I think that it is likely that the existing power supply can be optimized to achieve the desired low ripple. First of all, it should be verified that the input ripple never takes the regulator input voltage close to the dropout level. Secondly, the ability of the regulator to reject ~100Hz input signals must be checked to see how much we should expect it to reduce the ripple. It may be warranted to perform the regulation in two stages. I would also look at possibly using an inductor plus capacitors on the input and increasing the filtering capacitance. If you increase the capacitance without adding inductance, then the current drawn from the input supply will become ultra-short, very high current pulses only right around the peaks of the AC waveform, which is bad for thermal reasons and possibly for hum production, too, since it may introduce high-frequency components to the ripple which are not well rejected by the regulator. The inductors not only perform additional filtering but they also smooth out the diode bridge conduction. I very much doubt that an SMPS will do any better at hum rejection than a linear regulator. On Fri, Mar 6, 2015 at 12:38 AM, Ryan O'Connor wrote: > Theoretically this would reduce ripple after a SMPS too, correct? > > Ryan > > On 6 March 2015 at 17:34, Bob Blick wrote: > > > It wouldn't work the same. As the load on the circuit increases, the > > base current also increases, so the voltage drops both due to the > > overall voltage drop, and also due to the increases base drive. > > Hopefully keeping the transistor from dropping out (and letting ripple > > through). > > > > If you used a MOSFET, you could get gobs of ripple reduction (the > > resistor can be much larger), more than with a BJT. But as you increase > > the load, the output voltage only tracks the voltage drop, not the > > increased load current. > > > > A 78xx regulator would need the ground lead to be less than the output, > > unlike a BJT or MOSFET. So you'd need to use a voltage divider (off the > > input, to ground). It would track the input voltage, but due to the > > voltage divider, to a lesser extent than a BJT or MOSFET. > > > > Friendly regards, Bob > > > > On Thu, Mar 5, 2015, at 07:25 PM, Ryan O'Connor wrote: > > > Does it have to be a BJT? What about a mosfet or a 78xx regulator wit= h > a > > > cap on its adjust pin? > > > > > > Ryan > > > On 6/03/2015 4:17 PM, "Bob Blick" wrote: > > > > > > > Hi Ryan, > > > > You are right! It does not increase the storage of the filter > > capacitor. > > > > If you have an unregulated power supply with a filter capacitor (an > > > > old-school DC wall-wart, for example) there will be ripple under > load. > > > > The greater the load, the greater the ripple. What the described > > circuit > > > > can do (with appropriate transistor/resistor/capacitor choices) is > > > > maintain the output voltage a little under the lowest peak of the > > > > ripple. The ripple on the output is less than, but follows, a > multiple > > > > of the circuit's capacitor. The nice thing about it is that the > output > > > > voltage drops with load, so it almost tracks the lower ripple peak, > > > > something a 78xx regulator can't do. It also has less insertion los= s. > > > > Cheerful regards, Bob > > > > > > > > On Thu, Mar 5, 2015, at 06:50 PM, Ryan O'Connor wrote: > > > > > That page is really informative. However, I'm left wondering abou= t > > the > > > > > physics. How can a capacitor which stores charge possibly be > > "multiplied" > > > > > without any additional storage? > > > > > > > > > > Ryan > > > > > > > > > > On 6 March 2015 at 02:46, Adam Field wrote: > > > > > > > > > > > On Wed, Mar 4, 2015 at 12:17 PM, Bob Blick > > wrote: > > > > > > > Hi Joe, > > > > > > > You could use a one transistor pass element to multiply a > > capacitor. > > > > Let > > > > > > > me see if I can find an example on the web... nope. here's > napkin > > > > art. > > > > > > > > > > > > > > Bob > > > > > > > > > > > > ESP has a good page on the concept: > > > > > > > > > > > > http://sound.westhost.com/project15.htm > > > > -- > > http://www.fastmail.com - Access your email from home and the web > > > > -- > > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive > > View/change your membership options at > > http://mailman.mit.edu/mailman/listinfo/piclist > > > -- > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .