Connect with us

Switcher impulse noise reduction...

Discussion in 'Electronic Basics' started by Dave, Nov 8, 2006.

Scroll to continue with content
  1. Dave

    Dave Guest

    Does anyone have any experiance removing the switching noise from the
    output of a switch mode power supply? I have tried in vain but think
    there is a way to actively filter it out somehow. We're talking about
    a 20mv peak to peak noise spike at ~ 60 KHz or so. A picture is worth
    a thousand words but I cant see a way to attach one in this group. Any
    help would be great! I tried this circuit on the problem but it had no
    effect...

    http://www.wenzel.com/documents/finesse.html

    Thanks.
     
  2. I've not done anything along these lines, but it seems that many of
    the techniques used to clean up alternator noise in cars would apply.

    So the first thing would be to make sure the supply is well shielded.
    Don't just bring the wires out of the case through a hole, seal it
    and bring them out through feedthrough capacitors. Figure out where
    the noise is being radiated from, if it's via the line cord, then that's
    where you'll need to do the work. Another trick for cleaning up alternator
    noise was to put a trap (a parallel LC network) in series with the
    voltage coming out of the alternator, and by adjusting it you'd null
    out the noise. I would think the same scheme would work if the issue
    is noise coming off the output lines. Some external commercial supplies
    for radios allow the frequency of the supply to be shifted a tad, to move
    the harmonic off a desired frequency. That might require more work,
    since you need to deal with the existing supply, but it would be a solution
    if the problem is on a specific frequency rather than a widespread hash.

    Michael
     
  3. Before attempting external filtering, you should do what you
    can to reduce the noise produced, internally, and also what
    you can to maximize the noise containment inside the
    switching process. Are either of these possibilities
    available to you?
    You can post picture attachments in the
    alt.binaries.electronics.schematics group, if you have
    access to an ISP that provides a news server. Otherwise you
    might post them to some photo service on the web, and
    provide us with a link.

    60 kHz noise sounds like it comes from a switching
    regulator. Is this what you are working with, or is there
    such a beast, nearby? The closer you can get to the source
    of the 60 kHz, the easier the fix will be. Once that noise
    gets out into the surroundings, it gets harder to surround
    and contain it.
    More important than a picture of the noise would be a
    schematic of the supply system (not only the output you are
    worrying about, but any other parts that might be making
    noise that is getting through yours).
     
  4. Guest

    Dave,

    What type of SMPS is this (buck, boost, etc)? Did you design and build
    it? A schematic would be very helpful. But a lot could also depend on
    the PCB layout, even with a "good" schematic design.

    I don't call myself an expert, but have learned a few things by reading
    manufacturers' application notes (AND http://www.genomerics.org), and
    designing and building several medium-power boost-mode SMPS units.
    I'll try to generalize some of what I think I know:

    First of all: 20 mV p-p is not considered to be too terrible. Many
    commercial units have specs with "100mv ripple plus noise". However,
    if you're careful, you "should" be able to get it down into the
    several-hundred-microvolt p-p range, fairly easily.

    To not induce noise, or have noise induced, certain loops should have
    as small an enclosed area as possible. Look especially at the loops
    with large and/or highly-dynamic currents. Look up Maxwell's Equations
    (and/or Faraday's law).

    PCB traces should be "wide", and as short as possible, usually. YMMV.

    Use a "star" grounding scheme, with no grounds running together,
    anywhere. Instead, they should only touch each other at the ground
    side of the main input cap (assuming this is an "off line" supply, i.e.
    it runs from AC mains, with a transformer, rectifier bridge, and
    large-ish input capacitor).

    Keep the feedback (and ALSO the ground of any feedback voltage-divider
    used) away from everything else, except at the star ground point.
    Also, if possible, take the feedback from a point that's AFTER any
    output cap and output filtering.

    Your inductors should be toroids. Look at the JWMiller 2300 and 2200
    series, if your output pushes less than 15 amps or so. (Unless you
    like winding your own, you may have to base other parts of the design
    on what uH are available at the currents you need.)

    If you have a large diode connected to the SW output of a switcher IC,
    try designing an RC snubber to go in parallel with the diode; maybe
    something like .0022uF in series with 470 Ohms, with the resistor
    connected to the diode's cathode and the (film) cap to its anode. The
    actual R and C values will depend on the rest of your circuit.

    You might also want to consider using a CLC lowpass "Pi" filter on the
    output. If there's already a large-ish capacitor from output to
    ground, you could add a high-current inductor in series with the
    output, and another of the same cap, to ground, after the inductor.
    Choose L to make 1/[(2Pi)LC] small; certainly less than 5 or 10 Hz,
    while still comfortably carrying your maximum output current.

    Look at your big diode(s). Try something faster (or maybe even
    slower?), or with lower Vf, etc etc. (And they shouldn't get very
    warm, if they're not spending too much time in-between the on and off
    states.)

    Look at your big electrolytic capacitors. You want them to have low
    ESR (Equivalent Series Resistance). If this is through-hole stuff,
    maybe look at something similar to the Nichicon "UHE" series (see
    www.mouser.com, at al).

    You could also try adding a linear regulator to the output (with
    appropriate input, output, and feedback Cs and Rs, of course).

    You could also try adding a film or ceramic cap in parallel with your
    main input cap, say 0.22uF (or whatever helps).

    Get LT-Spice (free from www.linear.com), and simulate the circuit!
    Remember to add realistic series resistances to all of the capacitors
    and inductors. You can get a good idea of what "should" work
    well-enough, by trying it in LT-Spice. But your PCB layout and the
    types of components used will still need to be correct, to get good
    results from your actual circuit.

    Good luck.

    - Tom Gootee

    http://www.fullnet.com/u/tomg
     
  5. Guest

    Dave,

    (Are you running the SMPS with little or no load current? My data
    below suggests that at smaller load currents, some SMPSs could have
    increased high-frequency output ripple voltage.)

    I just checked the LT-Spice simulation for one of my SMPS designs, for
    you. This one is an off-line supply, with a 56-Watt 36v-Center-Tapped
    transformer (with circuit switchable for 115vac/230vac), bridge
    rectifier, NTC inrush-limiter (CL11, 0.7 Ohms --> 0.14 Ohms), large
    cap, and then my LT1270A-based boost-mode converter circuit running at
    60 kHz, which takes the approximately 28vdc from the filter cap and
    generates 36vdc, which it puts through an LD1084 (simulated with
    LT1084) Low-Dropout (LDO) Adjustable Linear Regulator, to get 35 vdc,
    and then uses an LM1875T Power Amplifier (simulated with a tweaked AD
    OP275 opamp model) to make a "rail splitter", creating isolated
    +17.5vdc, a (new/pseudo) Ground, and -17.5vdc.

    I ran a simulation with a 20-Ohm resistor from each output rail to the
    new ground. For a reference, I attached the Spice "GND" to the neg side
    of the main input cap, for this run, so the numbers would correspond to
    using it as a 35v/17.5v/Gnd dc supply, with 40 Ohms between 35v and
    Gnd. (So, actually, the ripple voltage numbers below would be halved,
    if looking at each of my new +/-17.5v rails wrt the new Gnd.)

    Anyway, at the main filter cap (33000uF Panasonic ECOS1EP333DA), with
    the stated load, there is roughly 27.4vdc, with about 386 mV p-p of
    120Hz ripple with about 41 mV p-p of 60 kHz ripple riding on it.

    Then, after the LT1270A-based SMPS circuitry, which has an
    appropriately-tuned snubber around the TO-220 Schottky (and a
    specially-tuned RC network from the Vc pin to Gnd), and a 2200uF/50v
    Nichicon UHE-series output cap to Gnd, and a 10uH (JWMiller 2201-V
    toroid, 12.5A, DCR .007) inductor in series with the output, followed
    by another 2200uF UHE cap to Gnd (i.e. forming a low-pass filter),
    there is 36vdc at the output, with about 7.2 mV p-p 120 Hz ripple with
    about 150 uV p-p of 60 kHz riding on it.

    But, AFTER the subsequent LT1084 regulator, which had an average output
    current of 916 mA, for this 40-Ohms-total load (and 219uA p-p 120 Hz
    ripple current with 335 uA p-p 60 kHz ripple current riding on it),
    there is 35vdc with about 270 uV of 120Hz ripple voltage, with about 5
    or 10 uV of 60 kHz ripple voltage riding on it.

    SO, the LDO post-regulator circuitry really helps a lot! It smoothed
    the 120 Hz ripple voltage from 7.2 mV p-p to 270 uV p-p, and smoothed
    the 60 kHz ripple voltage from 150 uV p-p to about 5-10 uV p-p (too
    small to really measure accurately, with the LT-Spice cursor, even with
    the display fully magnified).

    Here are some details of the LT1084 regulator circuit's configuration,
    in case you need them (assuming you're just wanting to make an "add-on"
    for some existing SMPS's output; You'll have to redesign some/most of
    it, of course.):

    The LT1084 TO-220 three-terminal regulator's input is preceded by the
    2200uF Nichicon UHE-series cap to ground, as mentioned farther above
    (part of the SMPS's output filter). The regulator also has a resistive
    voltage divider from Out to Gnd, to set the output voltage, consisting
    of a 1K between Out and Adj and 25.5k in series with a 19-turn 1K
    trimpot (523 Ohms nominal) from Adj to Gnd (to set it at 35v
    output-level). There is a 0.33uF Metallized Polypropylene cap from Adj
    to Gnd. There is also a 1000uF/50V UHE-series capacitor from out to
    ground, in parallel with a 0.22uF X7R ceramic cap. (Just FYI, after
    that, after the "rail splitter" section, but still basically between
    the regulator's Out and Gnd, there are some more caps: a 10uF
    electrolytic in parallel with a 0.22uF X7R ceramic from the regulator's
    Out to the middle rail, with another 10 uF in series with 1 Ohm, in
    parallel with a 0.22uF X7R ceramic in series with 1 Ohm, from the
    middle rail to Gnd. There is also a 7805-type regulator circuit
    operating between the top and middle rails, which creates +5 vdc wrt
    the middle rail.)

    Just for completeness' sake: The rail-splitter itself uses a National
    LM1875T (or LM675T) TO-220 5-pin power amplifier IC, with a preceding
    22k/22k divider between top and bottom rails (i.e. regulator's Out and
    Gnd), with its + input connected directly to the center of the divider
    and its - input connected through 1K to the same point (center of the
    divider). It has 15K, in parallel with a 33pF NPO ceramic cap, between
    its output and its - input. The LM1875T's + and - power supply pins
    are connected to the top and bottom rails (i.e. LT1084 regulator's Out
    and Gnd). The LM1875T's output pin is the new pseudo-Gnd, making the
    top rail the new +17.5vdc and the bottom (formerly Gnd) rail the new
    -17.5 vdc. [One "secret": To try to stop the LM1875T from dissipating
    so much power while idling (they get _quite_ hot, while idling, even
    with a 6.8degC/W heatsink, with no fan), there is a 470 Ohm 1W resistor
    from its output (middle rail) to Gnd (the bottom rail), which makes
    symmetric, and minimizes, the draw through the LM1875T's power-supply
    pins during low-current operation, in this configuration.]

    For fun, I changed the load resistors from 20 Ohms to 500 Ohms each,
    which made the LT1084's output only about 76.5 mA: With that (1K total)
    load, there was about 28.6 vdc at the main filter cap, with roughly 50
    mV p-p of 120 Hz ripple with about 15 mV p-p of 60 kHz riding on it.
    The 36vdc just BEFORE the regulator had only about 375 uV of 120 Hz
    ripple, BUT had about 4.5 mV of 60 kHz ripple riding on it. And AFTER
    the regulator, there were lower-than-120 Hz components (couldn't see
    any above about 100 Hz) of about 10 uV p-p, and a 60 kHz component of
    about 45 uV p-p.

    Of course, then I felt compelled to try it with something fairly-near
    the maximum continuous design current. So I stuck in two 10-Ohm
    resistors, one from each rail to the new Gnd: The 20-Ohm-total load
    makes the LT1084's output push an average of about 1.79A, with about
    750 uA p-p 120 Hz ripple current with 750 uA 60 kHz ripple current
    riding on it. The voltage at the input filter cap is 26.4 vdc average,
    with about 61 mV p-p 120 Hz and 44 mV p-p 60 kHz, while the voltage
    after the SMPS section, and just before the regulator, is 36vdc average
    with about 18 mV 120 Hz ripple and about 260 uV p-p 60 kHz ripple.
    AFTER the linear regulator, there's 35vdc average, with about 665 uV
    p-p 120 Hz ripple and about 15 uV p-p 60 kHz ripple.

    And that lead me to try it at almost double the maximum continuous
    design current, with two 5-Ohm resistors as the load. (I think I smell
    something burning.) It's definitely much worse, this way: The load
    resistors are carrying about 3.49A each, with about 3.5mA of 120 Hz
    ripple current and 12.3 uA of 60 kHz ripple current. The LT1084 is
    pushing about 3.53A, with 26.7mA p-p of 120 Hz ripple current and about
    4.7mA p-p of 60 kHz ripple current (SO, the rail-splitter must be
    smoothing the ripple currents, quite a lot, under these conditions.).
    The voltage at the input filter cap is about 24.75 vdc, with about 1.1
    Volt p-p 120 Hz ripple and about 42 mV p-p 60 kHz ripple. After the
    SMPS section, before the LT1084 regulator, the voltage is 36vdc average
    with about 48.8 mV p-p 120 Hz ripple and 550 uV p-p 60 kHz ripple.
    After the LT1084 regulator, the voltage is down to about 34.88vdc
    average, and still has about 35 mV p-p of 120 Hz ripple and about 95 uV
    p-p of 60 kHz ripple. That's more than _52 TIMES_ the 120 Hz ripple
    that we saw at half this load! Yikes!

    Sorry for the long post. I hope this helps someone.

    (By the way, I manufacture these and use them in a larger piece of
    equipment. But I could also sell them as separate power supply boards,
    if anyone is interested. They're currently done on a 4"x6"
    single-sided PCB (with 2.5-inch max height), and include two 40mm fans
    (.08A @ 12V each). One fan is on-board, for cooling the power supply
    itself, and one is for off-board, wherever needed.)

    Regards,

    Tom Gootee

    http://www.fullnet.com/u/tomg
     
  6. Chris

    Chris Guest

    Hi, Dave. The advice from other posts is worth following. I'd like to
    add a little bit.

    Many times, especially with mixed digital/analog designs, you may have
    to live with switching noise on the power supply lines (and 20mVp.p. is
    pretty good, actually). This can be a big problem, because the great
    power supply rejection ratios specified for many analog ICs are sadly
    specified for 100Hz/120Hz power supply noise. The rejection ratio
    drops dramatically as frequencies increase.

    Star ground layout and sufficient attention to high frequency
    considerations are both very important under all circumstances. But
    once you get past that, you can minimize the effect of power supply
    noise by locally bypassing the power supply at the IC with an L-C or
    R-C filter. A 22 to 47 ohm resistor and an 0.1uF ceramic or 1uF
    tantalum cap can work wonders here, like this (view in fixed font or M$
    Notepad):

    |
    | Vcc +5V
    | | |
    | .-. .-.
    | | |47 ohm | |47
    | | | | |
    | '-' '-'
    | | +|| | +||
    | o---||--. o----||--.
    | |\| || | |\| || |
    | -|-\ 1uF | -|-\ 1uF |
    | | >------)--- | >-------)---
    | -|+/ | -|+/ |
    | |/| ||+ | |/| |
    | o---||--o | |
    | | || | o--------'
    | .-. 1uF | |
    | | |47 ohm| ===
    | | | | GND
    | '-' |
    | | |
    | Vee ===
    | GND
    (created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

    The op amp on the left has split supplies, and the one on the right is
    single supply.

    This local bypassing at the IC is sometimes the most practical way of
    solving the switching power supply / power supply noise sensittivity
    conundrum.

    Doing this well is almost never easy, but many times, following general
    rules can solve most of your problems. Take the time to read the
    venerable

    AN-202: An IC Amplifier User's Guide to Decoupling, Grounding, and
    Making Things Go Right for a Change
    http://www.analog.com/UploadedFiles/Application_Notes/135208865AN-202.pdf

    also take a look at:
    AN-345: Grounding for Low-and-High-Frequency Circuits
    AN-214: Ground Rules for High Speed Circuits
    http://www.analog.com/en/dcIndex.html

    When you're bypassing ICs, make sure to take their advice under
    consideration. Just placing the R and C just anywhere can actually
    make your problems worse, not better.

    Good luck
    Chris
     
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

-