Hi,
What's the status on good decoupling electrolytes these days? Is there still a need for ceramics and 'lytts in paralell?
What I'm decoupling are CS8414 s/pdif receiver, PCM1704 and PCM1794 DACs, and opamps like 2134/2227/5534/2604.
The perfect decoupling, I guess, would be a cheap and compact SMD 'lytt that goes well into the MHz range. Any personal experiences you care to share with me on this one?
Greetings,
Børge
What's the status on good decoupling electrolytes these days? Is there still a need for ceramics and 'lytts in paralell?
What I'm decoupling are CS8414 s/pdif receiver, PCM1704 and PCM1794 DACs, and opamps like 2134/2227/5534/2604.
The perfect decoupling, I guess, would be a cheap and compact SMD 'lytt that goes well into the MHz range. Any personal experiences you care to share with me on this one?
Greetings,
Børge
I know, but why is the ceramic specified in the first place? If the electrolyte alone is able to produce high-frequency supply current at a low impedance, is there a need for the ceramic at all?
One good electrolyte means less routing and thus less parasitic inductance between decoupling and power pin.
One good electrolyte means less routing and thus less parasitic inductance between decoupling and power pin.
All capacitors have series inductance causing them to have a rising impedance above their self-resonant frequency. Unfortunately, an electrolytic's resonant frequency could be as low as a few tens of kilohertz. Bypassing the electrolytic with another capacitor is the traditional method of more nearly approximating to an ideal capacitor. Ceramic capacitors are usually specified by (American) IC manufacturers because the USA has a large ceramic capacitor industry. Plastic capacitors are equally useable as bypasses.
This thread may be of some interest:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=29259&perpage=15&pagenumber=1
http://www.diyaudio.com/forums/showthread.php?s=&threadid=29259&perpage=15&pagenumber=1
Most ludicrous post of the week?
"Ceramic capacitors are usually specified by (American) IC manufacturers because the USA has a large ceramic capacitor industry. Plastic capacitors are equally useable as bypasses. "
Do you really believe this? Does this include decoupling for digital circuits? Doesn't the US make film caps as well, and the Japanese ceramics? Are you just trying to be funny or are you really that confused? I think 20 minutes with a search engine could help you learn the difference between ceramic and film caps........ it might be time well spent. I will even give you a big head start.
http://www.faradnet.com/
"Ceramic capacitors are usually specified by (American) IC manufacturers because the USA has a large ceramic capacitor industry. Plastic capacitors are equally useable as bypasses. "
Do you really believe this? Does this include decoupling for digital circuits? Doesn't the US make film caps as well, and the Japanese ceramics? Are you just trying to be funny or are you really that confused? I think 20 minutes with a search engine could help you learn the difference between ceramic and film caps........ it might be time well spent. I will even give you a big head start.
http://www.faradnet.com/
Fred, if you disagree with a post, it is sufficient simply to give reasons why. Personal abuse is unnecessary.
The source of information that led to my remark was:
"There is an interesting geographical factor involved, though. US manufacturers are stronger in ceramics than in film, whereas in Europe the situation is reversed. Therefore, US design engineers will go for film. So far so good, but the majority of digital ICs sold throughout the world are of US manufacture, and their application notes are written by US applications engineers. These, quite naturally, specify ceramic capacitors when it comes to recommended practice for IC decoupling. As a result, the suppliers of monolithic ceramics are riding on the back of an enormous hidden subsidy as designers worldwide take the easy route of sticking recommended parts in their circuit designs."
This quote was taken from a discussion of ceramic versus film capacitors on p79 of: "The Circuit Designer's Companion" Tim Williams. Butterworth-Heinemann. (1991) ISBN 0 7506 1756 X
The source of information that led to my remark was:
"There is an interesting geographical factor involved, though. US manufacturers are stronger in ceramics than in film, whereas in Europe the situation is reversed. Therefore, US design engineers will go for film. So far so good, but the majority of digital ICs sold throughout the world are of US manufacture, and their application notes are written by US applications engineers. These, quite naturally, specify ceramic capacitors when it comes to recommended practice for IC decoupling. As a result, the suppliers of monolithic ceramics are riding on the back of an enormous hidden subsidy as designers worldwide take the easy route of sticking recommended parts in their circuit designs."
This quote was taken from a discussion of ceramic versus film capacitors on p79 of: "The Circuit Designer's Companion" Tim Williams. Butterworth-Heinemann. (1991) ISBN 0 7506 1756 X
Christer said:
Thanks Christer, what a nice little talk on grounding!
The whole reason I'm asking is that all those capacitors do take up quite a bit of board area.
The easy cases are decoupling caps that fit power pins at chip corners. Then I put the ceramics right next to the power pin and let ground be a large plane under the chip. I want as few holes in this plane as possible so that return current doesn't have to go around obstacles.
But I'd like to cut down on the number of caps on the pad sides of chips. Take PCM1704 and PCM1794 which need a little forest of capacitors each. And the more of them I have to put in parallel, the larger my ground-plane holes (to route between SMD caps and power pins) and the larger my series inductance between cap and power pin. Caps on the pad sides don't have the easy end-of-chip immediate access to the common power plane.
I've put a lot of care into my layout, so don't come here and tell me this is only an RF problem
Wouldn't it be nice with something like an SMD os-con with a ceramic 0805 embedded in the black base?
Greetings,
Børge
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.