For the skilled builder – John Channels
The following article is aimed at cooling hobbyists with specific knowledge and skills pertaining to electronics and electronic kit construction. Prospective builders of the Peltier PSU project should possess, or be willing to learn, basic skills such as soldering and reading an electrical schematic. Readers should also be aware of inherent dangers regarding working with electricity and soldering equipment.
Neither the author or Overclockers.com will be held responsible for any damages due to anyone attempting what is written herein.
In order to complete this project successfully, you should have a few things available to you:
- A Multimeter, whether it be a Digital or Analog, is very important for testing. It is hard to get by without one of these. They measure resistance (ohms), voltage and some can even measure amps. I use a Craftsman model which has additional features such as a capacitance meter, diode test, hertz meter, and a thermal probe input.
- Some method of measuring the temperature of components should be available. You must ensure that each device is operating within limits to prevent premature device failure.
- Soldering equipment. Most of the devices used here are not very sensitive to iron temperature, so precision equipment is not required. I used a 30 watt “el cheapo” from Radio Shack on my project. A damp paper towel or sponge should be used in order to clean the tip of the iron periodically. Fairly heavy gauge rosin (flux) core solder makes life easy.
- An oscilloscope provides the next level of troubleshooting ability to your arsenal. With our project, it is not necessary, but it will allow you to see how much voltage ripple you have on your PSU output lines. This is optional and in not required to build our PSU.
- A drill is needed if you plan to modify heat sinks or drill holes in the mounting case. A nibbler tool is also handy for working with cases. Using a rasp or file will make smoothing out any rough edges a breeze.
More recent advances in personal computers have led to increasingly more exotic and effective approaches to cooling them to a greater extent than factory specifications call for, in order to reach an end-user’s specific objective. With these creative methods, such as air cooling, water cooling, phase change and the method discussed here, peltier effect cooling, certain unavoidable caveats face the user.
These caveats must be approached logically, in order to assess the feasibility of a cooling project, and to determine our final choice as to which cooling method we will implement. This article intends to help a user overcome, or more correctly stated, minimize, certain problems we face when choosing peltier effect cooling, although there are aspects of peltier cooling which render it one of the less-likely candidates for cooling.
Weighing the options is obviously left to you, but a rather simple analysis of peltier cooling is presented here to hopefully aid your decision.
The order of importance of these considerations of peltier cooling will vary from person to person, so I present them here with no order of precedence, only as a general outline.
- Cost. Both initial purchase costs and operating costs must be considered. Operating costs of the peltier cooling method are considerably higher than any other method of cooling I know of, this stemming from the peltier (TEC) modules extremely low cooling efficiency. Power consumption of the device compared to its cooling ability is enough alone to steer many away from Peltiers.
- Difficulty. This is biased depending on skill level. Peltiers may to some seem a simple task, whereas phase-change would seem an almost impossible feat. Peltier cooling alone is not a complete method; another lower order cooling method must be used on top of the peltier method, such as water or air cooling. Consider the difficulty level based on both the peltier setup and the secondary cooling setup.
- Reliability. When developing a system, consider the cost of an unreliable system. When introducing more components and levels of complexity into an equation, there are many more points of failure that may not come into view until after the experience of some level of a catastrophic event. Consider the number of critical devices there are to fail in, say, air cooling. Compare this to the number devices that could fail in a peltier cooling setup.
- Timeframe. Time is an important consideration. Initial build time of a peltier system can be far greater than other means of cooling. When considering your individually suited cooling method, also consider the maintenance and upkeep that a particular system requires, and what outcome would be had if this maintenance would fail to be performed on time.
- Mobility. System mobility and size can be a considerable decision maker. If the system is likely to be moved, your possible cooling options may depend on this.
- Bragging Rights/Overall Performance. All realistic consideration aside, this alone may be the number one reason people use exotic, expensive, unwieldy cooling devices on their personal computers. Let your ego be your guide.
While building my peltier cooler, I had a primary goal in mind: Initial cost efficiency.
Considering the cost of my secondary cooling setup, a homebrew water cooler, the first stage TEC cooling must be cost effective in order to maintain a sane budget. The unavoidable cost: Copper cold plate, TEC module, affixing hardware, etc. The stickler: The power supply! I am NOT paying $130 to run this thing, period.
The alternative was clear.
With a little number crunching and a fairly solid understanding of linear power supplies, an effective solution was readily available at a fraction of the cost. While my personal solution was bottom of the barrel, I will present several options for the home user, depending on the level of power supply regulation desired. The process can be broken down into several easy to follow layers.
We must not rule out a commercial power supply. If you buy a commercial power supply, you will find a few benefits with this type of supply:
- Smaller: The mass-marketed switching PSU targeted at the peltier crowd is small enough to fit into a drive bay. This is a very attractive option to many people. Your homebuilt PSU will not even come close to fitting in a single drive bay, although the creative may find a good way to mount it inside their case. This does not apply to commercial linear supplies, as their size will be comparable to that of the homebuilt version.
- Better Line Regulation: I hope that if you were to purchase a commercial supply, whether it be switching or linear, that it has near perfect output voltage. Most commercial supplies are going to run in the range of $80-$200, depending on the model and specific features. Since Peltiers are not very sensitive to voltage quality, we can save ourselves money by not making such a robust filter section in our supplies.
- Efficiency: If you buy a switching power supply, its efficiency will be far greater than the homebuilt PSU. If you buy a linear PSU, The efficiency will be comparable. You could also group heat output in this category. A linear PSU wastes energy by converting some of it into heat, while a switching supply is more efficient and produces lower heat levels.
Now, some benefits we might see with our homemade Peltier PSU:
- Affordable: In most all cases, you can build your power supply for a fraction of the cost of a commercial model.
- Customizable: You can build it to suit your own needs, in every aspect. Whether you need two voltage rails, or a custom built case for the PSU, it can all be accounted for while building your own.
vBuilding it yourself will give you a nice feeling of satisfaction when you are through. This can enhance the bragging rights of your cooling system.
Let’s take a look at how a mains (The power coming into your house, appt, dorm, etc) powered linear supply works. Study the drawing below. The following paragraphs break down the fundamental sections of a simple linear unregulated power supply.
In the first step, we take the mains voltage, ~120 Volts AC (Grey), and convert it into anywhere from 12 to 30 Volts AC (Green), depending on application. This is done by means of a Step Down transformer, represented by the two thin vertical lines and squiggles of wire on either side. This will in most cases be your largest component.
A transformer is made by wrapping two coils of wire around a common iron core. The first coil of wire that accepts the input voltage (120V) is called the primary. The output side is called the secondary. Transformers can have more than one primary and also have multiple secondaries.
A transformer works using principles of induction. The AC input voltage creates a constantly changing magnetic field into the iron core, which then transfers its energy into the secondary, producing a voltage.
Transformer manufacturers vary the number of turns (wraps) in a coil of wire in the primary and secondary coils to achieve the proper input and output voltage specifications. The size of wire they use to wrap around the iron core determines how many amps that a particular transformer can supply.
Transformers also supply us with an isolated source of power, separate from the mains supply. This is crucial in making a power supply safe, so that the full 120V AC signal is not present on the output of our power supply. Without this isolation, a power supply can be very dangerous because of the shock hazard that the mains voltage presents.
In the second step, we see something called a diode bridge (blue). This device uses discrete components (diodes) to give the electricity a limited path to follow. This in turn produces a crude DC signal, represented by DC Ripple, and the graph to the side of the drawing. The peltier module uses DC, but we want a little better “quality” DC before we can hook it up to the TEC.
Taking a closer look at the diode bridge section, you’ll see that there are arrows on each of the four diodes, pointing towards one end or the other. There is also a line at the tip of each arrow. The arrows point in the direction that voltage is allowed to flow through it.
A diode is simply an electrical check-valve; it only allows electricity to flow in one direction. The line at the tip of the arrow just reminds you that if you try to go the wrong way, you’ll be stopped by the diode. The line is like a brick wall to electricity.
The reason we need diodes is fairly simple: The AC output from our transformer is constantly changing directions, which means there is no positive (+) or negative (-) wires. Each wire switches back and fourth between positive and negative up to 50 or 60 time each second! Obviously, devices like peltiers and DC fans need a + and a – wire to work correctly. The diode bridge controls where the electricity can flow, and only allows the electricity to flow one way, hence making a DC output.
Step 3 is the Orange “Filter Section.” Here, a few different methods can be used to convert the DC ripple into a smoother voltage output. For our purposes, a simple circuit using capacitors alone can be used.
Capacitors, simply put, resist a change in voltage, or can be said to store an electrical voltage. What they do is charge up when we are at our ripple “peaks” and then release their voltage back into the circuit during the “valleys.” This should be easy to apply visually to our DC Ripple graph. Just take a hold of both ends of that bouncing line and pull, there, now you’ve got a straighter line.
Our filter will never give us a completely straight line. A straight line is the optimum high quality DC, just like you would get from a flashlight battery. We don’t need a perfect line to run our equipment, any fairly straight line will do.
In part 2 of this article, I will explain how to build a better filter section, and also discuss building a regulator section in order to achieve a more steady output result. Part 1 of the article will be enough to get most to a point where you can finish your peltier supply. Part 2 will discuss more advanced topics that will be unnecessary for most of us.
I spent about $30 building my supply, which is fully covered is part 1 of this article. It is by no means a perfect power supply, but it does its job well and is perfect for powering my 226W peltier and a 120mm fan. It has quite a bit of voltage fluctuation, upwards of 2 volts of ripple, between 12 and 14 volts output range. This is fully dependent on your filter section, and can be further reduced with a regulator section, discussed in part 2.
We need to look at the requirements for our Homebuilt PSU. We need to know all of the devices that we plan to power with it. We also need to know what kind of voltage each device is going to need. We need the current draw, in Amps, of each device.
This should be a fairly straightforward process. Peltiers can be tricky to determine the optimum voltage however. For a 15V Vmax Peltier, I would suggest a 11-14V power supply. For the 24V Peltiers, You can range them from 12-22 Volts with good results. 18V seems to be a nice solid number for most applications.
Our vast arrays of 12V DC fans can live happily up to 15 volts with little problems in most cases. When building a power supply that runs a 12-14V output, running fans from the same source should be no problem. While many of us use rheobusses to control our fan voltage, I will suggest a few alternate means you could use to power fans from a higher voltage source, such as an 18V application.
Our final consideration has to do with our input voltage. In the United States, our nominal supply voltage is 120 volts AC, 60Hz. In other parts of the world, you may have to deal with other voltages, ranging from 100 to 240 Volts AC, and could come in a 50Hz variant. While the difference between 50 and 60 Hertz (Hz) is not critical, the voltage is.
The input voltage will determine what power transformer we buy. If you are here in the US like me, choosing your components might be easier because I’ve already found sources for my PSU parts, which you can in turn use in your design. In other parts of the world you might need to check out some local electronics retailers to find the parts you need.
- Here is my list of PSU requirements:
- 120 Volts AC, 60Hz Input Voltage
- 12-14 Volt output voltage
- At least 24 Amps of load rating
Now, To Go Shopping
Using this article as a guide, you can go to an online electronics retailer and pick up most of the things you’ll need. I would suggest checking out the following retailers in the US:
- All Electronics
- Marlin P. Jones & Associates
- Jameco Electronics
- Hosfelt Electronics
Depending on your PSU requirements, you may need to buy from more than one retailer. I’ve included several practical examples and parts lists to aid you in building your own supply. In order to make this selection process easier, just take it a step at a time.
In order to make our first selection, we need all the data we compiled from the above section, Determine your PSU Requirements. Try to find a transformer that closely matches your requirements. The input and output voltages are important, but not as critical as the Amp rating of the transformer. You must find one that meets or exceeds your power requirements, in Amps.
The voltage ratings of transformers can be a trickier matter. A little knowledge can go a long way here. My Ideal transformer would say:
120V Primary, 12V/24Amp Secondary
In reality, it is not likely that you will find a transformer that meets your specific requirement 100%. Fortunately, this is fine and will cause us little problems. Looking at the transformer I bought, let’s see how this affects my project:
All Electronics product: Power Transformer, 100/115V Primary, 12.6V/25Amp Secondary. $20.00
Look at the amp rating first: 25Amps. This is acceptable for my project. Output voltage is actually dependent on both the input voltage rating and the actual voltage that you have in your house. In my house, I have 121.3 Volts at the wall receptacle I use for my computer. If you have a multimeter or any kind of AC volt meter, you should measure the actual voltage that you will be working with.
Now look at the input voltage ratings of the transformer: There are two: 100 Volts and 115 Volts. What happens when I apply my 121.3 volts to the 115 volt primary? The expected output voltage, 12.6 volts, will change accordingly. In my application, I got 13.55 volts output when hooked up to the 115 Volt primary coil. This is still a very acceptable number. Darn near perfect, actually.
Now what if I hook to the 100 Volt Primary? My output voltage jumps up to 19.26 Volts! This means that the “12.6 Volt” transformer would work perfectly for both my 226 watt, 15V Peltier, or any number of the 24V Peltiers that are around! So, these voltage ratings can be deceiving and should be carefully considered when choosing your transformer. I speculate the above transformer will be perfect for 90% of the US audience.
Other transformers to consider from the above mentioned retailers:
MPJA: 24V CT (12-0-12)10A Has 110/220V Primaries… Center tapped means this transformer can supply 20A at 12V
Online catalogs sometimes do not contain all of the products from a certain retailer. If you have trouble finding parts, order a print catalog from these sources. Often times, you will find a wider selection in the print catalog.
These are the two main transformer selections I would expect to find in a homemade power supply, as they are the most readily available application-friendly units I have found. The first from All Electronics is an excellent choice, although lacking a 220V primary for international use.
One other important consideration is the use of multiple transformers. If you are unable to find a transformer large enough to suit your needs, you can use two ore more operating in unison to supply your needs. With two of the All Electronics transformers, you could make a 50A rated power supply. If you need more than 50 amps, you’ve got one serious project going.
This should provide you with enough information to be well on your way towards finding a suitable power transformer. Finding sources for various components can often times be one of the most time consuming tasks in a project such as this. If all else fails, try large distributors such as Newark or Digikey.
A diode bridge is a much easier part to find and is very easy overlook the proper implementation of them. Bridge units that are large enough to handle the current output of a peltier power supply generally come in a nice big case that has good provisions for mounting. This is important because these diodes emit an extremely large amount of heat.
The most common types you will probably use are either the 25 or 35 amp variants. The voltage rating of the diodes is not critically important, however, you should at least have double the rating of your power supply’s expected voltage. So for a 12V supply, I think the closest reasonable number would be a 50V rated diode. I have 200V rated diodes in mine, because they were what was available in the 35 amp package.
It has been my experience that these diode packages are a bit optimistic on their amp ratings. I chose to use 2 identical diode bridges in my design. They were both 35 amp units, and my expected power supply output was only 24 amps. This serves a dual purpose, however. Not only lowering the strain on each of the diodes, it also lowers the heat output that each diode set will produce, which is the single most important factor with these bridge diodes.
I used an Athlon type HSF to cool one, which lowered the load enough on the other diode set to use a simple passive aluminum heat sink. The case temperature of these diodes should not exceed 50C under normal operating conditions. Use your thermal test probe to ensure that your diodes are well within their limits.
Using a simple passive heat sink on a single 35 amp diode yielded very poor results for me. After running the peltier and fan on the power supply for a few minutes, I measured the case temperature of the diode bridge. It was at an astonishing 120C! There was a faint burning stench, which is a good indicator that all is not well.
Pay close attention to the diode’s temperature – consider this fair warning! A good sized passive heat sink on a single bridge, which was not operating at anywhere near its peak current limits, was quietly self destructing. I would never suggest building this type of supply without a fan of some sort, especially if you build it in an enclosed case.
As far as finding diode bridges, there should be no problem. All of the mentioned retailers carry them in a variety of sizes. I highly suggest over-rating these when you spec them out, and always go for multiple units in a high-amperage PSU, unless you plan on water cooling it as well as your PC.
Heat sinks with attached fans, such as CPU HSF combinations are my best recommendation for cooling the diodes. The stock coolers that come with many processors work well for this purpose. They do require some modification in order to bolt the diode bridge to it, but this is a very easy process if you own any sort of drill.
The filtering element is a crucial part of our Peltier power supply. It determines what the final quality of our power supply’s output. For our simple linear supply, a capacitor filter is all that should be needed in most cases. When searching electronics catalogs, look for “snap-lock” or simply snap electrolytic capacitors. Typically, you will find the largest value capacitors in this section.
As a general guideline, I suggest at least 1000uF (micro Farad) of capacitance per 1 amp of output. This would be a bare minimum and would not be acceptable in some situations. A mid-grade supply would use at least 2000uF per 1 amp.
My supply has 1280uF/Amp, which is acceptable for my situation, but by no means does it provide a high-quality output. My total capacitance is 32,000uF, for a ~25 amp supply. Had I gone the extra distance and used 50,000uF instead of 32,000, my output would be considerably better.
When purchasing capacitors, you’ll notice that they have a voltage rating. This is simply the breakdown voltage of the insulation in the capacitor. The higher the voltage, the better (and usually thicker) the insulation has to be. This is why high-voltage capacitors are so much larger than the comparable low voltage ones. Always buy a capacitor with at least twice the voltage rating of what you expect it to see in your supply.
For a 12V-15V supply, consider a 35V rating the minimum acceptable value. For up to 15-24V, use at least 50V rated capacitors. Capacitors can EXPLODE when they are subjected to voltage higher than or close to their ratings. This can be a very dangerous situation, and can cause blindness from flying pieces of aluminum, and also hearing damage, as the explosion is very loud.
I have had the misfortune of this happening to me, and I can say honestly that you should avoid this situation.
I bought my snap-locks from All Electronics, which only has 3900uF capacitors as there biggest value. I just browsed Jameco’s site, and found a jackpot of 10,000uF, 50V capacitors. These are more expensive, but look like a great choice for our power supplies.
Remember, if you use these capacitors, an optimist would spec 10 amps per capacitor, while a higher quality supply would only place 5 amps per capacitor. The Jameco part number is 157737. I would err on the side of caution and have too much rather than too little capacitance.
Once you get these three basic components, the transformer, diode bridge, and filter capacitors, you’ll have enough to build your very own supply – almost. Make sure you pay close attention to all the small accessory items that you will need.
Well, now that you’ve got all your major pieces, its time to gather up all the switches, wire, and even a case if you so desire one for your supply. Don’t forget heat sinks and fans as well!
When picking out a switch, it is always a good idea to put the switch on the high-voltage side of the transformer. This will cut all power going to the rest of the circuit, and will be the safest location if you forget to unplug the power supply before working on it. Any switch that is rated for 15A/120V will work well for this supply.
Wire consideration is especially important when dealing with this kind of high-amperage supply. If you own a peltier, take a look at its wire leads. Chances are, these leads are much too small to carry the kind of current that the peltier will use. If you hook it up and wires are warm, that is a good indication that they should have used larger diameter wire. Here is a table that I use to determine the maximum amperage that a particular wire should be subjected to:
18 Gauge <=5 Amps
16 Gauge 10 Amps
14 Gauge 15 Amps
12 Gauge 20 Amps
10 Gauge 30 Amps
8 Gauge 40 Amps
6 Gauge 65 Amps
For those with a 226W peltier, the wire suggestion would be 10 gauge. Go buy a length of 10 gauge wire and compare it to the power leads on your peltier. See any difference? Now, in order to minimize voltage drop to your peltier, I would suggest cutting the leads on the TEC device as short as possible before attaching them to the power leads from your power supply. I would suggest stranded wire, not solid, for all our applications here.
Many different methods exist for attaching wires together. For high-amperage, low voltage applications such as ours, I will only suggest one method of wire attachment. Buy yourself some heat-shrink tubing, and solder every wire to wire connection in your supply. This will minimize the voltage drop throughout your supply.
Make sure you place a length of heat shrink tubing over the wire before you solder them together, and always slide it back far enough so that the soldering iron doesn’t shrink the tubing beforehand! Using a cigarette lighter works well for shrinking the tubing once it’s in place. On a side note, Radio Shack’s heat shrink tubing is the worst quality, overpriced tubing I have ever found.
Hosfelt, on the other hand, carries some very cheap, generic heat shrink that is the best I’ve ever used, and it comes in any length you want and in many different sizes. You may have to order a catalog from them in order to find it, their online store doesn’t reflect on their full inventory well.
Cases are going to be a matter of preference for most people, so I won’t attempt to make any recommendations here. You can buy cases from the electronics stores mentioned, build it into your case, or come up with your own creative solution. If you find something that really works well, share it on the Overclockers forums!
Don’t forget to buy some sort of line cord to plug into the wall and attach to the high voltage side of your transformer. You can find sockets like you see on the back of your PC’s power supply, and use a normal computer power cord for your supply. This gives it a very professional finished appearance.
Some sort of breadboard should be used to mount your capacitors on. You could alternatively make a PCB from a kit that is customized to your specifications. If you have experience with this, it is a very good choice, but be aware that any current-carrying traces should be made as wide as possible. You could also mount your diodes on the PCB, if you can incorporate a good way to mount the heat sinks.
Make sure to use a fuse in your design. For most PSUs a 10A/250V glass-type fuse will work nicely. Use this on the high voltage side of the transformer. You can buy both inline and panel-mount fuse holders for these types of fuses.
Any number of lights, panel meters, buzzers, or whirligigs can be added to your design. I encourage you to add whatever you want to your PSU, and make it unique to your own needs. Now, all I have for you is a few notes on assembly, and that will wrap up part 1 of this article.
Actually building the power supply may be the easiest step. As long as you pay attention and have some decent soldering skills, you should be able to assemble it fairly quickly without mistakes. It is important at this stage to familiarize yourself with each device package, and understand which components leads do what. Errors are going cost us money at this point, so be careful and use all of the available information to guide you through it.
Starting with the power transformer, you must identify which leads are primary leads and which are secondary leads. Sometimes they are clearly marked, other times, you will have to make an educated guess. Take a look at the power transformer from All Electronics that I used in my supply:
Notice that there are three leads coming around from the back of the transformer. Since there are two primaries on this transformer, a 100V and an 115V, it was easy to identify these as the primary leads. They each have a female quick disconnect on the end. There are three colors, white, brown, and black.
The white lead is common, which means it is used by both the 100 and 115V primary. The black wire is the 115V primary, and the brown is for the 100V primary. You will either use the white-black or white-brown combination for hookup. These hook straight to your line cord that plugs into your wall outlet. Make sure to wire in a fuse holder on one of the leads, and wire in a switch if you want one.
AC wiring does not have polarity, (+) and (-) like DC wiring. AC wiring has two wires, one hot and one neutral in the United States. It does not matter which one you hook to which lead on the transformer. Transformers do not have a connection for the ground wire. Hook the ground wire, if your line cord has one, to the case that you are building your power supply in.
If you look at the front of the transformer, you will see two silver tabs sticking off to the right. These are the secondary output quick disconnects. You could use a female crimp on quick disconnect here, but I would suggest soldering. These will attach to the two wires that you run to the diode bridge(s). These are also AC, and it does not matter which is which, i.e. not positive/negative leads here. Make sure you use big enough wire to handle the current.
Now, lets move on to hooking up the diode bridge. These come in a square case, often with a chamfered corner, indicating the DC (+) positive output. Here is a picture of a common case style, and the one that I used with my supply:
These are very straight forward to hook up. When using multiple diode bridges, always make sure to hook them up identically, each wire to the same location on the two bridges. From this point on, you will be working with polarized DC outputs, so care must be taken to distinguish between positive (+) and negative (-). I always use black wire for (-) and red wire for (+), this helps me keep everything straight and minimizes the chance of errors.
There is a hole in the center of the diode bridge package; this is where you can place a bolt to attach it to a heat sink. If you use a CPU cooler, drill a corresponding hole through the center of the heat sink to mount the diode. You might be able to get away with using a thermal adhesive, like Artic Silver Thermal Adhesive, but I would trust a bolt over an adhesive. I use regular white-goop thermal paste with these diodes.
The electrolytic capacitors we use look like small cans, and have two leads. Snap-lock electrolytics are polarized, which means they have a positive and negative lead, and must be hooked up correctly. A white stripe down one side of the can indicates the negative lead. They are all wired in parallel, which means you will have two wires between each capacitor, a positive and a negative. Follow the provided schematic and you should have no problems.
I have drawn up a schematic that can be used for many purposes for your use. This schematic does not include any of the advanced features that will be added in part 2 of the article. I will include schematics of those things in part 2. Please feel free to request additions, corrections, or suggestions on things to add to part two of this article.
Part 2 will deal with building a regulation section, dealing with multiple voltage rails, adjustable voltage rails, and fan control. This schematic is an unregulated, 12-18V, 25 amp power supply. The cost of building this particular supply is about $55 with the specified parts, plus the cost of an enclosure if you wish to use one. This is not including the cost of any heat sinks, since I could find none that I would specifically suggest, other than a CPU cooler.
Looking at the schematic, you can see how truly simple it is to build your own Peltier power supply. I look forward to your feedback on the forums or via email. If you need any sort of specific help, I will try to help you as much as possible.
In order to benefit others, I would ask that most questions be directed to the Overclockers Forum, where it can be a matter of public record, rather than having to answer 30 emails about the same problem. I truly hope this provides enough information to convince some of you that building your own power supply is indeed worth it.