On Saturday 16 June 2007 19:43, Randall Nortman wrote: > It seems that the lower the average input > voltage, the higher the average efficiency of the regulator is going > to be, so I would ideally aim to have the largest input ripple > possible,
You have to consider things like loop filter lag etc. A more stable input will give a more stable output. > which coincidentally allows me to choose a smaller, cheaper > input capacitor. Do not over look the "ripple current" rating of the capacitor, when used in a switching regulator. Each topology of regulator places different demands on the input and output capacitors. Using a capacitor that does not have a good ripple current rating (we are not talking about your 120 Hz here, but the 500 kHz that is reflected back to the input) generally lets the magic smoke out of them. Once the magic smoke escapes you have to get a new one. Also the type of output capacitor matters. If the ESR is to high or to low it can cause stability problems. > But it can't be that easy. Smaller and cheaper is never better, so > what am I missing? At 500 kHz you are designing a radio, and RF Black Magick starts to apply. Follow the recommended layout in the data sheet as closely as you can, and if you can use the parts they use in there example that is even better. Unless you need small size, a lower frequency can be more efficient, due to lower switching losses. The major reason to go to the higher frequencies is to allow smaller components to be used. Right now Micrel holds the record at 8 MHz for one of their regulators. From the National Semi. Power Design Seminar http://www.national.com/appinfo/power/ a couple of months ago, I hope the post the manual on line, it has some of the best layout times I've come across, but in the mean time, look up these application notes: "National Semiconductor application notes: – Layout Guidelines for Switching Power Supplies, AN1149 – SIMPLE SWITCHER® PCB Layout Guidelines, AN1229 Layout Considerations Summary Here is a checklist of the layout considerations we discussed for noise reduction. Realize that PCB traces have impedance (both resistive and reactive) and current takes the least path of impedance." Nothing wrong with the TPS5420 choice based on the little information you have given, if you want to go with external FETs take a look at the LM5116. http://www.national.com/pf/LM/LM5116.html -- http://www.wearablesmartsensors.com/ http://www.softwaresafety.net/ http://www.designer-iii.com/ http://www.unusualresearch.com/ _______________________________________________ geda-user mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
