small resistor in series with bypass capacitor
Mahen K
in series with the bypass capacitor to the ground.
http://picasaweb.google.com/mahen.jnk/PostingImages#5405424967785689106
TIA
-Mahen
Bill Sloman
Both the by-pass capacitors - C1 and C5 - that have the series
resistor are 33uF parts.
It could be that the capacitors need to be protected against in-rush
current when the battery is plugged in, or it could be that the
battery lead inductance is high enough that R1 and R5 are required to
critically damp the resonance between this inductance and the filter
capacitance. This implies an inductance of the order of 100nH which
isn't unreasonable.
--
Bill Sloman, Nijmegen
markp
"Mahen K" <mahe...@gmail.com> wrote in message
news:dc668d51-cfab-4a2c...@x6g2000prc.googlegroups.com...
Probably to add some small ESR (effective series resistance) to the cap to
keep the loop stable. Some supply chips especially LDOs require some ESR in
their output capacitance and some actually specify particular manufacturers
for that purpose.
Mark.
John Larkin
None. This person is obviously a victim of one of the stranger bypass
religions.
John
Tim Shoppa
Specified ESR of a small cheap electrolytic 33uF 10V capacitor would
be an ohm. Yes, there are ones better and ones that are worse.
Tantalums are gonna have a lower ESR, there are some 33uF 10V
tantalums that are less than 0.1 ohms ESR.
Some folks in the past have used tantalums as power bypass caps and
had them blow up on power-up. Some have recommended putting small
resistors in series to prevent blowing up the tantalums but that's
stupid because you put a tantalum in to begin with because low ESR was
important. Like "the old lady who swallowed a fly".
Tim.
Phil Hobbs
It's stupid unless you put the resistor in series with the input supply
and take the output across the cap. You lose a bit in the regulation
department, of course.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net
Hammy
wrote:
>
>"Mahen K" <mahe...@gmail.com> wrote in message
>news:dc668d51-cfab-4a2c...@x6g2000prc.googlegroups.com...
>> In the schematic below, what is the purpose of the R1 and R5 that are
>> in series with the bypass capacitor to the ground.
>>
>> http://picasaweb.google.com/mahen.jnk/PostingImages#5405424967785689106
>>
>> TIA
>> -Mahen
>
Everybody else pretty well stated the obvious its stupid.
Ideally for bypass/ decoupling you want an ideal capacitor these
don't exist so you use the next closest thing polymer for bulk and low
esl / esr ceramics and/ or arrays.
An example of a low esl cap.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=490-4321-1-ND
The cheaper alternative is 0603 or smaller general purpose MLCC's and
standard Aluminums or Tants for bulk.
Dave Platt
markp <map.n...@f2s.com> wrote:
>Probably to add some small ESR (effective series resistance) to the cap to
>keep the loop stable. Some supply chips especially LDOs require some ESR in
>their output capacitance and some actually specify particular manufacturers
>for that purpose.
It can help to have some ESR in the bulk bypass cap path even if one
is not using an LDO which requires this for stability.
If you build a bypass system with only low-ESR capacitors, you can end
up with a resonance effect - the capacitance interacts with the
inductance of the PC-board traces, creating an LC tank circuit (or a
set of them) which can resonate at one or more frequencies. This
resonance reduces the effectiveness of the bypassing, and can increase
noise and voltage fluctuations at the resonant frequencies involved.
The "better" your bypass capacitors (the lower their ESR) the higher
the Q of the resonance.
Adding bulk bypass capacitance, with a modest amount of ESR in the cap
(or in a resistor in series with it) "swamps" the resonance, reducing
the Q, and thus decreasing the noise level and voltage-fluctuation
problems. Once again, if your bulk bypass caps are "too good" (e.g.
tantalums with low ESR) adding a separate swamping resistor can help
matters.
--
Dave Platt <dpl...@radagast.org> AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!
Gerhard Hoffmann
>It's stupid unless you put the resistor in series with the input supply
>and take the output across the cap. You lose a bit in the regulation
>department, of course.
Not neccessarily. The series inductance of the 33u happily resonates
with the neighboring 0.1u and this is a parallel resonance. The better
the 33u is Ohm-wise, the worse the decoupling is at the resonant frequency.
Watch page 6 or 18 of this:
<http://www.hoffmann-hochfrequenz.de/downloads/experiments_with_decoupling_capacitors.pdf>
regards, Gerhard
Rich Webb
Very interesting! Thanks for posting this.
--
Rich Webb Norfolk, VA
Joel Koltner
news:1np8g5d1a5pr9oeoh...@4ax.com...
> Very interesting! Thanks for posting this.
Be on the watch for Larkin to jump in here and tell you that, if you have a
power and ground plane, none of what's in that presentation (interesting thing
that it is) really applies very much: The plane structure represents a very
broad, low reasonance source.
Bill Sloman
wrote:
> "Rich Webb" <bbew...@mapson.nozirev.ten> wrote in message
At low frequencies - an if the original circuit designer knew what
they were doing, that 0.1 ohm would damp a resonance at around a
couple of MHz - the power and ground planes would present a simple
capacitance and wouldn't provide any useful damping.
--
Bill Sloman, Nijmegen
Jamie
That is used as a snubber, to reduce transient pulses.
The cap is what responds to the narrow timed pulse and
the R limits the max current.
Since there is a zener there to cover most of the over
voltage on that input, conventional zeners may not be
fast enough to respond to some transient pulses, so, a
snubber in this case helps with the initial pulse.
In automotive circuits, noise from the ignition system would
most likely get absorbed in this circuit..
Hammy
wrote:
>In article <7mifjaF...@mid.individual.net>,
>markp <map.n...@f2s.com> wrote:
>
>>Probably to add some small ESR (effective series resistance) to the cap to
>>keep the loop stable. Some supply chips especially LDOs require some ESR in
>>their output capacitance and some actually specify particular manufacturers
>>for that purpose.
>
>It can help to have some ESR in the bulk bypass cap path even if one
>is not using an LDO which requires this for stability.
>
>If you build a bypass system with only low-ESR capacitors, you can end
>up with a resonance effect - the capacitance interacts with the
>inductance of the PC-board traces, creating an LC tank circuit (or a
>set of them) which can resonate at one or more frequencies. This
>resonance reduces the effectiveness of the bypassing, and can increase
>noise and voltage fluctuations at the resonant frequencies involved.
>The "better" your bypass capacitors (the lower their ESR) the higher
>the Q of the resonance.
>
>Adding bulk bypass capacitance, with a modest amount of ESR in the cap
>(or in a resistor in series with it) "swamps" the resonance, reducing
>the Q, and thus decreasing the noise level and voltage-fluctuation
>problems. Once again, if your bulk bypass caps are "too good" (e.g.
>tantalums with low ESR) adding a separate swamping resistor can help
>matters.
I think you would be better off inserting a small ferrite bead with a
small DCR if you have SRF problems? That is what I've been doing beads
are cheap and small. Or at the very least CRC pi filter instead.
To me it makes no sense to place a small R 1 ohm 0.1 ohm. Then you may
just be going from SRF problem to noisy rail large spikes generated by
the esr every time the gates switch.
John Larkin
OK, here it is:
A 0.1 uF cap, TDRd along an asymmetric coplanar waveguide (no via
parasitics) at various time scales. Note the ringing at about 1 GHz.
ftp://jjlarkin.lmi.net/TDR_0.1uF.JPG
ftp://jjlarkin.lmi.net/TDR_0.1_slow.JPG
And here's the TDR of a 3.3 volt power pour on a board we received
just this morning:
ftp://jjlarkin.lmi.net/TDR_3Vplane.JPG
The positive pulse is the 50 ohm hardline from the sampling head to
the board. The drop to near zero ohms is indistinguishable from a
hard-shorted SMA of similar geometry. The plane looks like an
essentially perfect 2.3 nF capacitor. Based on experience with similar
boards, loading bypass caps anywhere will make it look like a bigger
perfect capacitor.
On a multilayer board with good power planes, bypass cap resonances
will be undetectable.
I plan to use the SMA on the plane to check plane noise when the board
is running, too.
John
Joel Koltner
John Larkin
<zapwireD...@yahoo.com> wrote:
>Very nice job John, thanks for posting those results!
>
Actually, if you look at that last shot, at the little flat spot just
as the coax impedance transitions onto the power plane, one could
suspect that the current is propagating into the plane as a
point-driven sheet transmission line, around 1 ohm maybe, and the
little wiggles in the flat region are the various reflections off the
sides and corners of the board.
After all the reflections die down, the plane acts like a capacitor,
giving the charging ramp to the right.
John
Yzordderrex
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Thu, 19 Nov 2009 09:39:53 -0800, "Joel Koltner"
>
It's to move the output esr zero lower to stabalize the loop
regards,
Bob
N9NEO
Joel Koltner
news:pc1bg59ksuk8f3uaa...@4ax.com...
> one could
> suspect that the current is propagating into the plane as a
> point-driven sheet transmission line, around 1 ohm maybe, and the
> little wiggles in the flat region are the various reflections off the
> sides and corners of the board.
Does the shape of the wiggles change if you handle the board? (E.g., run a
few fingers along the edges.)
John Larkin
<zapwireD...@yahoo.com> wrote:
>"John Larkin" <jjla...@highNOTlandTHIStechnologyPART.com> wrote in message
>news:pc1bg59ksuk8f3uaa...@4ax.com...
>> one could
>> suspect that the current is propagating into the plane as a
>> point-driven sheet transmission line, around 1 ohm maybe, and the
>> little wiggles in the flat region are the various reflections off the
>> sides and corners of the board.
>
>Does the shape of the wiggles change if you handle the board? (E.g., run a
>few fingers along the edges.)
>
Our copper doesn't extend to the board edge, so touching the edge has
no effect. The 3.3 volt plane is layer 2, so putting my hand on the
top adds a little capacitance and changes the assumed reflection
pattern a little bit, but not in any obviously causal way.
The SMA-connected TDR doesn't work extremely well trying to TDR what
may be a sub-ohm sheet transmission line structure. This is a *very*
low impedance structure and I'm guessing that Z drops with frequency
as the "splash" spreads out radially from the SMA. Most of the
features are in the 3-10 GHz sort of range, way higher than the SRF of
bypass caps and vias.
I'll repeat it on the loaded board. I should have the first one this
afternoon, just about 24 hours after the bare board was received.
John
JosephKK
Could very well be. They might be "fuses" to protect against caps
that have issues with shorts.
John Larkin
Or they might just be silly. There is an amazing variety of really bad
schematics available on the Web.
John