myers@hpfclm.HP.COM (Bob Myers) (07/29/88)
> I've got one comment on power-amps when rolling-your-own. > In the power supply, the bigger the caps you use the better > the ripple filtering (60Hz stuff) BUT ALSO the bigger the > current surge through your rectifying diodes. > > As the ripple gets smaller the 'conducting-time' (forward bias) > of the rectifying diodes gets smaller. If the power to the > load stays the same then the current during foward bias gets bigger. > (lots bigger, the average current [over time] must be constant > for a given load.) > > So, for mongo(big) use mongo rectifiers. > > just a thought. :) > > > dsm > While this will work, there are better ways to handle the inrush current than overly-overrating your diodes (although a healthy bit of margin in the current-handling is always a good idea!). The biggest problem is that, while the diodes may now handle the current, you're still gonna need a MUCH larger fuse than your power supply really should have - and you ARE going to put a fuse on it, aren't you? Two common means of taking care of this problem: first and simplest, place an approriately-selected thermistor in the path of the inrush current. The initial current is limited, but then the resistance drops off so that you can get enough current through. It works, but there's a better - and more complex - method. The second method is to simply place a power resistor in the path of the current, again to keep the inrush current down to some reasonable value. Add a relay on a time-delay (switch the relay via a transistor with some suitable R-C delay on it) so that the relay will short the inrush limiting resistor after the main caps have charged up. Bob M. hplabs!hpfcla!myers
max@eros.uucp (Max Hauser) (07/29/88)
In article <61833@sun.uucp> klein@sun.UUCP (Mike Klein) writes: | | More important than the ampere-hour rating will be, as for any power source | you are using, the output impedance. ... Well, Mike, you're right, of course; but separating the supply-impedance issue from the ripple issue is no small accomplishment. And there are of course the incidental but possibly vital benefits of not having all of those 60*N Hertz AC magnetic fields around. Besides, one can always lower the *DC* output impedance of a supply as well, very satisfactorily, with a little series regulator. Common-emitter, of course (both for low voltage drop and, more important, to place the dominant pole at the output, the key to success). Max Hauser / max@eros.berkeley.edu / ...{!decvax}!ucbvax!eros!max "The ghost of Baron Rudolph von Guggenheim, the 16th century nobleman murdered by the Countess Rowena DuBois and her lover (believed to be the Duke of Norwood), falls into Edna's bean dip." -- The Far Side, 10/85
henry@utzoo.uucp (Henry Spencer) (08/03/88)
In article <1320005@hpfclm.HP.COM> myers@hpfclm.HP.COM (Bob Myers) writes: >> As the ripple gets smaller the 'conducting-time' (forward bias) >> of the rectifying diodes gets smaller. If the power to the >> load stays the same then the current during foward bias gets bigger. >While this will work, there are better ways to handle the inrush current >than overly-overrating your diodes... >Two common means of taking care of this problem: ...thermistor... power > resistor [shorted by relay] I think you've misunderstood; the problem is not the inrush current at powerup (although that is something to watch) but the current drawn on each half-cycle of the AC afterward. Current flows through the rectifiers only when the voltage behind them exceeds that on the filter capacitor (ignoring their forward voltage requirement). With big capacitors that don't drop much in voltage between half-cycles, flow can occur only at the very peak of the wave, so all the current needed by the load throughout the half-cycle has to come through the rectifiers during a brief period at peak. Hence the need for bigger diodes, or else some more sophisticated regulator downstream so that the capacitors can be made smaller. -- MSDOS is not dead, it just | Henry Spencer at U of Toronto Zoology smells that way. | uunet!mnetor!utzoo!henry henry@zoo.toronto.edu
markt@hpiacla.HP.COM (Mark Thompson) (08/04/88)
After reading all these recommendations I am suprised that no one recommended any of the designs in The Audio Amateur. This construction oriented audiophile magazine has had lots of amplifier designs published in it with may different design philologies. The last issue had an article on how to build a 40w mosfet amp that compared very favorably to the Rodger Mogeski RM9 tube power amp. In the past they have had a 25w class A mosfet, 100w mosfet, 60w mos many biopolar designs and many mods to current amplifiers. Its definitely worth checking out. You can contact them by calling (603) 924-6526 or (603) 924-6371. I don't have the address here. My only tip is to resist what science would tell you about passive components and use polyproplyne and polystryene where ever possible, they are worth the money. Mark
myers@hpfclm.HP.COM (Bob Myers) (08/06/88)
My stuff: >>While this will work, there are better ways to handle the inrush current >>than overly-overrating your diodes... >>Two common means of taking care of this problem: ...thermistor... power >> resistor [shorted by relay] Henry's stuff: >I think you've misunderstood; the problem is not the inrush >current at powerup (although that is something to watch) but the >current drawn on each half-cycle of the AC afterward.... Hence >the need for bigger diodes, >-- >Henry Spencer at U of Toronto Zoology Yep - I blew it. As Henry correctly pointed out (as have several others who sent E-mail), I read the original posting too quickly, and went off on a tangent about *turn-on* inrush. Which IS, by the by, something to be concerned about, and I'll refer you back to my original mistaken response for further details. Forgive me - I've been working recently on some power-supply-blow-fuse-at-turn-on complaints, and my mindset was in the wrong direction (and due to the sluggish response of my mind, loathe to change direction once set). Having said that, I note that there have in the interim been some excellent responses on the problem which is actually at hand, but will go ahead and add my (course-corrected) $0.02 to them: One of the mailed comments I received (and I'm sorry I can't recall the author right now, because it was a very well-thought-out response) pointed out the actual problem being discussed in terms of the capacitor's charge/discharge action on alternate half-cycles. While this is certainly one correct way to look at it, I want to point out that for analysis purposes, it may be more helpful to consider the output of a power supply as being the superposition of two sources - a DC source (the desired DC output) plus an AC source (the ripple). Your mission is to design a circuit which minimizes the AC voltage across the (assumed) resistive load. This can be done through a combination of BMF (a technical term, the meaning of which can be found in various electronics dictionaries :-)) capacitors and inductors (the filter). The filter circuit is amenable to simple AC analysis. (The DC analysis is even simpler, of course, except for some possible problems shown in the following....). As a previous poster pointed out, choke (inductor)- input filters hold down the repetitive inrush to the filter caps. The problem is that they're big, heavy, and can be more difficult to locate than big caps. Fortunately, they're not too hard to BUILD if you get desperate. When performing this analysis, though, don't forget that there ain't no such thing as a "perfect" capacitor or inductor. One important characteristic (which should be readily available from the capacitor catalogs) is the ESR of your BMF electrolytics. And you ARE going to put a fuse on this sucker, aren't you? Bob M.
ornitz@kodak.UUCP (barry ornitz) (08/06/88)
In article <4110007@hpiacla.HP.COM> markt@hpiacla.HP.COM (Mark Thompson) writes: >My only tip is to resist what science would tell you about passive components >and use polyproplyne and polystryene where ever possible, they are worth the >money. What goes on here? We "scientists" fully understand the results of nonlinear dielectrics, dissipation factor, and dielectric absorption. Speaking as one who has built custom sample cells and assembled apparatus to measure dielectric properties of polymers over extended frequency and temperature ranges (1 to 200 MHz, 0 to 200 degrees C), I wonder if your posting should have possibly been in wreck.audio. We make polyethylene, polypropylene [note the correct spelling], and numerous polyesters, by the way. If you want to spend your money wisely, use the right components at the appropriate places in your circuit. However, I would love to sell you 100,000 uF polystyrene capacitors for your power supplies.;-) [With their inherent high internal inductance you would still have problems.] Imagine the use of all low-loss film capacitors in boom boxes too! Barry ----------------- | ___ ________ | | | / / | | Dr. Barry L. Ornitz UUCP:...!rochester!kodak!ornitz | | / / | | Eastman Kodak Company | |< < K O D A K| | Eastman Chemicals Division Research Laboratories | | \ \ | | P. O. Box 1972 | |__\ \________| | Kingsport, TN 37662 615/229-4904 | | -----------------