Silent Fan With Thermal Controller

Although I don't need an air conditioner to live in my apartment in Milan, temperature in Summer reaches 30°C and more, so some of electronic devices in my living-room suffer for the absence of ventilation. Specially the hi-fi amplifier and the XBox360 are in a low cabinet in front of television. To avoid to see the three red leds of death on my console, I decided to add a fan to the cabinet back panel. The best will be have a thermal fan controller but... hey, we're on instructable: we can make it!

Step 1: First Try

My first try actually was to add a 220V AC 80 mA fan, but at full speed it was too much powerful and noisy, so I decided to build the temperature controller, and to simplify the process I changed the fan into a 12V DC one.
Since Arduino lets you to read a temperature sensor and to drive a power transistor with pulse width modulation to vary fan speed, that will be the best solution to make a professional and more personalized device. Anyway I preferred to go for the odd way, to reach a good goal with the simpler circuit and avoid any programming.

Step 2: The Schematic

After a research on internet I found this interesting website: Fan Noise Solutions where I found this very useful page about thermal controller: Thermal Control. That circuit is based itself on some ideas published on 4QD-TEC. There you can read about the use of a LM393 dual comparator to generate a triangular wave and compare it with the signal from thermal probe, this generates a square wave which activates the PNP transistor and drives the fan with PWM. 
You can also build a simpler version of thermal controller with a MOSFET, as described on the heatsink guide page, but that will be more power consuming since the missing of the PWM function, and it will need a heat-sink to cool down the MOSFET.  
With the ready to copy schematic I begun to load components into DipTrace, a simple circuit program with autoroute function. I prefer DipTrace to Fritzing, although this last has a beautiful breadboard section, because it's more clean to draw traces in autoroute results.

Step 3: The Pcb Layout

In these pictures you'll see some steps of the PCB-LAYOUT process which led me to have the traces ready to print on the copper board. Read the notes on the images to better understand each action.
I attach here the Schematic and PCB layout files from DipTrace.

Step 4: The B.o.m.

Here is the bill of material:

a 12V DC fan (the bigger it is, the quieter it will be)
a 12V DC PSU
an LM393 dual comparator
a PNP transistor that will take the fan current
a little LED
a NTC Thermistor of 100k @25degC

some capacitors:
C1: 68nF Ceramic or Mylar
C2: 47uF 16V aluminium electrolytic
C3: 220uF 16V aluminium electrolytic 
C4: 100nF (0.1uF) Ceramic or Mylar 

some resistors:
R1: R2 R3 R4 R6 100kΩ
R5: 10kΩ
R7: 1kΩ
R8: 510Ω

OPTIONALS (read further for the way to avoid these):
a diode 1N4148
VR2: 10kΩ potentiometer
VR1: 47kΩ potentiometer

Step 5: The Prototype

Now I have schematic and components ready and I want to test the operation of the device before etching the board and solder the components.
Since something was wrong with this "Bells & Whistles" controller version (probably I missed something or I inverted some component in the mess on the breadboard, and I also forgot to photograph it) and I wasn't satisfied with the pots behavior, I decided to build the simpler version described on the thermal control site, the one with no range settings. That is the test on the first picture, in the second picture you can see I added a little green LED to show the device is ON.
Feel free to build the improved version with my pcb layout, I already fixed a polarity error of the C3 capacitor label (which didn't affected the breadboard prototype).

Step 6: The Downgrade

So I removed from the schematic the potentiometers and some other component. The board is more simple and compact now. The more space you leave between the components the simpler will be the autoroute process, so if you see that the software inserts too much jumpers, enlarge the board and give him more space.
You can download the pdf with schematic and PCB layout, and also the corresponding files from DipTrace.

Step 7: The Board

It's time to choose the copper board and cut it with exact dimensions. You can see my instructable about the PCB cutter tool I built. This time I used a simple cutter blade since the board is more simple to cut. 
After cut it wash the copper side and clean it with alcohol, this will make a better adherence to the toner.

Step 8: Ironing the Toner

To transfer the toner on the copper surface use an iron. Read this my other instructable step where I explained ironing tecnique in detail. Be sure that iron is very hot, and to push all around the edges of the pcb, where traces are very adjacent to the perimeter..

Step 9: Finishing If Needed

I obtained two boards, one is simply perfect (YEAH!) and the other needs a bit of corrections along some traces. I suggest to always make a couple of boards at the same time, since it takes a short time and it increases the chances to make better results.

Step 10: Etching

When the toner is transfered on the boards, and after cleaning away the paper with a teeth-brush, you can immerse them in the ferric chloride. Again, see the other instructable to read more about etching tecnique. Forgive me for these external notes, but I think that adding again those details would make a too boring instructable, I don't want that my electronic projects seem too much similar each other.

Step 11: Removing the Toner

Clean the board with a metal sponge, but scratch it with sweetness, treat it like a woman: you can be a bit rude, but you can't hurt her ;-)
As you see from the pictures the traces on the top board are very good, I'll use that board to make my device. I'll keep the other one as souvenir or for another identical controller.

Step 12: Drilling the Holes

Time to drill the holes. Again I use a tool I built,  fast, cheap, precise, a pleasure to use, it's my column drill. You can enlarge the few holes for bigger pins with an hand-drill for hobby-modeling.

Step 13: Soldering

I begun to solder the bigger components, and I added three solid pins at which connect the switch, so it will need no other fastening.
Then I added resistances, capacitors, also the LED which has long legs because I didn't decide where to direct the light.
I've chosen two flexible long wires to link the temperature sensor to the board, so I can put it near any device in the cabinet. I braided them so to have a single cable.

Step 14: The Thermal Probe

Here I solder the temperature probe at the far end of the wires. Remember to add a pair of heat-shrinking tubes to avoid short circuits before solder the thermistor pins, then cover everything except the probe head with another tube.

Step 15: Ready to Test!

The device is now ready, you can already test it, just plug a 12V PSU and turn it on. Now keep the probe head in your hands, and listen the fan speed increasing. Indeed it works pretty good, at 25°C is about half speed and silent enough, it speeds up with your hand temperature (30-32°C?) and it will go full speed at 40°C or above (try to warm up the probe with a lighter, but not burn it!)

Step 16: Attach in Position

I attached the controller near the fan on the pvc panel on the back side of the cabinet, exactly behind the console's grille (read next step to avoid my mistake).
The switch is behind the vertical brace I added to add more stability to the cabinet, and it could be easily found following that metal bar with the hand.
The LED goes into an hole so that it lights the cabinet compartment and it could be seen from the front.
Attach the pcb to the panel with a double face tape, avoid glues as they could corrode the copper traces.

Step 17: Fixing Mistakes

This is pretty funny! Installed the console and turned it on with fan already in operation, the three red leds of death appeared!! Fortunately I turned it off and I understood that the fan worked opposed with the console's ventilation, so it pushed air backward! I had to reverse the fan, so it helps the air to escape out of the cabinet (it enters from the big hole behind the amplifier). Everything worked good one time turned on again... it was apparent death! ;-)