Skip to main content

Flight test success!

This weekend I was blessed with some good (great) weather. Saturday morning was cold, but very calm and clear.

20091114-DSC_5358

My friend Tom agreed to be my test pilot for the first (well, second 😉) flight. He got it all trimmed and ready to fly. I say it was the second flight because while I was taxiing it around, I decided "what the hell" and decided to fly it… Well it didn't end that well.

Takeoff cartwheel]

Takeoff cartwheel

Looking on the bright side, the result is a testament to quality of my construction (read: luck). The only damage was a little dirt on the wingtip.Not too bad. :)

Cartwheel"damage"

Cartwheel "damage"

After that bit of excitement, I decided it would be best to let my test pilot try again ;) This time, the plane flew quite well.

First(controlled) flight!

First (controlled) flight!

Above is one of the first moments of controlled flight for my humble plane. I have to say, it was very exciting!

I don't have a lot to say about the rest of the photos, Tom took most of them while I was flying (including the lovely cartwheel photos). Enjoy them!

20091114-IMG_406720091114-IMG_404520091114-IMG_4042

Really cool reflections and refractions

Really cool reflections and refractions

20091114-IMG_3970

Finished building!

I've just finished building the first radio controlled airplane kit since, I don't know, over a decade. I tried to build one 8 yeas ago, but never finished. Anyway, It's pretty exciting. I'll share some of the building process with you. My intent from the beginning of the project was to build an arial platform for autopilot development, photography, and, perhaps, an attempt at FPV (first-person video).

With those goals in mind I decided to select a "trainer," which is a very stable plane intended for teaching new pilots. There are several advantages to a trainer in this context. In a way, I'm training an autopilot to fly, so it's a fitting. In addition, a typical trainer has a simple, boxy fuselage. I think that the simplified geometry should be easier to build around. The individual plane I settled on is the SIG Kadet LT-25, and my mom was nice enough to get it for my for christmas a few years ago.

I wanted the plane to be on the small side because I'm 100% electric for all my planes and helicopters. I don't have any investment in nitromethane power, and I don't want any. It's stinky, loud, dirty and annoying. At the same time, electric power is getting better and better. The batteries are increasingly amazing in their capacity and power. Unfortunately, I don't have any early building photos. But, I'll talk about some of the photos I do have:

Wing servo mount

Wing servo mount

This is the only photo I have of the wing being built. It went together really well, the laser-cut parts are amazing. I'm not all that great at making the wing halves meet up right, but lots of epoxy and filler remedied the problem(hopefully). In this picture I haven't added the fiberglass reinforcement around the seam. Finally, notice that the servo hole is a strange shape. I added a brace in the middle of the opening because 2 micro servos are slightly more narrow than one full-sized servo. Using 2 servos allows me to make the ailerons behave as flaps as well as ailerons. This way, I can adjust the lift of the wings to compensate for the additional weight of payload packages.

Tail mounted servos

Tail mounted servos

As I mentioned, I wanted to make it as easy as possible to add payload modules to the aircraft. To make this easier, I moved the servos for tail feathers surfaces to the tail itself. This has 2 advantages, I don't have to waste space in the radio bay, and it should offset the weight of the comically giant motor I bought. There is some additional parasitic drag, but I hope it isn't too much.

Brushless motor

Brushless motor

Speaking of the beast, this is the comically huge motor. It's a scorpion, which claims to be better than the others because the magnets are rated to 200degrees celsius while also having the strongest magnetic field (which is apparently rare). It's rated to 80 Amps and 1200 Watts continuously. That is WAAY more than what I need for this model. I'm not sure what I was thinking when I bought it, but it is what it is. I spent some time on the motor calculator to figure out a prop/battery combination that seems appropriate to the weight of the model. Most people have theirs coming in at a finished weight of 4 pounds, so I assumed 5 pounds as a starting point and went from there.

Scorpion motor calculator

A common rule-of-thumb for electric flight is that for "reasonable" sport flying 40 to 50 Watts/Pound are necessary for brushed motors and 35 to 45 for brushless motors. In the performance results notice that the current (24 Amps) for a 3-Cell LiPo (11.1Volts) ends up with about 266 Watts and 53 Watts/Lb, which is just about right. My speed controller is rated for a 4-Cell LiPo, which would enable about 400 Watts of power. I doubt I'll need that much, but I guess it's nice to have the option.

Spinnerbinding

Spinner binding

Several companies, such as APC, developed special props for electric motors. They are much more efficient than the older models. There is renewed interest in propeller design, even in the general aviation circles. One of the leaders of the movement is Paul Lipps. He has an interesting article on EAA's (experimental aircraft association) Experimenter magazine. Some qualities of his designs are also present on this propeller. As demonstrated in the photo above, the angle of attack at the root of the propeller is much greater than other model props. This caused a problem for the spinner I planned on using. I had to carve the opening larger to accommodate the prop and give some room for comfort.

Power system installed

Power system installed

I skip ahead here a little, but here is a photo of the motor and speed control installed in the fuselage. At this point, the motor is just sitting there but it doesn't look any different bolted in. I later moved the speed control slightly. In the photo it's sitting above a balsa panel, I was able to move it underneath. This left plenty of room on top of the panel for the battery pack.

Top of the GPStray

Top of the GPS tray

I wanted to build a place for the GPS and autopilot circuitboards. I knew thatI could install them in the radio compartment, but it would have been covered by the wing. This photo is looking through a lightening hole on the top of the fuselage, just behind the wing. I found some plastic 'U' channel and thought it would allow me to slide electronics trays into this bay. It's necessary because once the covering film is installed, I won't have access to the bay from this angle.

Front view of the GPS tray

Front view of the GPS tray

Here is another view of the GPS tray. I had to take this picture from the firewall to get the right angle. You can see the bracing I made underneath and alongside the plastic channels to add rigidity. On the bottom of the fuselage, you can see 2 balsa sticks. Those are the mount for the tailwheel steering servo.

Tailwheel steering servo mount

Tailwheel steering servo mount

Here, the tailwheel steering servo is mounted. I decided that I could sacrifice this amount of space in the bay. I didn't want the servo to be mounted upside-down on the bottom of the plane because it's rainy in Oregon, and I didn't want water to splash up from the main gear onto the servo. I also thought that if I had a landing onto tall grass there was an increased risk of damaging the servo because there would be more places for things to get caught. The funny-looking bulge on the right is there so I can remove and install the servo. The sticks are just wide enough to fit the body of the servo, but there is a wire coming out that makes it hard to mount otherwise.You can see that the control wires rub against the side of the former. I later carved them a little to allow for clearance.

Partially covered

Partially covered

As things move along, here is a photo of the plane partially covered. The entire fuselage is covered, and the top of the wing is, too. You can't really see, but the tail surfaces are covered and installed. I haven't bought the wheels at this point, either.

Perfectly balanced

Perfectly balanced

Now, things are just about finished. I installed an eye-bolt in the top of the wing at the Center of Gravity (CG) to check the plane for lateral and longitudinal balance. Notice that there is a metal tube sticking out of the leading-edge of the right wing. I decided to install a pitot tube for use later.It's made out of aluminum and goes into a silicone tube down the length of the wing to another aluminum tube that exits out the bottom of the wing. I usedBernoulli's formula for determining air speed using a pitot tube and determined that I need to get a 1500 Pa (1.5 kPa) differential pressure sensor, assuming 50M/sec as a maximum speed (which is about 100Mph).

Bernoullis formula, Kpa versus M/sec

Bernoullis formula, Pa versus M/sec

Anyway, back to the plane, I've finished covering it. I added a yellow stripe along the whole bottom of the left wing that wraps around the front of the right. This is a high-contrast indicator of the orientation of the plane.These aid in the pilot's ability to determine the orientation of the plane.

Silhouette disorientation

Silhouette disorientation example

The idea is to make the 2 possible orientations of the plane look very different through the use of color. In this example, my plane would be different because the image on the bottom-left would have a stripe along the length of the wing and on the one in the bottom-right image, there would be a single diagonal stripe on the right wing

Top of the plane, fully covered

Top of the plane, fully covered

Bottom of the plane, fully-covered

Bottom of the plane, fully covered

Well, long story-short, I'm done building the airframe. I need to work on the electronics now. Luckily (I'm trying to be up-beat about this) the weather is absolute shit, so I'll have plenty of time to work on it. Hopefully there will be a break that I can use to take it on it's maiden flight.

All dressed up with nowhere to go.

All dressed up with nowhere to go.

There it is all finished. Note that the deck isn't wet because I sprayed it off. We're supposed to have "torrential" rain this weekend. 🙃

iPhone controlled TV Lift

Video of the iPhone TV Lift controller working!

Hardware

This was probably one of my most time consuming projects. Hopefully I can do some justice to the time and work spent through this post. I should probably begin by describing the whole system. First, the TV is lifted up & down by a"Lift Tech" lift. By the way, they get the prize for the most original company name ever. Their controller box has a port that allows home automation systems to control its operation. They did a really good job of making it extensible. You can control it in a variety of ways: short one pair for going up, another for going down; short a pair for down, open for up; etc. I had intended to use the first mode, but I think I fried a channel on my opto-isolator. I ended up using the second method because it only required one channel. Unfortunately that means I can only have the TV up, or down. I wasn't too disappointed, though, because I don't expect to want it any other way very often ;) .

TV Lift controller box

TV Lift controller box

The board I designed to interface the TV Lift to the server was fun to design and build, even though it was prone to error. I had originally intended to use the Microchip ENC28J60 ethernet controller. I thought it'd be cool to have the iPhone app connect directly to the board's ethernet controller and microcontroller. Unfortunately, I screwed up the interface from the ethernet controller (specifically the physical layer circuitry) and the magnetics. This interface is harder than it looks, trust me. I thankfully included a serial port on the board (which I had other plans for) and used that instead. This choice made the microcontroller software extremely simple. All it really has to do is wait for a 'd' character over the serial port and lower the TV, a 'u'character lifts the TV, and a 's' character queries the controller for the current state. I've included an image of the schematic for the board, in case you are curious about my lofty intentions.

TV LiftSchematic

TV LiftSchematic

If you're interested, the board I used on my reflow soldering toaster oven page is the TV Lift board. Incase you don't want to go over there, here is a picture of it mostly finished up:

Controller board installed

Controller board installed

Using the RTS line for reset

Using the RTS line for reset

There is one more thing that I had to change. For some reason the controller board that I made crashes after a while. I have to use a ladder to reboot it(power-cycle). Hauling the ladder around is annoying, and I don't like doing it. To allow myself to reboot it remotely, I added a diode from the MCLR line(reboot) of the microcontroller to the RTS line of the serial port. The RTS line is used for modems from back in the day. Now I can use it to force MCLR low, to reboot the controller, if it's "low" (-5 volts). It's a hack, but what are you going to do?

Software

The software for this project is deceptively simple. The iPhone software basically connects to the server using a socket, and if the "up" button is pressed it sends a 'u' over the socket, and a 'd' if the down button is pressed. That's it! Server software is almost identical, however it listens to the return of the serial port and expects my controller board to return a 'U'or a 'D' from the 'u' and 'd' command. If these are not received, the board is reset and the command is tried again.

Conclusion

Well, that's all there is. I hope you've enjoyed it, I know I have. Please, feel free to ask any questions in the comments section!

Weather Station Online!

I've finally gotten my home weather station online!

I'm using a Oregon Scientific WMR918 system with anemometer (wind speed), wind direction, temperature, humidity, and barometric pressure sensors. Click Here to access the current and recent readings, or the Weather page of an integrated page (also available in the navigation bar below the blog title)and "read more" to find out how it works.

The WMR 918 has a serial port that sends-out all the data that it collects(and lightly processes). I have this signal going into my home server, which is running the open source wview website. Because my weather setup is a bit different than what is typical, I had to make some changes to the source code. (probably the best part about open source!) I guess most weather stations come with a "mushroom" sensor which is a thermo hygrometer (temperature and humidity) that is encased in a radiation shield (so direct sunlight doesn't affect the readings). I don't have one of those. My outdoor temperature sensor is reported differently on the serial output. The wview software was expecting a certain kind of data packet, and having never gotten it, it never initialized. Once I changed the source code so that it ignored the difference I was all set. B.T.W: The wmr918 init issues thread about this process is on the Wview google groups page.

I've gotten the data up in several personal weather aggregation services. I've decided that I don't really like weather underground (way too cluttered), butI'm up on PWS (Personal WeatherStations), and APRS (Automatic Position ReportingSystem). APRS is a weird amateur radio thing. Basically, what it was designed to be is a way for small bits of digital data to be passed around so that, especially during search-and-rescue, operators can know where one another are. Over the years it has been expanded to include many interesting uses. One of which is the sharing of personal weather station observations.Eventually, NOAA even started using the data in their models. One of the many things I want to do, one of these days, is install a APRS node in my car.I know it is geeky to the max, but that is what I am.

Anyway, back to the point, everything works! I hope you enjoy the fascinating(not really) weather trends from Philomath Oregon! Oh, and hopefully soon I'll have a handy widget that displays a digest of the current conditions in this sidebar. But for now, feel free to check the current conditions page.

Toaster Oven Temperature Profile

Reflow Oventemperature profile

Reflow Oven temperature profile

I've decided to spend some time profiling the toaster oven. In addition, I've made a simple controller board that may get the brains to manage a reflow operation on it's own one day. For now, I'm satisfied learning more about the oven's capabilities. An interesting thing to notice is that it (appears) incapable of satisfying the recommended profile from Kester. I wanted to start with a very simple controller that gets most of it's brains from an attached computer. This is a somewhat central tenant of the way I build most projects. The reason for this is that computers have really huge screens, and nice keyboards so it's easy and cheap to have a cool UI. For this project I'm still working on the OpenGL for the realtime graphing of the temperature profile, but you can see what I have designed so far:

Oven ControllerUI

Oven Controller UI

The left side of the window is where the realtime graph window goes, the oven system status is in the upper right, with the reflow profile defined by the user in the bottom right. Anyway, back to the controller hardware.

The controller is built around a PIC microcontroller and the Maxim 6675. The 6675 is a thermocouple conversion chip that handles all the analog complication of dealing with thermocouples (including cold-junction temperature). The microcontroller interfaces with it through a simple synchronous serial protocol. One thing that you have to keep in mind with this chip is that if you request a sample more than 4 times a second then it doesn't finish the conversion. I decided to let the microcontroller handle waiting for the appropriate amount of time, taking a sample, and sending the data over the serial port.

Because computers don't have serial ports anymore, I have to include serial toUSB converters on everything. In this case I used the FT232RL schematic block that I copy and paste whenever I need it.

Reflow OvenController Schematic

Reflow OvenController Schematic Full size PDF

Like almost all my projects, I got a board made at BatchPCB. It turned out great:

Oven ControllerBoard

Oven Controller Board

Unfortunately, I wasn't able to find a nice socket for the somewhat standard plug they put on the end of thermocouples. In this case, I just used a screw terminal. The posts on the left go to the thermocouple wires. I've included another pair for adding a relay that could be used to switch power to the oven. Also, I've included a port that could control a R/C servo that could open the front door to facilitate rapid cooling of the oven. Eventually,I'd like to allow this controller to begin a reflow cycle when the button in front is pressed, runs until the target temperature is reached, shuts off power, and opens the door at the perfect time. For now, I still have to watch and stop when I feel it's right.

The thermocouple is installed just by using the friction of threading it through the oven rack that is installed above the board.

Oven thermocupleinstallation

Oven thermocouple installation

To keep everything relatively tidy, I double-stick taped my controller to the side of the oven:

Oven ControllerInstallation

Oven Controller Installation

Finally, I'll analyze the graph from the beginning of the post. For your convenience, here it is again:

Reflow Oventemperature profile

Reflow Oven temperature profile

The blue graph is the recorded profile used while soldering an actual board based on my observations. I stopped the oven cycle based on the advice I gave in my other post about the Toaster Oven Reflow Soldering post. For the Green trace I ran the oven until the thermocouple read 210°C. I decided to run it without a board in it because I didn't want to smell burning FR-4 ;). I'm not saying that it would have actually burned it, but I realize now that I really need to calibrate my thermocouple against a trusted reference thermometer.

To compare the actual results against the specifications in the Kester datasheet, I've included a line for the maximum and minimum profiles. For the preheat phase of the profile all that is required is that the temperature increases less than 2.5°C/second, up to 150°C so I didn't include it. Once it reaches 150°C the temperature needs to increase between .5 and .6°C/second up to 180°C. From 180°C to 210°C temperature needs to increase at a rate between1.3 to 1.6°C/second. As you can see the toaster oven almost keeps track with the 150-180° rate, but can't achieve the 1.3-1.6°C/second ramp rate. I'm not sure what the exact effect of this is, but I've had good results so far.

Anyway, it was an interesting exploration. I do kinda wish I was able to match the correct profile, but I'm glad in a way that it doesn't because I don't have to implement a PID temperatue controller :).

Adding a serial port to the Netgear WGR614L

Netgear Serial Port

I added a serial port to my wireless router so it could monitor the weather. I know, why does a router care what the weather is like Well, it doesn't, but Ido. Actually, I want to add a weather widget to this blog and also upload it to weather underground (not the one with Bill Ayers ;) ). <this postfor details. >> I haven't finished any of that yet, but it came in handy later when I bricked it! When they designed this AP, Netgear was nice enough to make it easy to open.All that you need to do is pull back the rubber feet and use a T-9screwdriver. The little feet even stay captive during the process, mighty convenient.

Removing the screws

Removing the screws

The bottom plate comes right off, revealing the mainboard, which lifts out.

Disassembled

Disassembled

Unfortunately, I don't have a device without headers installed, but you can see them installed in the lower left corner. The single-row one is the serial port, the double-row is the JTAG. From the factory mine had solder blobs on the serial port. I had to use solder wick to remove it. Once the holes are cleaned out, installing the header is a snap.

Now that it's easy to attach to the port, we need to convert the 3.3v CMOS serial signal to true RS232. The typical way you do this is to use a RS232transceiver. I had a SIPEX transceiver lying around, so I used that. Just follow the datasheet for whatever you use.The pinout for the serial port on the netgear is:

  1. VCC
  2. RXD
  3. N/A
  4. N/A
  5. TXD
  6. GND

I've attached the approximate schematic that I used. There are a couple of important things to note, because the router uses 3.3v and the RS232transceiver uses 5v we need to negotiate the difference. Referencing the transceiver's datasheet, the minimum voltage necessary for a '1′ is low enough that a 3.3v CMOS signal is sufficient, so we can just hook it up. The signal going toward the router, however should not exceed its VCC. To achieve this using a minimum of parts we use a pull-up resistor fed from the device's VCC and a diode biased toward the transceiver. This way, when the transceiver is outputting a '1′ the 5v are blocked, but while outputting a '0′ it will drop the voltage sufficiently enough for the router to see it as a logical low.

schematic

schematic

Top view of theadapter

Top view of the adapter

Bottomview

Bottom view

If you want the router to be the "slave" of the serial connection (like a modem) then reverse the TXD/RXD pins on the DB9 connector (pins 2 &3). Is pecifically wanted the router to be the master to hook it up to other stuff(my weather station, remember).

To get the 5V necessary for the transceiver I tapped the 12V power from the wall adapter and fed it through a 7805.

Stealing 12V from the wall adapter

Stealing 12V from the wall adapter

Adapter installed on mainboard

Adapter installed on mainboard

Adapter board installed

Adapter board installed

Mainboard and adapter in case

Mainboard and adapter in case

Connected with aNULL Modem cable

Connected with a NULL Modem cable

In the above photo I'm using a NULL Modem cable because, remember, the router serial port was installed as if it's a computer.

Netgear preboot console

Netgear preboot console

It works! The photo above is the preboot console of the router. Notice that the MAC address is FF:FF:FF:FF:FF:FF. I'm not just paranoid, this is one of the failure modes of these routers. The DD-WRT firmware overwrites some of theNVRAM in a weird way and this happens. It's pretty screwed, but at least with a serial port you can fix it :)

Anyway, enjoy!

House Construction Time-lapse movie

It's been a while since my house was completed, but I've finally gotten around to finishing the video compilation. It's a bit of a long video, but I think it's pretty interesting. There is one frame per minute for over a year. I've removed all the time where there wasn't active construction to try to keep it short. If you "read more" I have details about the construction of the outdoor camera housing, and mounting. Also, I talk a bit about the details of the wayI captured the frames themselves.

I intend to add more details here when I get a chance to take some proper photos, specifically, I will add a section on the construction of the outdoor video camera housing one wiring.

System Schematic

Schematic ofTime-lapse capture system

Schematic of Time-lapse capture system

I used SecuritySpy by BTV to capture the frames. This software was intended for businesses interested in "Loss prevention," otherwise known as catching shoplifters. For that reason, it's a little spendy and the features aren't perfect, but it does work rather well, and it's reliable. There are 2 primary features I was looking for: Storing a time-lapse video, and providing web access to the camera. Security Spy does provide these functions, and it let me set a schedule for beginning and stopping the recording for each day on a weekly schedule. Also, you can modify the settings over the web interface. I had some trouble with SecuritySpy crashing, and by contacting support they provided me with an AppleScript that monitored the application and restarted it.

The problem with daily recordings, however, is that they need to be combined.About once a week, I would download the videos for that week and remove the times the crew wasn't there, and combine them in iMovie. Now that the project is over I could never imagine the pain doing all that work in would sitting would cause.

Toaster Oven Reflow Soldering

Soldering Surface Mount Devices (SMD) can be challenging work, especially with the no-lead (legs, not pb) parts that are so popular with device manufactures these days. A useful and surprisingly easy solution is "Reflow" soldering. Reflow is when you use a paste of solder to hold down parts while the board is heated in an oven. During heating, the solder melts and surface tension makes the parts align themselves in a much more uniform way than is possible with an iron. I've got more information about the toaster here, including a controller I made for it.

Board used in this example.

Board used in this example.

Preparing for reflow is a simple process. The most important part, as usual, is having a good foundation. In this case that means a good circuit board. I have made my own boards in the past, but BatchPCB is just so damn convenient. They create 2-layer boards with soldermask and slikscreen for $2.50/inch. The turn around time is pretty long, ranging from 3 weeks to as long as 6 weeks. Here is an image of the board I'musing for this example, and one I got from BatchPCB.

While it's not completely necessary to have a solder mask, I find that it really helps keep the molten solder where it is supposed to be. If you make your own board, make sure you get a complete etch and that there are no specks of copper around pads. They could prevent the solder from staying where it belongs.

Laying out the parts

Laying out the parts

I've just started to lay out all the surface parts before I begin a large project. It's kinda nice, all the stress of finding the parts is isolated to a time where you aren't in a hurry. I printed out the sheet twice as large so there would be plenty of room for the paper tape that carries passives.

Solder paste syringe from Kester

Solder paste syringe from Kester

Once everything is laid out, ready to go, It's time to break-out the solder paste. I keep my solder paste in the fridge, but I've heard of some people that just leave it lying around, and have for months, and say that it's just fine. I like applying it cold, because it's a bit more viscous; though that means you have to press a little harder on the plunger.

Solder paste applied to board

Solder paste applied to board

Placing the paste on the board is probably the most delicate step. If you put too much on some problems can arise. While the board gets warmer the paste spreads out (like a cookie baking) and if there are several components close together the blobs might join. If this happens the surface tension may drag them together as the solder melts. This has only happened to me once, though.Another reason to avoid too much solder is that you'll create solder bridges on IC parts with close leads. To be fair, these are nearly unavoidable.Finally, if you add too much the extra may squeeze out in a really strange way. If you put too little on you'll need to add more solder with an iron.There is clearly a balance, but I'd err on the conservative side. The amount I put on in this picture was about right. Basically, it should look a bit like a Hershey's kiss. For ICs just put a bead along the leads. Put on a little as possible while covering all the pads evenly. I had to remove some with solder wick from the IC's but not very much.

Placing the components on the board

Placing the components on the board

Putting everything the board is easier than with standard soldering. You don't need to make sure that things are lined up very accurately. However, for the reason outlined above, it is important to make sure that the solder paste doesn't touch between the small components.

I place the boards in a cold oven, and bake at 450°F for about 8 minutes. You should watch to see the solder melt. Don't worry about thermometers, or controllers, really. Trust your eyes. If the solder melts you know that you have achieved the proper temperature. Leave it in there for about another 30seconds with the heat on to make sure that you've at least reached the critical temperature everywhere (while hopefully not exceeding it much). Turn the heat off and leave it in for about another 20 seconds. Open the door and wait for about 5 minutes and remove the part. That's all there is! Good luck!

All done!

All done!

ATX Bench Power

See, it works!

Finished product

With many devices these days, power comes in the form of DC 5 or 12 Volts. It seems that battery chargers are especially bad because they require upwards of10 Amps and don’t include a wall-adapter. Probably one of the best solutions to this problem comes in the form of a standard computer power supply (ATX).These devices have several nice qualities and only a couple caveats. They are capable of providing plenty of current: 10-15A 12V, 10A 5V, etc. are small in size, and generally efficient. However, they also require a couple of considerations: cooling needs, voltage levels (12V is really more like 11.5V),and Switching Noise. Switching Noise comes from the way they are designed to be small, efficient, and cost-effective. Switching-Mode Power Supplies generate a given voltage by turning on-and-off a higher-voltage source so that the average of the output is the desired voltage. This is then filtered to smooth the result enough such that it is appropriate for the application.Computers don’t require voltage sources that are that clean, they are almost always run at 3.3V and lower. They generate these voltages in a different way.Anyway, enough justification for the choices made, on with it: First, though: Do not follow these instructions! I have no idea what I’m talking about. If you follow these instructions, you’ll die, you pets will talk and the world will cease to be. I take no responsibility for the idiotic things you do.

Step 1: Disassemble the power supply:

Be careful for latent energy in the capacitors!

Be careful for latent energy in the capacitors!

First, you want to find the screws that hold this thing together. They are often in a bit of a strange place, on this model they are on the top, far away from where the case separates. Notice that the sides are held in by the lip punched into the steel. Also, you may need to cut-off the ends of the long wires at this point (the connectors that attach to your motherboard, CD-ROM disk drives, etc) as they will no longer be needed.

There are a couple things to pay attention to about your power supply. You'll want to check to see if the main power connection and the fan have removable connectors on the main circuit board. If they do, you're in good shape. If not, it's still possible to continue, but it'll be a little more challenging.You'll have to carefully solder them once everything is installed in the box.

The next step is to remove the "IEC" socket and switches (On/Off and 110/220)from the case. You may have to do more or less than I had to, depending on the specifics of your salvaged power supply. I had to desolder the wires from theIEC socket to the On/Off switch, and the wires from the switch to the harness that connected to the mainboard. The 110/220 switch was mounted fully on the inside of the case and was left alone.

What was required on my box

What was required on my box.

Now, we need to prepare everything for mounting in the new wooden box. Because these power supplies are produced in mass quantities, the manufacturers do whatever they can to save money. This usually includes using single-layerPCBs. What that means it they have to move the ground plane off the PCB. This means that the metal case is the ground plane. So, we have to keep it. In this case, I measured exactly how little case I could get away with and broke out the dremel. Here is all the parts you need from the case (not including the main board):

All you need fromthe case

All you need from the case

Step 2: Preparing the box

Now, we need to find a wooden box to encase the power supply. I found mine from the local craft store. In the next image you can see that I've cut holes for ventilation, the switch and IEC socket. We also need to re-solder the wires from the socket to the switch. Make sure that the HOT wire (black) goes through the switch.

Ventilation, switch, and wall power

Ventilation, switch, and wall power

Power wires soldered back on

Power wires soldered back on

Glue the mounting plate into the box. Make sure that the plate won't chafe the wires leading to a short, and that the wires can reach. I used super glue (CA)to glue the plate to the box. I was happy to note that it bonded to the metal pretty well.

Mounting plate glued down

Mounting plate glued down

Step 3: Installing the Power Supply

Screw the power supply mainboard onto its mounting plate. Make sure you mechanically and electrically attach the ground wire to the plate. During this step, it is imperative that you make sure the ground wire doesn't short anything out. Now is also the time to attach the wires to the mainboard.

Wiring the input to the mainboard

Wiring the power input to the mainboard

I glued the 110/220 voltage switch to the side of the inside of the box, using CA again.

The 110/220 switch glued to the case

Now, we strip and gather the 12V wires (yellow, usually) and an equal number of ground wires (black) and attach either spade lugs or something else convenient.

spade-lugs

Get your hands on some banana plugs to use. They're available at RadioShack, but they suck, so try to get them somewhere else. Drill some holes on the front and mount them.

Mounting of the12V wires

Mounting of the 12V wiresIn order for the power supply to work, the "PS_ON" wire must be tied to ground. This is usually the green wire, but you should probably check for yourself. Also, now would be a good time to "Smoke test" your installation.Sometimes, switching power supplies need to have a non-inductive load to turn on, such as a light bulb or something.

Step 4: Finishing up

AirflowBaffle

Airflow BaffleThese switching power supplies are really efficient, but often need some active cooling. I decided to make a baffle plate to mount the fan in it. I traced around the fan and cut it out, and center punched the screw holes. I think it fit nicely. The writing on the lid is the voltage and amperage rating copied off the original case.

Below, I've included a a photo of the whole device. Everything seems to fit really well.

All done (only12v included)

All done (only 12v included)

See, it works!

IR Remote Control Transceiver

IR Transceiverbuilt using home made PCB

IR Transceiver built using home made PCB

For use with my home theater PC I developed an IR Transceiver by combining 2projects (Receiver, Blaster). Note that this device maybe taxing of your serial port, I take no responsibility for any damage you cause to your equipment. That said, I’ve provided PDF’s of the silkscreen, copper layout, and the Eagle PCB files. But first, the parts list:

  • Common - PCB Board: FR4 Single-plated copper PCB
  • SERIAL: 90° through-hole female DB-9 connector- Transmitter
  • D2: 1N4148
  • R2: 1kΩ Resistor
  • XMT: Infrared LED- Receiver
  • D1: 1N4148
  • R1: 4.7kΩ Resistor
  • IC1: 7805 Voltage regulator TO-220
  • REC: IR Receiver (Radioshack cat.276-640)
  • C1: 4.7μf Capacitor

irxcvr

IR Transceiver files