difference between 100nf capacitor and 1 micro farad capacitor

Thread Starter

Kadav

Joined May 11, 2018
130
I was using this setup to turn the motor on and the software was bugging after some time, that the motor was running but when i connected the capacitor C1 ( in my mind for removing the AC noise of the software ) the software was running very well.
My question is how is to chose the capacitor to connect when you want to filter , how to know if it's 100nf , 1 micro farad or 1pf , what happens when you connect a small capacitor or when you connect a big capacitor

Thanks .
1596657146944.png
 

jpanhalt

Joined Jan 18, 2008
10,095
The smaller the capacitor, the lower its ESR (equivalent series resistance) and the faster its response. Thus, one will sometimes see 0.1 uF, 1.0 uF and 10uF (or another series) in parallel. The question, why not use just a 10 uF, as the tolerance of a 10 uF includes 11.1 uF usually? The answer is that response time.

If in doubt, you can do that too. One rarely sees anything less than 0.01 uF (10 nF) insuch designs. Never pF . For one reason, your routing/wiring is probably a few pF.
 

Papabravo

Joined Feb 24, 2006
14,244
The smart alack-y response to the literal question is: "An order of magnitude". That said it should be noted that capacitors behave differently according to the frequency of what ever is across the terminals. All capacitors look like open circuits to a pure DC signal with no AC content. No circuit means no current flow. No current flow means no voltage drop across the capacitor. Are you still with me? The other end of the frequency spectrum is just below blue light. Almost any capacitor you can find in the wild will a have a comparatively low AC resistance (also called impedance) to the flow of AC current at high frequencies. If you are interested in looking at this further we can put some numbers to the qualitative description.
 

dl324

Joined Mar 30, 2015
11,282
My question is how is to chose the capacitor to connect when you want to filter , how to know if it's 100nf , 1 micro farad or 1pf
It depends on the frequency spectrum of the noise you're trying to filter. Sometimes a 100nF cap will be placed in parallel with a 10nF cap (or 1uF).

All capacitors have some parasitic inductance, so while the capacitor's impedance increases decreases with frequency, the inductor's increases.

EDIT: corrected capacitor's impedance which obviously decreases with increased frequency.
 
Last edited:

Papabravo

Joined Feb 24, 2006
14,244
It depends on the frequency spectrum of the noise you're trying to filter. Sometimes a 100nF cap will be placed in parallel with a 10nF cap (or 1uF).

All capacitors have some parasitic inductance, so while the capacitor's impedance increases with frequency, the inductor's increases.
Impedance of a capacitor DECREASES with frequency. At least you got the inductor right.
 

MrSoftware

Joined Oct 29, 2013
1,865
Impedance will decrease as frequency increases initially, then after a specific point the impedance will increase as frequency increases further. The frequency points depend on capacitance and inductance of each capacitor.

 

MrChips

Joined Oct 2, 2009
21,666
First of all, know your units.

Increasing in steps of 10x
1p
10p
100p = 0.1n
1000p = 1n
10n
100n = 0.1μ
1000n = 1μ
10μ
100μ = 0.1m
1000μ = 1m
10m
100m = 0.1
1000m = 1
10
100
1000

Next know that capacitances in parallel add.

C1 || C2 = C1 + C2

Hence 100nF + 1000nF = 1100nF.
But be aware of the following:
A 1000nF capacitor with ±10% tolerance can range from 900nF to 1100nF.
So you would think that putting C1 = 100nF in parallel with C2 = 1000nF makes no difference.
That is not true.
Capacitors have internal resistance and inductance. This makes their response to high frequency signals to be very different. You can put different capacitances in parallel specifically to be effective as a filter at different frequencies.

In other words, it is good to know the frequencies that you want to attenuate.

1596827096691.png

Read more here:
Choosing and Using Bypass Capacitors
https://www.renesas.com/us/en/www/doc/application-note/an1325.pdf
 

BobaMosfet

Joined Jul 1, 2009
1,177
Most of the time it's just trial and error. If a 100nf gets the job done then problem solved.
Steve G
No, no, no.... it is not trial and error. Engineering is about understanding what is going on, and designing to solve specific problems.

1596831078181.png
A smaller capacitor discharges quickly in the face of a pulse of a given frequency. Making the capacitor larger, prevents it from discharging fully so it acts as a buffer against the pulse frequency.

Choosing the correct capacitor is about determining that frequency and choosing a capacitor that cannot react fast enough to sustain the ripple through it, at a basic level. There are other factors, but this is what the TP can grasp at this time.
 

nsaspook

Joined Aug 27, 2009
7,478
...
So you would think that putting C1 = 100nF in parallel with C2 = 1000nF makes no difference.
That is not true.
Capacitors have internal resistance and inductance. This makes their response to high frequency signals to be very different. You can put different capacitances in parallel specifically to be effective as a filter at different frequencies.

In other words, it is good to know the frequencies that you want to attenuate.

View attachment 214214

Read more here:
Choosing and Using Bypass Capacitors
https://www.renesas.com/us/en/www/doc/application-note/an1325.pdf
"parallel specifically to be effective as a filter at different frequencies"

This is a important point. On this board design I could have used two smaller physical sized decoupling caps on separate lands but for ease of hand soldering I used the stack method to parallel bypass capacitors for the 120MHz controller cpu clock and lower frequency I/O switching bypass duties.

Topside


Stacked power bypass caps on the bottom:
 

MrSoftware

Joined Oct 29, 2013
1,865
No, no, no.... it is not trial and error. Engineering is about understanding what is going on, and designing to solve specific problems. <..SNIP...>
I think it depends a good bit on what you're designing. If you're dealing with something like quieting down a a SMPS, or a few of them in the same circuit, sometimes in practice you do just take an educated guess and throw a few different sizes at it and see what happens, and often good-enough is good-enough. I can see how for more sensitive situations you would want to be more precise.
 

sghioto

Joined Dec 31, 2017
1,334
No, no, no.... it is not trial and error. Engineering is about understanding what is going on, and designing to solve specific problems.
The TP doesn't appear to be an engineer so for his specific problem "trial and error" is his best approach. Which he figured out when installing the 100nf cap.
Steve G
 

nsaspook

Joined Aug 27, 2009
7,478
Glad to see someone else does that. I find stacking easier than hand soldering the first component.

Wonder how that works with reflow?
I've never tried reflow stacks. :eek: Two smaller side by side 0402 (or smaller) packages would be better for high frequency bypass duty but I've seen no problems with the added 0805 package stack inductance on four layer boards with a ground plane at the PIC32 controller clock speeds. Just make sure the loop between chip and bypass cap is as short as you can reasonably make it.
 

MrAl

Joined Jun 17, 2014
7,758
I was using this setup to turn the motor on and the software was bugging after some time, that the motor was running but when i connected the capacitor C1 ( in my mind for removing the AC noise of the software ) the software was running very well.
My question is how is to chose the capacitor to connect when you want to filter , how to know if it's 100nf , 1 micro farad or 1pf , what happens when you connect a small capacitor or when you connect a big capacitor

Thanks .
View attachment 214065

From a perspective of a new design with various inductances the cap has to work with those inductances to help to reduce changes on the supply line that might interfere with the turn on and turn off of the MOSFET. Without some sort of filtering there could eb oscillation.
However, sometimes it is just a matter of following guidelines that were created long ago and just going with that.

This circuit looks like it may have been a new design so the designer would have to estimate inductances and the current draw and power supply impedance and stuff like that in order to estimate the size of the cap to be used, or it could have been worked on on the bench.

A good example of following guidelines however is when dealing with TTL logic. The guideline was always use 0.1uf in parallel with 10uf (one for every certain number of TTL chips). TTL is not used as much as it was but that's just an example.

In the experimental scenario we also want to keep radiated EMI down also and capacitors can help there too. Motors can be noisy especially when they have brushes.

To tell you the truth though i would have expected a bit more capacitance in this circuit. Like 0.1uf in parallel with 10uf or even 100uf. It could be that the power source here is every good and the wiring is short. Long wiring tends to require more bypass capacitance because it comes with the price of higher inductance and more EMF radiation.
 

MrAl

Joined Jun 17, 2014
7,758
At my age, I stick to 805 size for hand soldering. Maybe if my first reflow is successful, I will begrudgingly go to 603.
Hi John,

Have you ever tried solder paste, even for hand soldering? That stuff is amazing even when you use a soldering iron. It amazes me that this ugly gray green gook turns into shiny silver. The most significant and important point though is that it flows into the joint very very well, and even if you solder two pins at once it seeks the two pins not in between pins unless you use a REALLY lot of it. I think it is surface tension that makes it flow to the pins and not in between them like you would think, as well as the very free flowing nature of the heated stuff.
 
Top