Connect with us

Dimmer through mosfet?

Discussion in 'Electronic Basics' started by [email protected], Oct 4, 2004.

Scroll to continue with content
  1. Guest

    Would a construction like this:

    220V AC --> Fullwave rectifier --> Capacitor --> 1xPowerFet --> Capacitor -->
    Resistive load (bulbs, heaters)

    PowerFet/Mosfet etc.. feeded with pwm pulse.

    Be able to get significantly reduced RF emissions than standard tyristor style
    dimmers ?

    Computer control is planned.
  2. Ban

    Ban Guest

    Don't they have 230V AC in Sweden, as in the rest of Europe?
    EMI problems will be significantly worse with the setup you desribe, as you
    have to switch much faster with the PWM, the capacitors won't help much
    either, what you need are inductors.
    Standard dimmers use triacs, not thyristors.
  3. Hmm sine wave and reverse phase control dimmers dimmers like: look for Compusine

    bit more

    Guess some of this technology comes from AC motor drives, but the very few
    manufacturers who make them kind of indicates that it seems close to rocket

  4. SCRs tend to be a bit more resiliant to a big lamp blowing on them where
    triacs can blow before the breaker trips, lot of older and big expensive
    installation dimmers still use thyristor or solid state relay with SCR
    (thyristor) outputs.

  5. CFoley1064

    CFoley1064 Guest

    Subject: Dimmer through mosfet?
    RFI is caused by di/dt rather than switching element itself -- to reduce the
    effect, good lamp dimmers will typically add a small (20 to 30 uH) high current
    toroidal choke in series with the lamp load to soften turn-on.

    It's potentially hazardous to rectify line voltage and expect to control it
    with a computer without an optocoupler somewhere in your control chain. I kind
    of get a feeling the reason why you're doing this is that you feel comfortable
    with switching FETs and you don't know too much about TRIACs and thyristors.
    The best way to do the job you're describing is with standard TRIACs and
    optocouplers, if for no other reason than having a current path between
    hazardous voltages and a computer is asking for disaster.

    Take a look at Fairchild App Note AN-3003, Applications of Non Zero Crossing
    Triac Drivers Featuring the MOC3011. Since you're controlling 220VAC, you'll
    need to use a MOC3051/3052 instead of a MOC3011.

    Good luck and play safe
  6. Guest

    Any tip on a semiconductor that can turn on the load in each phase with a
    slower di/dt without becomeing an electrical heater :) ..?

    Optocoupler is so natural for me safe-wise that I didn't put it in the generic
    When working with dangerous voltages I always check things twice, and make
    sure all parts are properly fastened at safe distance from eachother. And
    analyze what happens should a component fail, or if thunder, spikes etc..
    would occour.

    I know very well about triacs, my idea was to find an alternative that will
    avoid emitting RFI.
    Maybe generating pulsed sinewave with a mosfet would be better?, as the di/dt
    might then be lower (U=220*abs(sin(w))).

    Nice chip.. now the thing is to get it ;)

    Btw, is it feasable to modify an of the shelf dimmer by replaceing the dial
    with a transistor-optocoupler in some way?
  7. CFoley1064

    CFoley1064 Guest

    Subject: Re: Dimmer through mosfet?
    The key to reducing RFI from dimmer switching is to just take the edge off the
    very fast transistion from no current to full current, particularly when
    dealing with cold light bulbs, which grab a lot more current than ones that are
    already on. A small choke will do this well. There are many chokes, both
    toroidal and non-toroidal, which are made for this application. Another thing
    is, you only need one for all the triacs -- just put it in series with the line
    before your loads like this (view in fixed font or M$ Notepad):

    L1 ___
    | | | | |
    | | | | |
    ,---. ,---. ,---. ,---. ,---.
    | X | | X | | X | | X | . . . | X |
    '---' '---' '---' '---' '---'
    | | | | |
    _|_ _|_ _|_ _|_ _|_
    V_A V_A V_A V_A V_A
    / | / | / | / | / |
    | | | | |
    | | | | |
    L2 | | | | |
    created by Andy´s ASCII-Circuit v1.24.140803 Beta

    All the opto stuff has been assumed here.

    You can get a 10uH 11A choke for $3.99 and a 25uH 5.5A choke for the same price
    from Jameco:

    Jameco #: 208266 25uH, 5.5A
    Jameco #: 208258 10uH, 11A
    $3.99 USD ea.

    If you have enough current to require the 10uH chokes, you might want to put
    two in series if 10uH is still giving you too much RFI. The easiest way to
    tell is to get an old AM transistor radio, and dial over the frequency range
    while your dimmer circuit is working. If you can't hear the buzz when your
    dimmer is working, you're probably OK. Try various antenna positions and
    orientations, as well as different power settings on your dimmer.

    Looking at MOSFETs for this is probably the wrong direction. A fast switching
    MOSFET which would give you a simulated sine wave would have many more
    opportunities per line cycle to generate current pulses which would cause RFI.
    The above is the simplest, easiest way to do this. Also, the thermal constant
    of light bulbs and heaters is such that you won't gain anything by using high
    frequency PWM for voltage control.

    Is it possible you're concerned about finding the zero crossing? If you're
    looking at using computer control for TRIAC phase control, all you need is a
    way to mark the line voltage zero crossings, and then knowing if you've got 50
    Hz (10 ms) or 60 Hz (8.3 ms), you can time out to where you want to turn on the
    TRIAC. Most people use the low voltage AC coming off the power transformer for
    the computer power supply to generate the zero crossing interrupts. If you're
    using a PC, you will need a separate transformer, as well as a way to get the
    interrupt into the PC (I'd try using assembler/C and one of the status pins on
    the printer port to generate interrupt, if you don't have a separate card for

    The MOC3051 is available at Jameco as their #: 277860 for $0.74 ea.

    Good luck with your project. Try the TRIAC/opto setup. It's what everyone
    else does, it's more efficient, costs less, and it's really what it's made for.

  8. CFoley1064

    CFoley1064 Guest

    Subject: Re: Dimmer through mosfet?
    Nope. Most "off-the-shelf" dimmers operate on a principle of charging an RC
    delay with line voltage, and using a DIAC or other SBS (silicon bidirectional
    switch) to discharge the C once it reaches a threshold voltage. If you know
    the zero-crossing point and can use an optocoupler that will source enough
    current to turn the triac on, you can replace the potentiometer, a couple of
    Rs, the C, and the DIAC. But in that case, why not just make it yourself?

    Good luck
  9. Guest

    The key to reducing RFI from dimmer switching is to just take the edge off the
    Some further reading pointed me in the direction of IBGT. One approach is to
    let it conduct at the beginning of the sinewave. And then shutoff when desired
    power has been reached (area under the graph..).
    Another method is to let the ibgt gate be pwm modulated with a few tens of
    kHz and by varying the duty cycle. Put together a sinewave through a series
    of pulses with varied duty cycle. The addition of a capacitor or similar will
    then smooth it to a near perfect sinewave.
    ( Dimmer Technology.pdf)
    Won't the wire between dimmer and load radiate RFI, as most power cables don't
    have foil, or twists..?

    I prefer if the unit takes care of zero crossing etc.. by itself. A simple
    PIC 12Fxxx could do the job. Being fed what to do over some multidrop
    With a central computer doing the intrisics there might be problem if the
    computer is on another phase than the dimmer+load. Or there's a software crash.
  10. Guest

    Some further reading pointed me in the direction of IBGT. One approach is
    Sinewave pwm is bad?
    I thought the smaller surges would be easier to filter out.
    Sinewave pwm is noisy?
    I know :)
    What I would like to achive is not a cost reduction per se. But a reduction of
    RFI compared to forward/reverse phase dimmers.
    Well I don't mind ethernet/usb, as latency won't be an issue for me.
    Infact a IR link from the computer to the devices would be useful to eliminate
    the need for wires, radiolinks, x10 etc..
  11. Nope refering to normal phase control, sinewave dimmers are virtually
    Reverse phase dimmers are quieter as well because theyre switching off
    rather than on.
    But the few manufacturers in the arena do indicate there is difficulties
    with sine and reverse phase dimmers, not sure what, but at least one
    manufacturer showed a prototype at a trade show couple of years back and
    have still to release a production version.

  12. Guest

    Any suggestions on filter?
    I read somewhere that it might have to do with some patent....
    Ofcourse if could be some purely technical issue.

    Reverse phase dimmers is an improvement. I think sinewave pwm is the best (now)
    so I will try that route. Althought I think I will do some basic testing first
    to see if the concept works at all.
  13. Bigger chokes, per channel, slow the rise time as it bangs on, is generally
    whats used , and x y y capcitors in a delta formation across the output,
    though these can cause nuisance tripping on earth leakage protection.
    Could always go scouring through U.S. patents it has the full
    published patents online.
    The concept works , because people are making them , but not very many are
    making them including some top end dimmer makers.

    Mebbe looking at Class D amplification might give some insight because it
    seemed to take ages before anyone got that to work outside the lab as well.

    Remeber to let us know how your getting on.

  14. Guest

    Nope refering to normal phase control, sinewave dimmers are virtually
    Chokes as in inductance in series with the ac input?
    What about not connecting the earth to the switching part, only let it through
    to the load. And especially what if neither ac source nor load has earth?
    (as to avoid any earth leakage circuits activation)

    (As a sidenote, in .se 4-wire system is used (PEN-wire). And 5-wire system is
    done through connecting neutral to ground in the central fuse box.)
    I don't live in US, and I won't massproduce it so I think it's a non issue. ;)
    Well management can make bad decisions stay around a long time..

    Any ideas concering active power factor correction circuits?
    I'm thinking on the case I would later like to apply the ibgt-sinewave stuff
    to a inductive load.. either DC motors, or AC->DC->Inverted for AC motors.

    Btw.. I'm thinking on how to use the ibgt-sinewave technique to make
    DC 12V -> AC 240V.. could produce sinewave for less $$ I hope. =)
  15. Per channel , yes

    has typical output channel
    Motor suppressors use a dlta form of Y rated , live to earth and X rated
    Live to Neutral to try and dumop some of ythe chash.
    At 220/240V unless its double insulated you definately want anything metal
    issue. ;)

    Not to get a patent but to look at what has bee patented to get some ideas
    It was a technical thing , sure Sony released and withdrew a ClassD car amp
    because of unrelibailty in field.
    Motor drives is a whole big subject that would be better asking someonelse
    Seen some cheap, `true sine` inverters about, in addition to the `quasi
    sine` 2 step square wave about, usual far eastern stuff , buy one and pull
    the lid off.

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day