------------------------------ From: manfredi@rockwell.com (Bert Manfredi, 747-6735) Subject: Re: Help : Shunt Capacitors Date: Mon, 28 Mar 1994 15:44:00 GMT In article <2msl1e$i5l@introl.introl.com>, "Donald S. Chiu" (AED-EWD) writes... >Can someone explain to me what is "Shunt Capacitors" ? >In my preamp, it has a 20,20,20,20 uf electroyltic >capacitor and was told that I should put shunt >capacitors on the electroyltic section to improve >the sound. I need to know how to connect them to the >electroyltic and what type and values should I use? > >Thanks, >Don Chiu >dchiu@pica.army.mil As frequency goes up, electrolytic capacitors start behaving more like resistors than capacitors. For example, electrolytic capacitors in a power supply do not ground RF very effectively. One good solution to this problem is to connect a decent film capacitor in parallel with the electrolytic. Because the film cap is only being used to handle the higher frequencies, its capacitance is not very critical. Typically, 0.1 or 1 uf capacitors are good choices to bypass electrolytics. But make sure the voltage rating is at least as high as the rating of the electrolytic. Not usually a problem, since electrolytics are normally not rated at high voltages, but there are plenty of 50 V film caps out there which would not be adequate in most power supply applications. Bert ------------------------------ From: DPierce@world.std.com (Richard D Pierce) Subject: Re: Help : Shunt Capacitors Date: Mon, 28 Mar 1994 18:00:43 GMT In article <2n71mb$1qvh@introl.introl.com> manfredi@rockwell.com (Bert Manfredi, 747-6735) writes: >In article <2msl1e$i5l@introl.introl.com>, "Donald S. Chiu" (AED-EWD) writes... >>Can someone explain to me what is "Shunt Capacitors" ? >>In my preamp, it has a 20,20,20,20 uf electroyltic >>capacitor and was told that I should put shunt >>capacitors on the electroyltic section to improve >>the sound. I need to know how to connect them to the >>electroyltic and what type and values should I use? > >As frequency goes up, electrolytic capacitors start behaving more like >resistors than capacitors. For example, electrolytic capacitors in a >power supply do not ground RF very effectively. Actually, Bert, it can be worse than that. Depending upon the manufacturer and grade, some electrolytics show a significant inductive component, with the impedance rising above 10 kHz or so, for a variety of reasons. They act only as reasonably pure capacitances below 1 kHz or lower. One solution, now availbale since there are manufacturers of large-valued film capacitors for crossover use, is to simply dispense with the electrolytics altogether where the capacitance (100's of uF) is doable. What you might spend for a high-grade electrolytic, you'll spend 20x that on an equivalent film cap. I have contructed several simulated speaker loads for testing amplifiers under realistic conditions, and the difference in behaviour of the load using electrolytics and film caps is, uh, not subtle. Like under high powers, the film caps don't explode. -- | Dick Pierce | | Loudspeaker and Software Consulting | | 17 Sartelle Street Pepperell, MA 01463 | | (508) 433-9183 (Voice and FAX) | ------------------------------ From: torgd@tolstoy.lerc.nasa.gov (Russell DeAnna) Subject: Re: Help : Shunt Capacitors Date: 28 Mar 1994 20:17:07 GMT In article <2n71mb$1qvh@introl.introl.com> manfredi@rockwell.com (Bert Manfredi, 747-6735) writes: >As frequency goes up, electrolytic capacitors start behaving more like >resistors than capacitors. For example, electrolytic capacitors in a >power supply do not ground RF very effectively. One good solution to >this problem is to connect a decent film capacitor in parallel with the >electrolytic. Because the film cap is only being used to handle the >higher frequencies, its capacitance is not very critical. ----------------------------------------------------------------------- Please help me out here. If two capacitors are connected in parallel, their effective capacitance would be the sum of the individual values: Ceff=C1+C2. The effective impeadance would be proportional to 1 ------- . (C1+C2)f So adding a small capacitor in parallel to a much larger one would not affect the effective capacitance or impeadance. I don't see how frequency has any effect. What does the "film cap is only being used to handle the higher frequencies", mean? I'm assuming the capacitance of an electrolytic changes with frequency. Is that the part I'm missing? -- Russell DeAnna NASA Lewis Research Center, Cleveland, Ohio torgd@tolstoy.lerc.nasa.gov ------------------------------ From: Craig Stark Subject: Re: Help : Shunt Capacitors Date: Mon, 28 Mar 1994 23:51:53 -0500 torgd@tolstoy.lerc.nasa.gov (Russell DeAnna) writes: > Please help me out here. If two capacitors are connected in parallel, > their effective capacitance would be the sum of the individual values: > > Ceff=C1+C2. > > The effective impeadance would be proportional to > 1 > ------- . > (C1+C2)f > > So adding a small capacitor in parallel to a much larger one would not > affect the effective capacitance or impeadance. Unfortunately, you're assuming your capacitor is an ideal one. Were electrolytics ideal capacitors, we'd all be quite happy. However, if you take a look at the impedance curve of an electrolytic (or any cap for that matter) you'll find it is quite non-linear, even in the audio band. (If you've got a Digi-Key catalog, the Panasonic low ESR section I believe shows their curves. Keep in mind, these are some of the best of the electrolytics.) Electrolytics are some of the prime offenders and can show substantial difficulties in the top end of the audio range. By putting a small film in parallel whose impedance remains low at the higher frequencies, you aleviate the problem by lowering the effective impedance. Craig Stark ------------------------------ From: torgd@tolstoy.lerc.nasa.gov (Russell DeAnna) Subject: Re: Help : Shunt Capacitors Date: 29 Mar 1994 20:31:34 GMT Craig Stark writes: >By putting a small film in parallel whose impedance >remains low at the higher frequencies, you aleviate the problem by >lowering the effective impedance. Oh! So that's the part I was missing. When the impedance of the electrolytic rises with frequency, the film cap in parallel still offers a lower impedance. Since impedance in parallel adds like X1*X2 ------ X1+X2 The small film-cap impedance dominates the much higher electrolytic. Thanks, now it's all clear. I'll get out the soldering iron. -- Russell DeAnna NASA Lewis Research Center, Cleveland, Ohio torgd@tolstoy.lerc.nasa.gov ------------------------------ From: manfredi@rockwell.com (Bert Manfredi, 747-6735) Subject: Re: Help : Shunt Capacitors Date: Wed, 30 Mar 1994 19:51:00 GMT In article <2n7e5b$1ndg@introl.introl.com>, torgd@tolstoy.lerc.nasa.gov (Russell DeAnna) writes... >Please help me out here. If two capacitors are connected in parallel, >their effective capacitance would be the sum of the individual values: > > Ceff=C1+C2. Right. When you parallel your big electrolytic with a smaller film capacitor, the overall capacitance and therefore the theoretical impedance will not change much at all. >So adding a small capacitor in parallel to a much larger one would not >affect the effective capacitance or impeadance. > >I don't see how frequency has any effect. What does the "film cap is only >being used to handle the higher frequencies", mean? > >I'm assuming the capacitance of an electrolytic changes with frequency. >Is that the part I'm missing? Right. If the electrolytic capacitor is rated at 10000 uf, a typical power supply filter for example, it's capacitance at high frequencies is nowhere close to that. Dick Pierce talked about DC blocking caps in the audio path too, but let's stick to power supply caps for now (because the effect is more obvious). In a power supply, ideally you'd expect the 10000 uf capacitor(s) to ground just about anything that is not DC voltage. Of course, any 50/60 Hz or 100/120 Hz ripple must be attenuated, but also any RF grunge. You don't want RF mucking about in the amp chassis. You will instead find that if RF is present in the power supply, the 10000 uf electrolytic capacitors will happily radiate the RF. I don't know what capacitance they actually represent at several hundred KHz, but it's not 10000 uf! As Dick Pierce pointed out, this basic effect is also visible at frequencies as low as a couple of KHz. So, in order to shunt away these high frequencies, assuming you need to use electrolytic capacitors, you can install a film capacitor, typically of much lower value, in parallel. Now you'll have the large capacitance for good low frequency filtering, for example, and the film capacitor will take care of the higher frequency filtering. Electrolytic caps are most often used when "a large capacitance" is needed. DC blocking and power supply filtering are obvious examples. In both of these applications, one wants to ensure that high frequencies sail through unaffected. The small film capacitor does the job admirable at the higher frequencies. Bert