On Mon, Apr 11, 2011 at 8:20 AM, Olin Lathrop wr= ote: > Did you read what I wrote last week? I went into this in some detail. > The following is the only reply that I had received: You are going a bit off the deep end. You need to go back to the basics an= d > see what this is all about rather than relying on rules of thumb with lon= g > forgotten derivations. > > This issue is that the efficiency (final resulting cooling power per > electrical input power) of Peltier coolers goes down the harder they are > driven, although the total cooling power still goes up to some maximum > point > specified in the datasheet. Let's say you want 50% of the total cooling > power the device is capable of. You can get that cooling power by a 50% > duty cycle of full power or a steady input resulting in 1/2 output power. > Both cause the same cooling, but the first uses more electrical power > because the device is more efficient at the lower power operating point. > > Taken to the extreme, any deviation from running the cooler with a steady > DC > current causes inefficiency. You can contort this to say that "ripple" i= s > bad in that it causes inefficiency, but small deviations from the average > operating point aren't going to significantly matter. Any regulated powe= r > supply is going to be good enough. > > If you are doing active temperature control, what you don't want to do is > the otherwise obvious PWM control of a Peltier cooler (like would be > appropriate for a resistive heater). You essentially want to smoothly > control a reasonably filtered analog current. However, that presents som= e > problems in itself. If you are going to derive this current by linearly > regulating a fixed voltage power supply, then you're just wasting the ext= ra > power elsewhere. Maybe that helps because it's physically in a different > place where it may not matter as much getting hot or it's easier to get r= id > of the heat. However, if power efficiency is the goal that's not going t= o > help. > > This is why Peltier coolers are usually driven by what amounts to a > variable > switching power supply, but with enough filtering to keep the total rippl= e > to a small fraction of the average output. Anything that would qualify a= s > a > regulated power supply for other purposes would likely be good enough. > Yes, I did read it, but all I got from it is that I should use a smooth DC source. I don't care about efficiency, since it doesn't consume much power anyway, and it will get power from a computer power supply (12V rail or wahtever). And for starting purposes, since the goal is to get the drink as cool as possible as quickly as possible, a direct, uncontrolled 12V line is perfectly fine, with an on/off switch. But if I'm going to be controlling the cooling power, I still did not understand exactly which way I should go about it. Linear regulators that can supply 6A are hard to come by, but I can get them. What other ways are there, short of designing your own variable voltage high-power power supply? --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .