crippledchicken said:
You may find that self-defeating. Generally, you try to place a bypass/decoupling capacitor as close as you can to its target. The enormous size of a polysterene compared to, say, a multilayer ceramic may add an excessive amount of trace and/or lead inductance. Polystyrene capacitors are great but I would use them elsewhere.
hi johnf, i just wanted to verify that we are talking about the same type caps and uses. i was thinking of using them to bypass +/- op amp power pins to ground. are these the type you are refering to? thanks crippledchicken http://www.mouser.com/index.cfm?han...oductid=313905&e_categoryid=5&e_pcodeid=02310
Just for fun, I measured the 100nF X7R ceramics I have been using. These are so small, I can place them in a 0.1 inch gap between IC sockets and even in the voids inside an IC socket. I calculate they have about 1/40th the volume of the 10nF parts you were considering. I don't want to imagine how big a 100nF polystyrene would be, or how much space ten smaller values would take.
hi, seems i've read that ceramics were not the ideal choice to use that polypropylenes such as the orange drops, stacked pp caps and wimas, sometimes film and foil, are the better choices. i guess maybe it might just be pretty much circuit dependent , individual preferences etc. i just didn't realize the polystyrenes were so much larger than the rest. i read so many articles, and different opinions on whats the best, that it gets pretty confusing hehe but, oh well i'm learning and appreaciate all the great help on the forums thanks, crippledchicken 
Multilayer ceramic caps are a wholly different animal than "regular" ceramic disc capacitors. They are excellent for power supply bypassing purposes. Polypropylene, polystyrene, polycarbonate, etc., film caps are not as good for this application due to their large physical size as stated. They are better for audio signal coupling (if you must have them at all), or for use in filters, etc.
-Ti
-Ti
hey guys, I was able to get a scrap PCB with square yellow ceramic caps in them (about 4mm square with a dab of green paint on the top, much like a regular ceramic disk but square and yellow). dunno what kind of ceramic it is really but I used them in my 5 way LR crossover by Rod, 22nF each and don't have any problems.
I have attached a picture comparing a couple 100nF multilayer ceramics to a 100nF metalized polypropylene, the most compact, stable capacitor you might use in an active filter. (25.4 mm = 1 inch). Just look at all that extra inductance on the right side of the picture.
I am trying the decoupling scheme suggested by Douglas Self. If it checks out, I will probably switch to even smaller SMDs soldered across the V+/V- supply rails.
Out of curiosity, it a good idea to use parts with an X7R dielectric for decoupling audio opamps? I have read that some Z5U parts are microphonic, though I would think an opamp should manage to reject such noise.
I am trying the decoupling scheme suggested by Douglas Self. If it checks out, I will probably switch to even smaller SMDs soldered across the V+/V- supply rails.
Out of curiosity, it a good idea to use parts with an X7R dielectric for decoupling audio opamps? I have read that some Z5U parts are microphonic, though I would think an opamp should manage to reject such noise.
Attachments
Take a look at the info on www.avxcorp.com.
Somewhere there is a table showing the properties of various types of caps. Of interest to this thread is the comparative bandwidth. If you go by the ideal caps in many Spice models or various graphical representations of the response curves generated by filter circuits, you might get the impression that the action of a particular value cap just goes on forever as you go up the frequency scale. Apearently this is not the case - rather after some point the AC resistance begins to rise again. Linsely-Hood (The Art of Linear Electronics) shows a graphical example for some electrolytics where the rise begins below 20kHz. Electros, Al or tantulum are most limited in this respect. The AVX data are not as exact but make it clear that all the "poly" caps have a bandwidth much more limited than ceramic or silver-mica.
This may not be important if you have a metal enclosure that is earthed, thus providing RF shielding. Nonetheless, if you want maximum bandwidth as a precaution against RF in the rails, ceramic or silver-mica gives you a greater margin of safety. The standard arrangement is an electro parralelled with a ceramic to maximize the range affected from very low AC up through RF.
Somewhere there is a table showing the properties of various types of caps. Of interest to this thread is the comparative bandwidth. If you go by the ideal caps in many Spice models or various graphical representations of the response curves generated by filter circuits, you might get the impression that the action of a particular value cap just goes on forever as you go up the frequency scale. Apearently this is not the case - rather after some point the AC resistance begins to rise again. Linsely-Hood (The Art of Linear Electronics) shows a graphical example for some electrolytics where the rise begins below 20kHz. Electros, Al or tantulum are most limited in this respect. The AVX data are not as exact but make it clear that all the "poly" caps have a bandwidth much more limited than ceramic or silver-mica.
This may not be important if you have a metal enclosure that is earthed, thus providing RF shielding. Nonetheless, if you want maximum bandwidth as a precaution against RF in the rails, ceramic or silver-mica gives you a greater margin of safety. The standard arrangement is an electro parralelled with a ceramic to maximize the range affected from very low AC up through RF.
I would not use a Z5U for anything. The problems involve more than just microphonics.
XR7s.......ok for digital........not so sure that I would for use them for audio.
The effect that sam9 refers to is called "self-resosnant frequency".
All caps will begin to look inductive at some frequency. I would advise trying to find that data for any cap used in any bypass application.
Of course.......with an RF generator and a 'scope.....you can measure it yourself.
I would not use silver micas for bypass, either. Their primary purpose is in tuned circuits. The values large enough to be useful for bypass will be physically large.........and unless you buy them at a surplus store.......very expensive.
Jocko
XR7s.......ok for digital........not so sure that I would for use them for audio.
The effect that sam9 refers to is called "self-resosnant frequency".
All caps will begin to look inductive at some frequency. I would advise trying to find that data for any cap used in any bypass application.
Of course.......with an RF generator and a 'scope.....you can measure it yourself.
I would not use silver micas for bypass, either. Their primary purpose is in tuned circuits. The values large enough to be useful for bypass will be physically large.........and unless you buy them at a surplus store.......very expensive.
Jocko
I know a bit about self-resonance, loop inductance, and such from past digital design work. I was just curious about everything else - in Rumsfeld-speak, the known unknowns (e.g. microphonic effects) and the unknown unknowns - that would be new to someone getting into audio.
But I think you guys have answered my question in an indirect way.
But I think you guys have answered my question in an indirect way.
I've used the Xicon polystyrene caps (from Mouser) and they work well in filters- for tone controls, phono RIAA, DAC output filter etc. A good improvement over polyester. But because they are big parts, you really can't use them elsewhere. I don't recommend using these for power supply decoupling or high-frequency >5MHz stuff because they have a fair bit of self-inductance.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.