and thought it was the most horrible product compared to the PCS switches. Not only did it consistantly turn on and off randomly, It also lost its house/device code twice in the 2 years I have had it.

                                The local dimming control was also not as nice as the PCS switches.

                                I am not saying all switchlincs are horrible. But I was so pleased with the PCS switches I stuck with them.

                                I've read many users that have no problem with the switchlinc's.

                                Why oh why didn't I just leave things alone, they had been working.