jjr@ut-ngp.UUCP (Jeff Rodriguez) (07/27/87)
I'm building a general-purpose power supply based on a 12 V/450 mA transformer. The secondary feeds into a full-wave bridge rectifier, which feeds into an LM317T adjustable voltage regulator (1.2 V to 27 V output at 1.5 A). How do I choose the size of capacitor to place across the output of the rectifier? I once heard that the rule-of-thumb is 1000 micro-F per 1 A of current. Is that reasonable?
larry@kitty.UUCP (Larry Lippman) (07/27/87)
In article <5705@ut-ngp.UUCP>, jjr@ut-ngp.UUCP (Jeff Rodriguez) writes: > > I'm building a general-purpose power supply > ... > How do I choose the size of capacitor to place across > the output of the rectifier? I once heard that the > rule-of-thumb is 1000 micro-F per 1 A of current. > Is that reasonable? The size of the filter capacitor is determined by two factors: (1) the maximum output current; and (2) the maximum permissible ripple voltage. The usual design equation is given below [@ 60 Hz]: C = (I / dV) * (8.33 x 10^-3) C = capacitance in farads I = maximum output current in amperes dv = maximum ripple voltage in volts Hence, there is no hard and fast rule which is solely determined by output current. Furthermore, the above equation is valid only for a 60 Hz power supply; a more generalized equation is used for other frequencies when dealing with DC/DC converters and switching power supplies. The real selection criterion is how much ripple will the application circuit tolerate. In the real world, that can cause C to vary from 100 to 10,000 uF per ampere. Personally, regardless of the equation or application, I would never use LESS than 1,000 uF for any power supply with an output voltage less than 50 volts. <> Larry Lippman @ Recognition Research Corp., Clarence, New York <> UUCP: {allegra|ames|boulder|decvax|rocksanne|watmath}!sunybcs!kitty!larry <> VOICE: 716/688-1231 {hplabs|ihnp4|mtune|seismo|utzoo}!/ <> FAX: 716/741-9635 {G1,G2,G3 modes} "Have you hugged your cat today?"
heppell@cory.Berkeley.EDU (Kevin G. Heppell) (07/27/87)
In article <5705@ut-ngp.UUCP> jjr@ngp.UUCP (Jeff Rodriguez) writes: > >I'm building a general-purpose power supply... > >How do I choose the size of capacitor to place across >the output of the rectifier? I once heard that the >rule-of-thumb is 1000 micro-F per 1 A of current. >Is that reasonable? In one of TI's applications manuals they list a graphical solution: O.H. Schade, "Analysis of Rectifier Operation", Proc. IRE., Vol. 31, 343, 1943 Basically, the desired filter (capacitor plus shunt resistance of the regulator) depends on several factors, including whether you use a full or half wave rectifier, line frequency (usu. a constant, but who knows :-), peak input (transformer voltage), and desired output voltage. I'm sorry, but I haven't worked with the method enough to give you a specific answer. The TI manual "Linear and Interface Circuits Applications", Volume 1, 1985 gives a good, _general_ discussion of basic linear regulators, their own parts notwithstanding. Disclaimer: I have no connection with TI except someone gave me a bunch of their books for free. -------------------------------------------------------------------------------- Kevin G. Heppell USNail: 784 Santa Barbara Rd. USENET: ucbvax!cory!heppell Berkeley, CA 94707-2046 arpa: heppell@cory.Berkeley.EDU (415) 528-6396
muller@sdcc7.ucsd.EDU (Keith Muller) (07/27/87)
Except for "brute force" power supplies, depending on the capactors to perform filtering only works on steady state conditions, which except for test bench measurements quite often do not exsist in a real application. The large capacitors are placed on the raw DC supply before the regulator circuit. Smaller ones are place on the other side. Although the capacitors do lower ripple, they have terrible response to dynamic load swings (which is the job of the active regulator).
jeffw@midas.TEK.COM (Jeff Winslow) (07/28/87)
I have a quibble here: In article <935@sdcc7.ucsd.EDU> muller@sdcc7.ucsd.EDU (Keith Muller) writes: > Although the capacitors >do lower ripple, they have terrible response to dynamic load swings... In the cases I have worked with, the capacitors actually have much better response to dynamic load swings than the rectifier-transformer-60Hz line behind them. The problem is that they can only store so much energy, so eventually they can no longer cover up the shortcomings of these other components. Jeff Winslow
bmaraldo@watdcsu.UUCP (07/29/87)
The usual modification to an audio amplifier power supply is replacing
the old caps with higher quality caps of the same capacitance or
higher. Also, the addition of a .47uF and 5uF polyproylene bypass cap
helps to clean up the sound.
I have added as much as .25F to a power amplifier and notices an
incredible improvement in the low end transient responce. These
modifications are of course audiophile related and my not make any
sence in the real world. As you know, the world of the audiophile
operates under a different physics :-)
I generally put as much capacitance on the dc rails as possible. Under
some operating circumstances, this may not be advised, but these
unusual circumstances are unknown to me. Also remeber to put a
drainage resistor across the caps that'll drain them when the power
shuts down. On large supplies it is best to have a relay close which
that places the resistor in the circuit. I have welded a 3/8 inch
drill bit to a quarter with my 40V .25F power supply.
Brett L Maraldo
--
-------- Unit 36 Research ---------
"Alien Technology Today"
-------------------------------------------
bmaraldo@watdcsumuller@sdcc7.ucsd.EDU (Keith Muller) (07/29/87)
In article <1666@midas.TEK.COM>, jeffw@midas.TEK.COM (Jeff Winslow) writes: > In the cases I have worked with, the capacitors actually have much better > response to dynamic load swings than the rectifier-transformer-60Hz line > behind them. The problem is that they can only store so much energy, so > eventually they can no longer cover up the shortcomings of these other > components. > > Jeff Winslow On good quality supplies, the RAW supply (before the regulator) does have large capacitors for what can very loosely be called an energy store. But the actual performance of the supply is solely the function of the closed loop feedback system that consists of the reference supply , difference amps and the regulator. For example a typical low voltage small wattage supply would have something like 1800 uf between the bridge and the regulator. The output cap is on the order of 51 uf. The whole system follows simple control system theory in terms of performance mesurements (which shows why the output cap has to be kept small). Take a look at a good supply circuit diagram and you can see (lambda, ac/dc, power designs, hp ...) what I mean. Keith Muller UCSD
gene@cooper.UUCP (Gene from EK Enterprises) (07/29/87)
in article <935@sdcc7.ucsd.EDU>, muller@sdcc7.ucsd.EDU (Keith Muller) says: > The large capacitors are placed on the raw DC supply before the regulator > circuit. Smaller ones are place on the other side. Although the capacitors > do lower ripple, they have terrible response to dynamic load swings (which > is the job of the active regulator). In general, when using a 3-terminal regulator like the 7[89].. series (7805, 7915, 7812, etc.), I *always* use about 5000 uF/A between the bridge rectifier and the regulator, and at least 10 uF/A at the output, and always in conjunction with a 0.1 uF ceramic bypass cap. In addition, *ALWAYS* put a blocking diode across the input and output terminals of the regulator; these buggers just hate it when the output voltage exceeds the input voltage (for example, on powerdown). --------|<--------- ______ | __________ | ----| ||| |--------+ +---*---+ 7xxx +----*----O + | ||| | |~ +| | |________| | | ||| | | | === | === | ||| | |~ -| | | | ----| ||| |--------+____+---*--------*--------*----O - diode xfmr bridge cap reg cap Gene ...!ihnp4!philabs!phri!cooper!gene "If you think I'll sit around as the world goes by, You're thinkin' like a fool 'cause it's case of do or die. Out there is a fortune waitin' to be had. You think I'll let it go? You're mad! You got another thing comin'!" - Robert John Aurthur Halford
jeffw@midas.TEK.COM (Jeff Winslow) (07/31/87)
In article <936@sdcc7.ucsd.EDU> muller@sdcc7.ucsd.EDU (Keith Muller) writes: >On good quality supplies, the RAW supply (before the regulator) does have >large capacitors for what can very loosely be called an energy store. Nothing loose about it. If no energy gets stored there, the regulator doesn't have a chance. >But the actual performance of the supply is solely the function of the closed >loop feedback system that consists of the reference supply , difference >amps and the regulator. As long as the regulator has enough input voltage to stay operating. All the control system theory in the world won't help a saturated regulator. This is where the capacitor vs. transformer response that I mentioned in my earlier article becomes an issue. We aren't really disagreeing, are we? >For example a typical low voltage small wattage >supply would have something like 1800 uf between the bridge and the >regulator. The output cap is on the order of 51 uf. The whole system >follows simple control system theory in terms of performance >mesurements (which shows why the output cap has to be kept small). There are many ways of designing regulators. What you want to do, usually, is minimize the output impedance, over as wide a frequency range as possible. Also, of course, the regulator loop has to be stable. There's no intrinsic reason that a small output capacitor is necessary to satisfy these requirements - you could have a small cap with a fast regulator (which is usually the case in linear regulators) or a large, low-ESR cap with a slow regulator (which is usually the case in switching regulators). Wasn't the question which started this discussion referring to the capacitors *behind* any regulator, that is, the energy storage elements I referred to above? >Take a look at a good supply circuit diagram and you can see (lambda, ac/dc, >power designs, hp ...) what I mean. I'm not sure if I'm supposed to read between the lines here, but I *have* designed several power supplies, both switching and linear, for Tektronix. Once again, I would say these are quibbles for the fun of techincal discussion, rather than indicators of any basic disagreement. Jeff [but now I have to re-learn transmission lines] Winslow
wtm@neoucom.UUCP (Bill Mayhew) (08/03/87)
To respond to Jeff at Tektonix (see, I know how to spell it): I think the the original comment about capacitors having poor dynamic response stems from the fact that aluminum electrolytic capacitors have a large surface area and use a very long coiled up set of plates. The coiled up plates give them a a relatively high inductance, which in turn prevents them from unleashing the current that they've stored up terribly quickly. To get around the inductance of electrolytics, most power supplys place small (usually less than 1 uF) capacitors in parallel. Typically, ceramic caps are employed since their small flat disc shaped plates have characteristically low inductance. In spite of what audiophiles would tell you, ceramics are better for this application than fancy expensive capacitors. Mica capacitors would be good for this application, but they are expensive if you don't happen to have a load in the junk box. Note that long leads tend to counteract the effect of capacitors, thus so-called bypass capacitors should be used at the ends of long leads to counterbalance the leads' inductance. Tantalum capacitors are polarized capacitors that have a large capacitance versus size ratio, giving them low inductance. The disadvantage of tantalum capacitors is that they are expensive and tend to be available in only low voltage ratings. The most popular use of tantalum capacitors is in aerospace and avionics equipment where physical volume and weight (mass) must be reduced. Bill Mayhew, NEOUUCOM (wtm@neoucom.UUCP)
wtm@neoucom.UUCP (08/03/87)
In article <642@neoucom.UUCP>, wtm@neoucom.UUCP (Bill Mayhew) writes: > > To respond to Jeff at Tektonix (see, I know how to spell it): ^^^^^^^^ |||||||| I kno I'm gunna get male on this won. Unfortunately, I didn't stop the last article before I caught on that, perhaps, I could spell "Tektronix", but I sure couldn't type it! The blindfold is in place, fire when ready. I am glad that this is a medical school and not an english dept. Bill
jeffw@midas.TEK.COM (Jeff Winslow) (08/04/87)
In article <642@neoucom.UUCP> wtm@neoucom.UUCP (Bill Mayhew) writes: >To respond to Jeff at Tektonix... Fair enough. Maybe we should preface articles with the range of frequencies we intend to talk about - most aluminum electrolytics are inductors above a few hundred kHz (and some considerably lower). But you have my company name confused with some geophysical theory or other... :-) >Tantalum capacitors are polarized capacitors that have a large >capacitance versus size ratio, giving them low inductance. The >disadvantage of tantalum capacitors is that they are expensive and >tend to be available in only low voltage ratings. There's also a controversy about whether their limited surge current capability affects their reliability in bypassing applications. Jeff Winslow
reh@ccd700.UUCP (reh) (08/05/87)
In article <3691@watdcsu.waterloo.edu>, bmaraldo@watdcsu.UUCP writes: > I have added as much as .25F to a power amplifier and notices an > incredible improvement in the low end transient responce. These ... > I generally put as much capacitance on the dc rails as possible. Under > some operating circumstances, this may not be advised, but these > unusual circumstances are unknown to me. Also remeber to put a > drainage resistor across the caps that'll drain them when the power > Brett L Maraldo During power-up, and even during operation the power supply currents will be much higher and the rectifiers (and even the fuse) may not be able to handle it. But it will sound better while it works. The transformer may also generate more heat. ...................................................................... Bob Harold 313-845-5404 Ford Motor Co., DPTC room B-206 ...!ihnp4!mibte!ccd700!eed090!bob 17000 Rotunda Drive Disclaimer: The views expressed might Dearborn, MI 48121-6010 not be those of my employer or myself Have questions about life? Read the original design manual - the Bible. ......................................................................