When looking at circuit diagrams from chip makers, I always see the following construct:
What is this for?
also if I were to replace the polarized capacitor in this diagram with a non-polarized cap what issues could this cause?
Thanks
Might be wrong, but looks like these caps are working as decoupling caps, filtering out unwanted power spikes from the power supply.
I wouldn't know about the non-polarized, I've only used polar.
-Englishscone
King_bob:
When looking at circuit diagrams from chip makers, I always see the following construct:
What is this for?
also if I were to replace the polarized capacitor in this diagram with a non-polarized cap what issues could this cause?
Thanks
If the same values, then there is no change. The polarized capacitor is just a whole lot cheaper!
Paul
Usually this type of bypass capacitor pair
is meant for broadband frequency response.
C10 is usually a polarized electrolytic and
C9 is usually a ceramic. The electrolytic
has inherent series resistance, ESR, and
series inductance. Both of which limit
its frequency response. The ceramic
cap has very good high frequency
characteristics and will make up
for the weakness of the electrolytic.
Herb
The answer to your question can be found if you do the math:
Capacitive Reactance ( in ohms)= 1/ 2Pif* C
| where f= Frequency(Hz), C=Capacitance(Farads)
The 100nF is for high frequency decoupling,
the 1 uF is a filter cap to provide stability.
The is at least one 3.3V breakout bd.(IMU?) that
won't work without the 1uF cap because it can't
source enough current. Caps are by nature
energy storage devices. The 1uF stores enough
energy to allow the load to source enough
current, but typically is simply a power supply
stability component. Google LM7805 or any
LM78XX or LM317 regulator and you'll see the same pair of caps.
combination of caps.
If you do the math, you'll see that at 3 kHz, a 0.1uF (100nF) looks like a 530 ohm resistor , but
at 30 kHz , it looks like a 53 ohm resistor. So if you put that cap on dc line that should not
have any of these frequencies, the higher the frequency, the more it approaches a dead short
for those frequencies. At 3 GHz, that same cap looks like a 530uohm resistor. (more than a
dead short to that frequency)'
The 1uF looks like a 159 ohm resistor to 1kHz frequencies so this cap is more of ripple filter
cap than a decoupling cap. It serves to provide voltage stability.
Not every perturbation is of the same frequency. A sharp spike from a fast output change will require a capacitor that works at a very high frequency to keep it from spreading to the rest of the circuit. A load change (LED going on / off) occurs at a lower frequency.
Every capacitor starts to loose its effectiveness at some high frequency. Technically it starts to look like an inductor.
Ref: Capacitor vs frequency Your should look at the graph that has been hand colored with the yellow showing the combination of multiple capacitors.
The Ceramic and the electrolytic loose effectiveness at different frequencies. So by combining one of each, the circuit benefits from the capacitors over a wider frequency range.
Note: some liberties were taken to simplify the concept.
For high frequency decoupling the stray inductance must be minimal, requiring MLCC ceramic cap for the 100nF decoupler, preferrably surface-mount (much less inductance when over a ground plane than any through-hole cap). And such decoupling caps are ideally a few mm from the chip pin at most.
High frequency decoupling is essential for any digital chip (whether the datasheet mentions it or not).
The larger cap is for lower frequency decoupling which is less critical. This is typically seen on datasheets for ADCs/DACs where the rails need lower noise at signal frequencies, not just high speed decoupling for the digital parts.
See:
http://www.thebox.myzen.co.uk/Tutorial/De-coupling.html
Paul_KD7HB:
If the same values, then there is no change. The polarized capacitor is just a whole lot cheaper!Paul
Some regulators need ESR of the output caps to be higher than some threshold. If you replace electrolytic cap with a ceramic one such regulator may oscillate.
Video on bypass capacitors you should watch:
