Connect with us

Analog switch question

Discussion in 'Electronic Basics' started by Anders N. Vinje, Sep 25, 2006.

Scroll to continue with content
  1. I am trying to make an audio router board for a mixer.
    I have found sevreal analog switches that can do the job.
    I decided to use the MAX4173 Clickless Audio/Video Switch.
    Its only one thing that makes me wonder.
    This IC works from a +5V supply and is SPI controlled.
    The thing i have seen with other analog switches including the 4066 is that
    they require a dual supply to work with a audio signal with refrence to
    In the datasheet it say the analog voltage swing may go from 0-5 volts.. So
    my question is this: Do i have to add a 2.5 V DC componet to the signal
    going into the switch so it may swing +-2.5 volts and then add a capacitor
    at the output to block the DC so it can be refrenced to ground again..?

    Anders N. Vinje
  2. tempus fugit

    tempus fugit Guest

    I think you do. There is an interesting article on this very topic on the site under 'Electronic Switching with the AD4053' article.

    I'm not sure if he suggests this because of the age of the chip used (i.e.,
    maybe newer chips don't have this problem) or not. I'm going to be working
    with some of these ICs shortly for use in an effects pedalboard, so I guess
    I'll find out soon enough. BTW, there are other ICs out there that have much
    lower on resistances (don't know if this is a concern for you or not), made
    by Maxim and AnalogDevices, among others.
  3. default

    default Guest

    Yes you do

    I understand what you mean. In the applications they show signal
    in/out and no reference to the signal levels, but in the absolute
    maximum ratings and in the electrical characteristics they state that
    the signal can't be lower than zero with single supply operation - and
    note one says that the signal will be clamped with internal diodes to
    within .3 volts of +V and Ground.

    So, this has the same limitations that a 4066 has and you'd have to
    reference the signal to 1/2 V, keep it to less than 2.5 volts on
    either positive or negative excursion then recover the signal and
    ground reference it on the output of the switch.
  4. Jamie

    Jamie Guest

    those that need a little - voltage can be done with a simple on board
    a little 555 or the like timer can generate a square wave pulse that
    then be rectified for a - voltage source, just connect the Vss on that
    out put.
    you may want to use the Cmos version of the 555 if your working with a
    5 volt supply , the bi-polar may have issues getting started at that
    to balance the voltage you could put a trimmer on the output of the
    inverter to select the correct - voltage to get near perfect zero output
    when the switch is on.
    just something to think about i guess!

    something to think about.
  5. default

    default Guest

    International rectifier has an interesting solid state alternative to
    relays. It is a AC/DC solid state relay with a four amp 20 volt

    These new 20V single-pole, normally open, solid state relays utilize a
    HEXFET® MOSFET output switch, driven by a unique integrated
    photovoltaic generator circuit. The output switch is controlled by
    radiation from a GaAlAs light-emitting diode (LED) that is optically
    isolated from the photovoltaic generator. The new series is available
    in 6-pin DIP, 6-pin SMT, and in tape and reel.

    On resistance 0.050 ohms
    On resistance DC only 0.015 ohms
    Amps 4.0
    Amps DC 6.0
    Input/Output isolation 4,000 volts
  6. Thank you.
    Starting to design PCB...

  7. tempus fugit

    tempus fugit Guest

    They look good except for the 3 000 000 nS turn on time. The on resistance
    is wicked though.

  8. default

    default Guest

    Yeah, well this is only the first of its breed (that I know of). If
    they are only trying to get into the mechanical relay market it seems
    like an alternative. Hopefully the switching speeds and price will
    come down.

    I kludged something together years ago with some surplus calculator
    solar arrays and four green leds powering them. It worked but the
    speed was nothing to brag about either, and it was bigger than a
    mechanical relay.
  9. tempus fugit

    tempus fugit Guest

    I don't really know that much about this type of device, but I have been
    pointed to a few others like it. So would you still need to bias the
    ins/outs to use them for audio? It does seem odd that in the case of the
    Maxim parts, which are marketed as having use in audio switching/routing,
    they don't really mention anything about setting up a bias point for the
    audio to swing around.

  10. Eeyore

    Eeyore Guest

    It requires that you undertand how the parts function.

    Electronic design wasn't ever meant to be be easy !

  11. Bob Masta

    Bob Masta Guest

    You might want to look into the 4051, 4052, 4053 series
    of CMOS analog switches. These are designed to switch
    bipolar signals if you provide bipolar supplies. They are
    really handy as line selectors since they have multiple
    poles in one package. They are not "clickless". If you
    switch on a signal and happen to hit it at peak, there
    will be a click, compliments of Mr. Fourier. However,
    I have not had problems with control signal feed-through clicks.

    Note that in any analog switch design there can be some
    "on" distortion because the switch resistance is nonlinear.
    However, you can minimize this by making sure that the switch
    resistance is not a big part of the total resistance, such as
    by having it look into a high impedance input of the next

    Best regards,

    Bob Masta

    D A Q A R T A
    Data AcQuisition And Real-Time Analysis
    Home of DaqGen, the FREEWARE signal generator
  12. tempus fugit

    tempus fugit Guest

    Thanks Bob.
    I guess it's kind of a toss up - which is more of a pain, biasing the
    ins/outs at the midpoint of the supply, or provding a bipolar supply. If I
    understand your comments about switching noise, you are saying that with no
    signal present there are no clicks, but (as one would expect) if there is a
    signal you will get a click because of the change in volume? From your
    remarks about on resistance and distortion, it follows that the loser the
    Ron the better?

  13. default

    default Guest

    The IRF gizmo is designed to replace an electromechanical relay
    directly and doesn't require the negative supply. They (and I) use(d)
    a series of small solar cells (photovoltaic array) that are lit by
    leds - since they can add (in series) solar cells and get any voltage
    they want so it is easy to come up with the 4-5 volts required by the

    Mosfets only require a few nano amps to turn them on but at high turn
    on voltage ~4-5 volts for saturation. Just as easy to develop a
    separate array to turn on each internal mosfet. So you could use them
    to switch audio or AC signals if the turn on/off delay isn't going to
    hurt your application. With a 4 amp 20 volt rating they were
    developed for power switching applications like relays.

    For that matter, if speed isn't an issue, you could use DIP style
    electromechanical relays yourself.

    The advantages of the cmos switches are: lots of them in a small
    package, and able to switch at relatively high speeds, low current
    consumption. Disadvantages are the high contact resistance, inability
    to handle voltages over/under the supply rails, low power switching
  14. Bob Masta

    Bob Masta Guest

    Well, basically true, but in reality it's the nonlinearity of Ron that
    is the issue here. I guess if Ron were low enough compared to
    your load, the nonlinearity could be ignored. But in my experience
    there are always other issues. I recall deciding never to bother
    trying another 4066, despite its low Ron, in favor of the 4016,
    which always seemed to behave better in my circuits... though it's
    been a few years and I can't recall the details now. (Sorry!)
    But the distortion issue wasn't ever a big problem, as I recall,
    no matter which switch I used.

    (Note that in really critical work you can often cancel distortion
    by running a second switch to provide an opposite distortion.)

    Best regards,

    Bob Masta

    D A Q A R T A
    Data AcQuisition And Real-Time Analysis
    Home of DaqGen, the FREEWARE signal generator
  15. tempus fugit

    tempus fugit Guest

    Hmmm... interesting about the lower Ron of the 4066 vs. 4016. I guess I'll
    just have to order a few different things up and see which works the best
    for my application. The parts I'm looking at have an Ron of 5 - 10 ohms (for
    the DIP) and some of the SMDs are 0.3 - 0.5. I'll probably start with the
    DIPs since they're easier to experiment with.

    Thanks again Bob.
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