rodgerwxh
Banned
- Joined
- Aug 24, 2012
- Messages
- 113
- Helped
- 25
- Reputation
- 56
- Reaction score
- 22
- Trophy points
- 1,298
- Location
- North Carolina
- Activity points
- 0
I am learning decoupling capacitor (DeCap) strategies. The basic rules are very straightforward to understand. "Making sure that noises carried in high frequencies will always pass through your decoupling capacitor so that the power supply of the IC has no ripples". Another function is to provide IC with fast and sufficient charge rather than ask the IC to pull charges through the long power lines which may introduce inductance and cause delay in phase. Correct me if I am understanding this wrong.
However, my concern is not about the purpose, its about practical strategy in using the decoupling capacitor.
1. DeCap values
Common suggestion online is use "sufficient", which means capacitance needs to be big enough, but how "sufficient", the bigger the better? Some give values like 0.1uF, 0.01uF, but we know Zc=1/(j*omega*C), which is a freq dependent value, a "fixed" or "sufficient" description is definitely not clear. Although manufactures of ICs suggest decoupling capacitor values, but how did they know that what kind of environment that I am going to use their IC? This is quite uncertain and everything has to be based on experience which is hard for newbies.
2. DeCap positioning
From YouTube or Google or some text books, it suggest to put local DeCap close the the power line of the IC. Also put a Bulk Cap close to the source. These all have its purpose. But I am wondering is it really always good to put DeCap as close as possible? Since it generate current loops and might cause resonances as well. So where to put the DeCaps seems to be another undetermined rules. What if I have many DeCaps on the board? Scatter them out or put them together?
3. With all above said, can I use a varactor as a DeCap? Since its capacitance can be controlled by the biased voltage. Is it more flexible? Any one can tell me if it violates some basic design rules or cause some extra troubles or resources? I also want to learn about the specific advantage and disadvantages of using varactor as DeCaps.
Thanks in advance guys.
However, my concern is not about the purpose, its about practical strategy in using the decoupling capacitor.
1. DeCap values
Common suggestion online is use "sufficient", which means capacitance needs to be big enough, but how "sufficient", the bigger the better? Some give values like 0.1uF, 0.01uF, but we know Zc=1/(j*omega*C), which is a freq dependent value, a "fixed" or "sufficient" description is definitely not clear. Although manufactures of ICs suggest decoupling capacitor values, but how did they know that what kind of environment that I am going to use their IC? This is quite uncertain and everything has to be based on experience which is hard for newbies.
2. DeCap positioning
From YouTube or Google or some text books, it suggest to put local DeCap close the the power line of the IC. Also put a Bulk Cap close to the source. These all have its purpose. But I am wondering is it really always good to put DeCap as close as possible? Since it generate current loops and might cause resonances as well. So where to put the DeCaps seems to be another undetermined rules. What if I have many DeCaps on the board? Scatter them out or put them together?
3. With all above said, can I use a varactor as a DeCap? Since its capacitance can be controlled by the biased voltage. Is it more flexible? Any one can tell me if it violates some basic design rules or cause some extra troubles or resources? I also want to learn about the specific advantage and disadvantages of using varactor as DeCaps.
Thanks in advance guys.