Paul, Thanks for the input. You've got me thinking again. I appreciate the response, as no one else responded. The Application is for a 1000 watt digital RF amplifier. I have a fixed switching power supply that puts out 50 volts @ 25 amps and 12 volts @ 8 amps. 12 volts will be going to a couple of thermoelectric coolers ( attached to output transformer ). The 7 to 10 volts adjustable is for the gate drive on the power mosfets (class-D). The 5 volts fixed is for the logic. A long ways away from a scanner. But, very interesting to hear what they have done, as well as its failure. Sounds like you've got it fixed-up better than their original design. The reason I wanted efficiency, is I don't want to use a heat sink if possible. The circuit board is all surface mount and small. Also, trying to keep the height profile low. Been doing a lot of data sheet looking since my post, and I am looking at the Linear Technology LT1086. This regulator is like the LM317, only is much more efficient by somehow taking a lot of the wasted power and dumping it back into the load....so they say :) I think this is going off topic....the PIC monitors the heat sink temperature as well as the amp input frequency and can shut down the amp. also, it provides for a pulse mode operation. Thanks, David - WA6LBX ----- Original Message ----- From: Paul B. Webster VK2BZC To: Sent: Sunday, July 04, 1999 7:03 AM Subject: Re: PIC Power Supply David Covick wrote: > 12 volts DC in - fixed (not a battery) > 7 to 10 volts DC variable output @ approx. 1 amp > 5 volts fixed @ 200 mA > Was thinking about LDO types of regulators, and would like the > regulators to be efficient enough not require a heat sink. A couple of misunderstandings here I think. You need a LDO regulator when the input-output differential is less than 2.5 volts. This *might* be a touchy point in deriving your 10V from 12V at 1 amp. It depends on whether you really want the 10V at 1 amp and whether the 12V is exactly 12V or always more than 12V. The common LM317 will *just* deliver 10V at 1A from 12V as long as its temperature is *above* 250C (and it's extremely unlikely to be cooler than that, isn't it?). If it really is critical, you may need a LDO device. But then you talk of "efficiency". Insofar as this term applies to series regulators, most LDO types are *less* efficient as they draw more current through the reference terminal in order to drive the output device, particularly at (near-)maximum load. The term "efficient enough not require a heat sink" really doesn't make sense. Series regulators dissipate the difference between input and output voltage, times current as heat and that's all there is to it. A *large* device may to some extent act as its own heatsink but your 7V regulator is going to dissipate 5 watts, and your 5V fixed device, 1.4W. I'd say the first will need a big heatsink and the second, a little one. > Can anyone recommend their favorite part number(s)? LM317 for both. > Also, would it be better to run the 5 volt regulator off the variable > 7 to 10 volt regulator, Definitely not! Firstly, the variable regulator is already dissipating the lion's share of power, so why add to that? Secondly, you are deliberately subjecting the fixed regulator to a variable input voltage - why would you want to do that? Thirdly, you are using the 5V regulator close to *its* dropout voltage if supplied from 7V. Finally, if the variable regulator shuts down, it will take the 5V supply with it (this may not be a bad thing of course, depending). > or from the 12 volts? Definitely. This said, I recently looked into my desktop document scanner and found it used a common LDO regulator to derive 10V or so from the "12V" supply, and both 5V regulators ran from that. The stepper ran from the 10V. Are you building a scanner perchance? Of course, you may ask why I examined the power arrangements in the scanner? Well, it overheated of course! Cooked the electrolytics and (there's more to the story). It shouldn't do that twice. I cut holes in the casing to clear *and* ventilate the new electrolytics which were bigger than the originals. -- Cheers, Paul B.