• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Noise generated by ordinary dimmers.

    Blog entry posted in 'Uncategorised', July 11, 2010.

    Well,

    This post is about noise generate by those simple dimmers, that use a control signal and TRIACs.

    In order to vary the power through a load that is fed by the outlet AC, we must control the sinoidal wave's conduction angle:
    Method 1:44539
    The conduction angle is how much of the wave were available to produce power (alpha) or just how much of the entire period were actually there.
    On that case, we have a 270º conduction angle, starting at 90°. (90° + 270° = 360°). Remember that 1 period = 360°.

    Take a look to another way to do that:
    Method 2:44541
    At this time, you wait for beta1, start the conducton, and stop at (360 - beta2).
    On that case, beta1 = beta2 = 90°, so beta1+beta2 = 180°, so we have, alpha = 360 - beta1+beta2 = 360° - 180° = 180° conduction angle.

    The above types of control can be achieved w/o any problems with an AC SS Relay.

    Usually, when working with TRIACs (and SS Relays as well), we obtain this sort of output (because it is simpler):

    Method 3:44540
    Observe that on that case, we have alpha1 and alpha2, each one in a half-cycle. Well, as the cycles are symmetric, we could just make the calculations for a half-cycle and then x2.
    So we just have to wait for beta1, then start the conduction. After the half-cyle ends, we wait for beta2 and then start the conduction...
    If we analyse well, the Method 2 is some sort of the conjugate form of Method 3.
    Well, looking at that specific picture, on each half-cycle we have 90º of conduction, leading us to a 180º of total conduction, so the total power delivered is similar to the power delivered by a half-wave rectified AC.

    For practical purposes, alpha1 = alpha2 for Method 3 and beta1=beta2 for Method 2.

    Those sort of power control are analog to PWM power controlling. PWM applies to square waves, which have duty cycle, you cannot say that a sine wave has a duty cycle. However, you can use PWM to control the conduction angle.

    But what about the noise, and which one to choose?

    If we think about signals, those controlled sine waves are obtained by multiplying a sine wave and a square wave (with DC level):
    44528
    Observe that 0 x anything = 0 and 1 x anything = anything.
    Square waves have high harmonic content, contaminating the spectrum, thus, injecting noise to the AC line.

    First conclusion: Every method is going to produce noise!

    Now let's take a look at the spectra (animated), I ploted for a 60Hz power line, but the results would be the same for 50 Hz (of course taking account that you are treating 50 Hz as well):

    Method 1: 44530
    We have all the 60 Hz harmonics, the 180 Hz and 240 Hz reach up to about 10% ~ 20% of the original magnitude. The 120 Hz reaches about 30% ~ 40% of the original magnitude, in some cases.
    And we have a DC level over there (observe the 0 Hz), for conduction angles that are less than 180°. For a complete resistive load, that would be no problem. But a DC level for AC apparatus is useless.

    Method 2: 44531
    Again, we have all the 60 Hz harmonics. The only advantage of this method over Method 1, is the absence of a DC level.
    But we have something worse, for a conduction angle < 180º the 120 Hz harmonic has a magnitude higher than the 60 Hz itself!

    Method 3: 44532
    With this Method, we only have the odd harmonics.
    The disadvantage is that we have a pretty high magnitude for the 300 Hz, 420 Hz and 540 Hz.
    The advantage is that the 60 Hz harmonic is always larger than any.

    Second conclusion: All the methods are rich in harmonic content and, looking to the spectra, harmonics higher than 960 Hz can be produced. So filtering the noise, to keep the power line clean is a must.

    Now lets take a look at the amount of power produced by the conduction angle, for each case. Remember: the amount of power capable to be produced by any signal is directly proportional to its RMS value.
    To create a relationship, just do x 100

    Method 1: 44535
    Method 2: 44536
    Method 3: 44537

    As you can see, Method 2 and 3 has practically the same power relationship. But very different spectrum.

    Third conclusion: None of the methods produce a linear power control, the worst is Method 1. But Method 2 and 3 are fairly linear from 108° to 252° conduction angle.

    So, it is up to you, choosing which method suits you the best. IMHO, the classic Method 3 is the best. Method 1 and 2 can be completely wiped out, due the harmonics content, DC level and bad linearity (Method 1).

    Of course, there are others ways to do power control, those 3 methods are just examples of what we can do.

    Comments
    Hayato, September 20, 2010
    [QUOTE=mneary;bt342]"method 2" cannot be built with simple TRIAC triggering circuits as they only turn off when their current falls to zero.[/QUOTE] Thanks. Yeah, you can achieve it with AC solid-state relays. The method 3 is the simplest, and the "Usually, when working with TRIACs (and SS Relays as well), we obtain this sort of output (because it is simpler):" phrase was a introduction to that.
    mneary, September 22, 2010
    AC solid state relays contain TRIACs and cannot implement Method 2.
    Hayato, September 25, 2010
    [QUOTE=mneary;bt344]AC solid state relays contain TRIACs and cannot implement Method 2.[/QUOTE] I do not agree with that. They do have TRIACs or MOSFETs, but they have a built-in control circuitry that make them behave like a electro-mechanical relay.
    mneary, September 20, 2010
    "method 2" cannot be built with simple TRIAC triggering circuits as they only turn off when their current falls to zero.
 

EE World Online Articles

Loading

 
Top