Opinions on small value film caps for non-audio path HF bypass

  • Thread starter Thread starter sweetbeats
  • Start date Start date
sweetbeats

sweetbeats

Reel deep thoughts...
I'm looking to add some 0.01uF film caps in parallel with the electrolytic filter caps for local power rail filtration in my mixer.

To clarify: my Soundtracs MX mixer has 330uF polar 'lytics to filter the +/-17VDC rails where the power comes into each channel card as well as where the power comes into the master section. I'm planning on bypassing each of those with a 0.01uF film cap.

I'm wondering how critical it is the cap I select...would something like this do, or is there some reason why it would be important to spend more money?
 
rattle-rattle...ping-ping-ping...

*THUMP*


Anybody?
 
I'll bite, but only because I am too stupid to be afraid of putting my foot in my mouth. From my college days (a looooooooong time ago), what you are describing was refered to as a 'hash filter'. All the kids used to smirk every time the professor said that.
Anyway, it was used mostly in RF circuits with high value filter capacitors, and a .1uF paper capacitor was common. Why?
The filter is constructed in such a way that it is wound like a cigarette. It acts as an inductor to really high frequencies. This depends on the value of the capacitor, and the frequency. 330uF doesn't sound all that high, and AF may not matter. But what you can do is 'scope out the power supply, and then add the cap. Does the ripple improve? Feed a 20KHz or even higher AF through. Now 'scope the rails again. If you can't 'see' an improvement, there likely isn't one.
 
You probably don't want inductance there! A plain jane polypropylene cap will work fine- I avoid "audiophile" caps and stick with mass-manufactured units. Wima is an excellent choice (there are many more). Since ripple currents at that point are low, an MKP will work well and is relatively compact. Expensive caps are for suckers. Or timing circuits, which this ain't. Polypropylene will give you better characteristics than polyester.

The power supply impedance limitation in 90% of the implementations I've seen is layout, not the caps.
 
No, no, what I meant was the capacitor can act like an inductor at high frequencies. All those capacitors are wound with alternating layers of conductors and insulators. As they're rolled up like a cigarette, it is like an inductor; a wound piece of wire. But the capacitance has to be high, like in the thousands of microfarads, and the frequency has to be high, like in the MHz. That's why RF circuits used ceramic discs; the disc is of course not wound, and any effective inductance is negligible at the much higher frequency.
Still, if you're that keen on changing capacitors, or adding capacitors, hopefully you also have some test equipment. Then you can 'scope the rails and 'see' what you're doing, and if it is effective at all.
 
Usually (meaning "every one I've ever seen in 45 years of electronics construction") the rolled caps have metallization on the ends, generally thermally sprayed zinc, which eliminates that source of inductance.
 
All true. But why should a tiny bit of inductance even matter in a power supply? You're only dealing with 50 or 60 Hz after all. But if you're that worried about it just stick any old 0.1uF caps in, it's not critical
 
It's not just 60Hz or 120Hz. The cap charging waveform is rich in harmonics (it's not a sine wave!). And there's the question of supply impedance as well- at high frequencies, most gain blocks' power supply rejection falls off significantly. If you don't want your amplifier to double as an oscillator, you need to keep power supply impedance low into the multi-megahertz range.
 
Yeah, and any old capacitor, esp ceramic caps, will be very low impedance up to very high freqs
 
Not any old cap. RF ceramics, sure. Polypropylene, polystyrene, mica, ditto. Electrolytics, tantalum, polyester, no, they have poor HF characteristics.
 
Yeah...the whole point of the bypass is to shunt UHF garbage so HF inductance IS a concern.

I went with XR7 ceramics...0.022uF to compensate for the aging characteristics of that type.
 
X7R are good for that. NP0 (C0G) are even better, but can be kinda bulky in high values.
 
You must log in or register to reply here.
Back
Top