Connect with us

What's the best bypass caps?

Discussion in 'Electronic Design' started by Chris Carlen, Aug 11, 2003.

Scroll to continue with content
  1. Chris Carlen

    Chris Carlen Guest

    Greetings:

    Before I go reading datasheets for hours and getting nothing done today,
    do you folks have any "favorite" bypass capacitor makes/models for
    medium speed digital stuff like HC and AC logic devices, CPLDs and FPGAs
    running no more than 200MHz, microprocessors of the less than 100MHz
    variety, etc.?

    I am looking to stock up on some 0805 and 1206 chips, as well as
    through-hole 0.1uF types, and perhaps some 100pF ones for higher
    frequency stuff, to put in parallel with the 0.1uF ones.

    Would you go for ceramic disks as opposed to "mono" caps for the
    through-hole types?

    Are there any quick ways to know that a cap is going to be an effective
    bypass without reading datasheets for hours? (Like getting someone to
    tell you the one they use?) ;-)

    Thanks for suggestions.


    Good day!




    --
    _______________________________________________________________________
    Christopher R. Carlen
    Principal Laser/Optical Technologist
    Sandia National Laboratories CA USA
    -- NOTE: Remove "BOGUS" from email address to reply.
     
  2. Ceramic multilayers.

    John Larkin made some comments about decoupling caps lately. Something
    about appnotes and manufaturers thinks their little 25 cents chip is
    the centre of the universe. I have another thing to add, take a nice
    board loaded with chips, make sure that it works, and then start cutting
    away decoupling caps until it stops functioning. Just do it. It's
    a nice experiment. Then *forget* the outcome of the experiment, and use
    plenty decoupling caps in your design ;)
     
  3. If you are in control of the PCB design, then you probably realize that the
    ground traces/ground planes can be just as important as the choice of bypass
    capacitors, if not more important.
     
  4. John Larkin

    John Larkin Guest

    I'd go for 0.1 uF 0805 as the universal cap. There's no reason to use
    a lower value for "faster" situations, nor any reason to put a smaller
    value in parallel with the .1s... they all have about the same ESL, so
    all look about the same at high frequencies.

    Leaded caps will have a lot more lead inductance, which will matter
    when dealing with fast digital stuff and extreme analog bits. The only
    real advantage of the mono parts is that they're smaller and available
    in square shapes, so you can keep the leads shorter to the board.

    Thru-hole is dead: time to move on. Even 1206s are becoming dinosaurs.

    For clean, fast stuff, use a multilayer board with a thin (say, 0.005"
    or less) dielectric between the Vcc and ground planes, then scatter a
    moderate number of the 0.1 uF 0805s around. The planes themselves are
    the best bypass cap.

    John
     
  5. I'm a big fan of Panasonic multi layer ceramics in your favorite
    flavor of SMT (so you can mount right at a pin and via straight to a
    ground plane). .1uF in 0402 - 0805 for me. Digi-Key has 'em.

    I would think that if you want to get super duper tweaky about
    individual ceramic cap specs, then focus might be more appropriately
    turned towards PCB layout - lead inductances, ground plane/paths, etc.

    Cheers,

    Buck Buchanan
     
  6. Isn't that how "old man Muntz" designed his tv sets? Start taking out
    parts until it stops working, and then put the last piece back in?

    Somehow this bit about datasheets seems a tad overkill. In the old
    days when we went to the wooden floored parts place, nobody asked
    to see the datasheets on the capacitors. I doubt they existed.

    What seems to be lost is common sense. "Oh, it's this frequency"
    we'd say, "then use a .01uF". "It's a bit higher", then it would
    be a .001". Only when you get into UHF would you switch to 470pF bypass
    capacitors, and then you'd try to get the button type (suddenly I can't
    remember the exact term, sort of looked like feedthroughs, but not actual
    feed through) that you screwed into the chassis. Or if the signal is
    going through a panel, use a feedthrough. You knew that as you went up in
    frequency, short leads became fairly important, and those other types made
    it easier to get a low inductance connection to ground.

    In tube days, getting a good ground return was often a bigger issue
    than the specific capacitor, at least in the shortwave and low
    VHF range. If you had to drill another hole, and expend a ground
    lug, one might try to use a longer lead on that bypass than necessary,
    or have problems with ground rings around tube sockets.

    Circuit board makes it easier to get a good ground, so long as
    layout is good. That's one of the points of a ground plane layer,
    so there is a really good ground layer, wherever you need it. Because
    it's a sheet of copper, the inductance between points is much lower
    than if narrow traces were used to wire up all the ground returns.

    I threw out all the paper capacitors a couple of decades ago.
    The ceramic capacitors I pull of scrap boards are fine for most
    bypass uses. Getting too fancy may simply complicate things.

    Michael
     
  7. qrk

    qrk Guest

    I like ceramic MLC's, surface mount in 0603. 0402 is too small for my
    tastes. Ground/power plane is a very important capacitor. Has
    wonderful high freq characteristics. Don't forget bulk capacitance,
    like 10 to 1000 uF - depending on what your doing. This goes for high
    speed digital junk.

    Last 957-pin FPGA we used only had about 20 local 0.1 uF MLC
    capacitors for all the power supplies. Bypass caps sat opposite the
    part.

    AVX has some very nice info on their MLC capacitors.
    http://www.avxcorp.com/SpiApps/spicap/

    Mark
     
  8. Active8

    Active8 Guest

    [snip]
    that's the way i remember the story.
    i think it only becomes important at UHF and above, or when you really
    need low ESR or low inductance. even then, a whole sheet is overkill.
    if you can resonate the cap with it's own L, it's an rf short to ground
    and you can forget about worrying. but that's a UHF thing. i've never
    seen this technique used at 40MHz.

    br,
    mike
     
  9. Bill Sloman

    Bill Sloman Guest

    The multilayer 0.1uF 0805 parts witn X7R dielectric have an impedance
    minimum at of the order of 100MHz, and - back in 1989 - we had an
    object lesson in what this means at 800MHz. One of my guys didn't
    really believe the application notes for the Gigabit Logic GaAs parts,
    and skipped the (1810) 1nF porcelain dielectric microwave capacitor
    (good to about a GHz) that we normally used to parallel the regular
    100nF 0805 decoupling cap.

    We had as much 800MHz on the supply rail at the part with the 100nF
    multilayer in place as we had with it removed. I think we replaced it
    on the layout with a 100pF single layer part with a COG dielectric,
    which at least had a visible effect. The next version of the layout
    had the full complement of capacitors.
     
  10. As all the volume shifts to 0603 and 0402 eventually, the 1206 parts
    of popular value will become more expensive (or you might have to buy
    higher voltage, higher value or better dielectric than you'd otherwise
    require).

    Example: 0.1uF/16V or 25V or 50V cap in high volume from my supplier,
    cheapest dielectric in stock:

    Mfr. Relative price Relative stock
    0402 TY 1.0 1.0
    0603 AT 0.85 94
    0805 AT 0.97 5.8
    1206 VI 4.4 1.1

    Another comparison (Digikey)

    Mfr. Relative price
    0402 MA 1.0
    0603 KE 1.08
    0805 MA 1.5
    1206 KE 2.5

    Not shown is the number of different suppliers, which is large for the
    0603 and 0805 and less at either end. You can see the "sweet spot"
    right now is 0603, with 0402 coming in the future.

    Best regards,
    Spehro Pefhany
     
  11. Both of these caps then seem to have about 25pH of ESL, which
    doesn't sound credible. The way you mount them is more important
    than the properties of the caps themselves. (How *did* you
    mount them then? 25pH of ESL would be quite a trick! A simple via
    between opposite sides of a standard 1.6mm PCB is ten times that!)

    I measured the ESL of a bunch of capacitors some time back.
    A 1206 ceramic standing upright on a ground plane measures
    about 900pH. I'd expect about the same for an 1810 component,
    since it's both 50% longer and wider. ESL does not depend
    on capacitance, nor on dielectric. It's all in the geometry
    of the part and its surroundings.

    Once you have a 100nF ceramic chip capacitor as bypass, there
    is no advantage in adding smaller-valued similarly shaped caps
    for HF. On the other hand, some manufacturers make, e.g., *0612*
    parts, that is, they're short and wide. Syfer comes to mind.
    That *does* help. (Their web site sucks, though :-( )

    Jeroen Belleman
     
  12. Blake

    Blake Guest

    Wow thanks for the thorough analysis. I did look at the Digikey catalog
    myself, and you are right, the 1206's cost a few cents more per cap. Also,
    I am amazed to see that ceramic caps of > 1uF can be found in a 1206
    package.

    Blake
     
  13. Jim Weir

    Jim Weir Guest

    It went further than that. Muntz engineers were amongst the most talented
    folks I've ever come across...

    So you want an RF amplifier? No problem, but since RF and audio are so far
    apart, do a bit of clever decoupling and use the same tube section for RF and
    audio. Hm. Audio only goes up to 3 KHz. if you don't care for hifi, so do a
    little more work and use it to split the sync at 15 kHz...and while you are at
    it, nothing in that tube at 4.5 MHz., so use it as the sound intercarrier
    amplifier. ONE TUBE section for several functions.

    Those people were CLEVER.

    Jim


    (Michael Black)
    shared these priceless pearls of wisdom:

    ->
    ->Isn't that how "old man Muntz" designed his tv sets? Start taking out
    ->parts until it stops working, and then put the last piece back in?
     
  14. I guess Asian sourcing and solid engineering trounces cleverness, as
    they're not around anymore. Or are they?

    Best regards,
    Spehro Pefhany
     
  15. Me neither. I don't think the OP was talking about this frequency
    though, was he? Anyway, whilst getting a short at RF is very nice,
    it's hardly necessary. I'm just left wondering what sort of maximum
    reactance is acceptable for bypassing purposes. Anyone know if there's
    a specific 'figure' to aim for?
     
  16. Sorry to engage in another shameless public display of ignorance, but
    what the hell do these four digit numbers refer to?
     
  17. Bill Sloman

    Bill Sloman Guest

    We didn't have the gear to measure the high frequency impedances, but
    whatever the loss mechanism, the 100nF multilayer X7R clearly had a
    much higher ESR at 800MHz. I always assumed that the dielectric was
    the problem rather than the inductance of the connections. American
    Technical Ceramics made a great deal of fuss about the porcelain
    dielectric in thier microwave capacitors.
    We mostly mounted them on the under-side of the board, on the opposite
    side to the surface-mount parts we were decoupling. At least four of
    the inner layers of the board were devoted to various power planes -
    it went component(IC) side, -2V, -5.2V, +5V/-3.4V, 0V and track
    (decoupling caps) side.

    The -3.4V was for the GaAs logic which was not mixed in with the TTL
    parts, so that ground plane split pretty naturally.

    The via's went straight to the relevant power planes, and were offset
    from the lands for the IC leads and the capacitor contacts - IIRR the
    boards came with the vias filled with solder and covered with
    solder-mask.
    That wasn't our experience - my colleague (technically my project
    leader at the time, though he didn't take his status all that
    seriously) shared your opinion, encouraged by a rather cramped layout,
    and had to recant.
    Sounds like a good idea. Back in 1989, we got what we could from the
    microwave suppliers.
     
  18. Jim Weir

    Jim Weir Guest

    So did I, sir. I started by sweeping out the TV shop in 1958 in the 9th grade.
    Beat the hell out of delivering papers in the rain. Started fixin' them later
    that year, and made a high school job of it. Got pretty darned good at it after
    my junior year and went on to other electroniker pursuits, but never forgot the
    smell of burning selenium (peanut butter) or the sight of horizontal output
    tubes that got so hot the glass sucked in to the plate.

    Nor how far into the plaster wall one's elbow will penetrate when getting bit
    off the yoke.

    As to the drifting parameters and sync buzz, yup. But you could replace every
    double/triple...duty tube in the damned thing with brand new for under $5 and
    bring it back to like-new.

    Jim


    Boris Mohar <>
    shared these priceless pearls of wisdom:

    -> You can stop a car with only one brake caliper engaging. Not very well.
    ->Why am I so cantankerous? I used to service those beasts. I can still
    ->smell the burned selenium stack.
     
  19. Guest

    Hi John,

    Just so you know, the core can (and does) suck 2.5A for a very short
    period of time. Just after the download is complete, as it sets all
    of the flops into their initial state.

    This <http://www.xilinx.com/xapp/xapp189.pdf> may seem like overkill,
    but if that 2.5A spike causes the voltage to droop enough, the chip
    will reset back to its uninialized state.

    Gary
    g w helbig -at- yahoo -dot- com
     
  20. R.Legg

    R.Legg Guest

    It seems a little odd, quoting metric dimensions when the imperial
    part size is quoted. This is not necessary.

    For clarity -

    1210 is ~ .120 x .100 (metric part sixe is 3225 ~ 3.2 x 2.5mm)
    1206 is ~ .120 x .060 (metric part size is 3216 ~ 3.2 x 1.6mm)
    0805 is ~ .080 x .050 (metric part size is 2012 ~ 2.0 x 1.2mm)
    0604 is ~ .060 x .040 (metric part size is 1608 ~ 1.6 x 0.8mm)
    0402 is ~ .040 x .020 (metric part size is 1005 ~ 1.0 x 0.5mm)
    0201 is ~ .020 x .010 (metric part size is 0603 ~ 0.6 x 0.3mm)

    Do we see a useful pattern here?

    Anything smaller than 0805 - check with fab house for their ability to
    handle parts automatically. There may be penalties, due to tape and
    reel incompatability (like trashing every second part on the tape).

    Cost, performance and real estate benefits below 0604 may not be
    naturally occuring either. This is particularly the case on .060 FR4.
    Anyone care to guess why this is the case?

    Check first.

    RL
     
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

-