Hi Garey, I believe I can answer your questions adequately.
For a garden variety NPN pass reg (317, 340, 7800, etc) a long input lead from the 'lytic would be >6", in which case you should stick a 0.1uF or greater on the input. Modern linear regs are less tolerant of lead length and may require a ceramic input bypass regardless because they have enough bandwidth to get themselves into trouble with oscillations even with modest input inductance from the lead. For switching regs you should bypass right at the pins with a substantial cap that has low ESR. Read the datasheet carefully regarding input and output caps (and inductors) for switching regs as it's real easy to get into trouble with those varmints. Regarding tantalums vs ceramics and 'lytics: A tantalum is usually chosen for small size at a given capacitance and better high freq performance. Occasionally it is chosen for effective freezing temp operation (auto makers love 'em). I would not trust one all the way to 30MHz, however. Even a 1uF is probably inductive up there, but perhaps not reactive enough to be an issue in most circuits. In general, you can increase capacitance on the output of NPN regs without upsetting the IC. Keep in mind that both 'lytics and solid tantalums have significant ESR that prevent the very low output Z of the reg from seeing a pure reactance at the higher freq's where oscillation tends to occur, hence these chips can handle just about any "vitamin C" you throw at them. The real question becomes at what point is it counterproductive for either cost of low self-resonance reasons. 1uF seems reasonable, but I would have preferred to see a 0.01 to 0.05uF ceramic in parallel in the Drake circuit. BTW, early solid tantalums were not as reliable as the makers thought they would be. Quality test equipment circa early 70's from makers like HP and Tek regularly failed from shorted tantalums later in life. Modern tantalums are much better and have built in fuses so when the fail they will fail open vs shorted like the early generation. Sure, I know about the 723. My company built the improved version (mainly lower noise). What question(s) do you have? Dennis AE6C On Sun, Apr 25, 2010 at 2:40 PM, Garey Barrell <[email protected]> wrote: > Dennis - > > I KNEW we'd get someone who actually knew what they were talking about! > :-) > > I hadn't ever heard about the lead length in the input side. How "long" is > long? Does putting chip caps ON the regulator leads really buy you > anything? Certainly wouldn't hurt!?! > > I understood the role of the output cap, I thought, and believed that it > shouldn't be TOO big, not so much because of inductance but to enable the > regulator to "regulate" the output voltage for varying loads. "Fighting" > the cap is a good description! > > The Low Drop Out regs came along after I was out of engineering, so that is > good information. > > The Fairchild data sheet I looked at for the LM78XX series showed a 0.33 uF > on the output. Drake seems to have used a 1 uF tantalum (solid) on ALL the > "ins" and "outs". The tantalum cap should be ok for 3-30 MHz, but is that > TOO much capacitance. My experience says that a 0.33 disc ceramic would be > less expensive than the tantalums, so I would assume that Drake thought it > would be better. > > Are you familiar with the 723 at all? > > > 73, Garey - K4OAH > Glen Allen, VA > > Drake 2-B, 4-B, C-Line& TR-4/C Service Supplement CDs > <www.k4oah.com> > > > Dennis Monticelli wrote: > >> My company makes IC voltage regulators. Let me pass along some >> information. >> >> The purpose of the "bypass" caps for a conventional NPN pass regulator >> (317, 340, 7800, etc) s as follows. The input cap is to prevent possible >> oscillations if the lead from the IC to the raw DC filter cap is on the long >> side. It is just insurance. The output cap is to improve the transient >> load regulation. IC regs use a feedback loop to steady the output, however, >> the bandwidth of the loop is finite. The output cap takes over for the loop >> beyond a certain frequency so as to maintain good transient regulation. >> >> A PNP pass regulator (a.k.a. a Low Drop Output regulator) uses the input >> cap for the same purpose as the NPN types. But it's output cap is not only >> there for transient load regulation, it is usually mandatory for obtaining >> loop stability. Because of this special role, you must pay careful >> attention to it's value and dielectric type as recommended in the datasheet. >> >> You generally do not bypass a regulator output to decrease internally >> generated noise because the output Z is so low that it will fight the cap. >> Sometimes an IC reg will give you a pin to bypass the internal voltage >> reference portion of the chip. But this is not common and optional to do >> anyway. >> >> The presence of the input and output caps can help in keeping RF out of >> the chip's internal circuitry. If you're worried about that use a small >> value ceramic. >> >> Dennis AE6C >> >> > _______________________________________________ > Drakelist mailing list > [email protected] > http://mailman.zerobeat.net/mailman/listinfo/drakelist >
_______________________________________________ Drakelist mailing list [email protected] http://mailman.zerobeat.net/mailman/listinfo/drakelist
