LED.EX - UV LED exposure tool for circuit creation

Hello everybody!
As many of you helped me on making the circuit of this project I really want to thank you all! I could probably never make this on my own!
This project has been used during my graduation exams at high school just a few days ago (July 2nd, 2013) and teachers had no words for it!

Here's the preview of the final product:


This being said, what's LED.EX?
LED.EX is an UV exposure tool that allows creating circuits (PCB) by using UV rays. The difference between my project and many others you'll find on the web is that mine uses UV LEDs instead of UV neon lamps, and, the most important part of it, it has a control board that counts the exposure time and allows to store it for further usage!
It works with special copper boards that use a light mask to impress the circuit traces. Once exposed to UV it can be immersed into acid to cut out the extra copper; only the part of copper exposed to UV light will be cut out, leaving the circuit traces behind.

Next there is the electric schematic. It's very simple so many could build it on their own even on a breadboard.
This is the main control board, and the PCB picture. Please note the schematic is newer while the PCB is from an earlier version of the circuit where I was having two 7812 voltage regulators (U3 and U4) and 4 diodes (1N4007) for rising the voltage up to 13.4V. In the final project I took out these parts and short-circuited the input and output pins of them, so all the power is tied directly to the transistors.


This is the LED board and the PCB picture. The two connectors on the sides are meant for power supply. in the PCB, LEDs are placed in a regular 16x10 grid and each LED is 1cm far from all others. That's very high density and ensures that the UV light does not lack any portion of the board. The resistors used are SMD in the 0804 package.

...and the components list:

Attached to this page there is a zip file containing the MPLAB project for PIC16F876. The source code of the project is written in C and has to be compiled with Hi-Tech's C compiler (works fine both Pro and Lite version).

The circuits provided can work with a variety of DC power supplies, starting from 5V up to 30V (just make sure it can carry out the total load).
The main board has a 7805 voltage regulator that will power the PIC and the LCD display, and the 4 switches. SW1, SW2 and SW3 use the R1, R2 and R3 pull-up resistors, and they are needed to interact with the software on the PIC. SW4 is a saftey switch: UV rays are dangerous, is better not to be hit by them if you accidentally open the top panel.
The LED boards are powered through the 2 NPN transistors. The transistors I have chosen (2SC4793) are set to power the load (LEDs) on the collector, and they have high hfe (min 100, max 350) so they're perfect for driving even huge loads directly with the low current output of the PIC (about 25mA max).
The main board has been made to control 2 spearate LED boards. Each LED board uses in parallel many series of LEDs. You may put up to 9 LEDs in series with a maximum power supply of 30V. In the final project the power supply I have chosen is a 13.5V switching power supply (already made) and I have built 2 LED boards of 160 LEDs each (4 LEDs in series, 40 series for each board), with 68 Ohm resistor for each series (around 15mA max per series) for a total current of 600mA per board (8W). The transistors though have a potentiometer (RV2, RV3) that it's used to control the maximum brightness of the LEDs. Whatever the power supply you'll be using, and whatever the load is, make sure the LEDs are not too bright.
In the project the 2 LED boards I made have been positioned to extend the maximum circuit size to 20x16 cm. Also, the LEDs used in the final project are flat-top with 120° light angle (best light between 30° and 90°); and the exposing circuit is placed on a glass surface at 2cm height up of the LEDs. This ensures that UV light coming from the LED boards is totally uniform and it helps when the acid cuts out the extra copper because there won't be areas where the process takes longer than others. If you'll be making this project try not to use round-top LEDs because they have a center light beam with very restricted covering area, and this will produce a "dotted-blur" exposure.

About the software, I just finished it a few days ago and I'm unable at the time of writing to translate (most of it is Italian) and comment the code. Moreover, the software is very rudely implemented, not optimized at all. Still.. it works . I'm planning to rewrite it from scratch; if so I'll be updating you here.
A very short help guide: there are only 3 buttons used in the project: UP, DOWN, ENTER. I guess there is not much need to explain what each does... anyway, use UP/DOWN to select a menu item, increase/decrease values or to select a message answer (some interactive messages on the display have "S/N" at the end, which is a simple Yes/No dialog - the choice blinks) and then press Enter to confirm.

Before I leave you I'll post some more images of the project (even early development). Unfortunately, I don't have the right copper boards to show a demo of the project in action, but I'll soon make one!

An early stage development breadboard prototype.... it scares me everytime I look at it!


The PCB actually used in the final project during the testing.


And a closer look to how it's made. For your interest, the box contained 3 bottles of wine... they magically disappeared to make room for anything else


That's all folks! I hope you'll enjoy this work as much as I do. Should you have any questions about it, or if you make a porting of the software for other compilers/PICs, just contact me!

LICENSE
The whole work is released under GPL version 3 License. Make sure you read the file License.txt included in the archive or read it online here:
https://www.gnu.org/licenses/gpl.txt

DISCLAIMER

Hi T3STY,

Thank you for sharing your project with us. This is a great project, very neat and well explained !!!!!!!!!! I have one question though, how much time does it take to cook a PCB in your design? Is there some kind of a calculation for it?

Regards

Thank you for your compliments, I'm glad to know you like it!

So... the calibration, that's something I forgot to say.
To cook a PCB the time is different between different PCB manufacturers.. you know, it's the same story for any same product made by different manufacturers.

This is why the software has a Calibration function. In the main menu, scroll down once to "AVVIA CALIBRA." (that stands for "avvia calibrazione", which means "Start calibration"). This function will make UV exposure sessions of a specified time, for a specified number of executions. These settings can be set by going from the main menu to the third item "IMPOSTAZIONI", and then selecting "IMPOSTA CALIBR.". The first thing you'll be required to input is the exposition time for each step. Then the next thing is to set the number of steps to perform for the calibration. The defaults have been set to 12 steps of 10 seconds each. Unfortunately, I tested this on my own and figured out it's not enough time. Try using 10 steps of 15-20 seconds each.
Now, there's another thing: to calibrate the PCB total exposure time you'll also need a calibration circuit. Make a circuit with a few horizontal traces starting from the smallest you can draw up to something large enough. Then divide the traces for the number of steps you set in the settings before. When running the calibration cover all the board. Then uncover one part of it at the time (one for each step), and keep going like this until the calibration ends (you're required to press enter each time a step has finished before it continues with the next one). The result in all this process is that each step of the board is exposed for a different time; the first step will be exposed exactly for the step time (with default, 10s), then the second is exposed for twice that amount (20s), the third is exposed three times that amount (30s) and so on for each step. When you'll be taking out the board from acid you'll notice that some traces will be clearly defined; they're somewhere in the middle usually, while on the extremes of the board either copper is not cut away at all (too short exposition time) or all of it is gone (too long exposition time).

Once you can define the amount of time you need for a proper exposition, go to "IMPOSTAZIONI" then "IMPOSTA TIMER" and set your final exposition time. From now on, every time you select "AVVIA ESPOSIZ." (it means "start exposure") from the main menu it will run for the amount of time you set.

Right now I have finished working on a nice calibration sheet. I'll attach here the zip, containig 2 PDFs. The copper layer PDF is the one you have to print use on the board during the calibration steps. The silkscreen layer has to be printed on a transparent sheet and placed over the copper board after you take it out from the acid, so you can check the results. The sheets have been made for a 16x10 cm board (6.3x4 inch), this is a standard size for the board you can buy from the stores. The silkscreen layer has both inch and cm measures, to make it easier to understand for people from around the world. The board includes standard traces used in PCBs, starting from 0.006 inch (0.15mm) up to 0.1 inch (2.56mm).

I'm not sure I've explained very well how to proceed into this calibration process, this is why later (or maybe tomorrow) I'll be going through the calibration process and post a nicer guide with photos and the result.