Redraw the echo-hello-world board, add (at least) a button and LED (with current-limiting resistor) check the design rules, make it, and test it.

I made two boards for this assignment. The first one was pretty much identical to one I found in the Academy archive. The second board, a little different process and it was a more custom layout.

Before beginning I needed to check that I had all the parts I needed for both these boards.

---> ATTiny44
---> 2 x 10k 1206 resistor
---> 100 ohm (blue) or 499 ohm (orange)
---> 1uf capacitor
---> smd button
---> 1206 LED
---> 20k crystal oscillator ---> 6 program header
---> 6 pin terminal


The first board I made for this assignment followed closely the version created by Loes Bogers(Fab Academy '2015)

I was very proud of the look of this board. Cleanest yet. But... I messed up the resistors. I had to desolder and change them for the right ones. In doing this, I burnt some traces and had to get creative using a 0 ohm jumper. This is what the final board looked liked.

Not only did I accidentally put the wrong resistors on, I put a 100 ohm resistor for the LED, thinking it was a blue LED. The LED is an orange one and actually requires 150 ohm or higher resistor.

---> choosing an LED & resistor

Ohm's Law: V = I x R /// V / I = R

Blue LED - SMD 1206
---> Part number: 160-1889-1-ND
---> ATTiny pin = 5v
---> V needed for LED = 3.3v
---> V drop = 1.7
---> 1.7 / 0.02a = 85 ohm

The board is sending 5v out of PA7 / pin 6 of the ATTiny that the LED is connected to. The LED needs 3.3v. We need to loose 1.7v at 20ma to protect the LED. This means we need a minimum of an 85ohm resistor coming from the pin of the ATTiny44 to the LED. We can round up to use a 100 ohm resistor. In this first board, I accidentally used an orange LED. The requirements for the orange LED are as follows:

Orange LED - SMD 1206
---> Part number: 160-1403-1-ND
---> ATTiny pin = 5v
---> V needed for LED = 2v
---> V drop = 3
---> 3 / 0.02a = 150ohm

My board only has an 100ohm resistor for the orange LED that was put on it by mistake. It works fine but the LED will eventually wear down. It should be changed to a higher resitance or add another 100 ohm resistor.

---> making a schematic in Eagle

Working with Eagle has been a lot of fun. I don't know if I will ever need to use more than just a fraction of what it can actually do. I have found that there are only a few basic features that I need to design the boards I have been making.

In the schematic view here are the tools I have been using most:

---> Adding components - use this to add the elements that you need from the Fab Lab Eagle Library. I learned how to do that by following the Fab Academy tutorialHERE.
---> Naming parts - in the schematic view you need to name the parts that you add. If not, you will have generic names that are impossible to keep track of. This step is done once they have been labeled with the label tool.
---> Adding values - it is also important to lable the values of your components. Especially resistors and capacitors.
---> Adding nets - the net tool is bsically how you wire your board. You add nets between the components or between the component and your label. This is how you know how to connect everything once you get to the board view---> Labeling parts - the label tool is one of the most important. You can identify what the different pins are and where you nets are connected to. You need to name your labels once they are done.

The finished schematic for this board looked like this:

---> designing a board in Eagle

Designing a board in Eagle is kind of like working on a puzzle. There are limitations to what you can do and you have to work around them. Like in the schematic view, there were only a handful of tools I actually used for this project.

---> Rotate - you end up rotating your parts to find best way to design your traces.
---> Drawing traces - the trace tool is at the heart of this process. I tried autotracing and hated it. I prefer manually tracing them.
---> Ripup - the ripup tool is essentially an eraser for your traces. If you need to start your board over you can type RIPUP ; in the command bar.
---> Vias - I never used vias but if you use them it certainly can help your layout.
---> Holes - you can drill holes in your board. I use this later on in Fab Academy to mark where I want to make to pass wires in my final project.
---> Ratsnest - the ratsnest tool is useful to clean up your board as you move items around. It will direct you to the closest GND or VCC and help you visualize when your parts are properly connected using airwires.

When using the traces tool you have options to help control how it behaves. I would often change the way the traces behaves by adjust the angle that it prefers. By far, the most important element here is to set your trace width. I found 0.016" to be the best. It matches size of the milling bits I use. When setting the other design rules by selecting the DRC tool in the Tools menu, I set anything that has to do with basic spacing of the traces to 16 mil. This seemed to work out nicely.

The last step is to export the board as an image. As was done in the Electronics Production week, I selected only the top traces in the layers menu, then exported it as an image with the following settings:

---> PNG
---> monochrome
---> 500 dpi
---> full area


When working on my first board, I tried different iterations. Eventually I just decided to play it safe.

With my board exported, I imported it into Illustrator to convert it to an SVG for the Carvey. Unfortunately, mods is not setup for the Carvey. It seems there is a weird glitch with Eagle and Mac. Marc in my group also had this issue. He discovered that by dividing your scale by 2 it will scale to the right size. Before this, I was always manually doing it. Now it was easier to simply go to the transform pane and add /2 to the width or height making sure the ratio is locked.

This is the difference between the Eagle output on the left and it scaled. I have not had any issues fitting components onto the board.

Another thing I learned to keep an eye out for is making sure that my traces do not wrap around pad. This would innevitably be a detail that Carvey and a 0.016" bit could not handle. The next two images illustrate the monochrome drawing and the Easel preview. Clearly the pad and the trace are fused together.

To avoid this, I simply made sure that my traces to sharper angles around pads and other features. while also giving as much space as possible. Here is the same part of the circuit with the trace adjusted.

The final drawing looked like this. I made it as simple as possible.

Carvey's software Easel is actually quite accurate at detecting problem areas in its preview. I always checked my drawing with the final tool path to make sure all the pads and traces have a distinct separation.

The milling process was straight forward. However, I continue to have this problem with the depth setting on the Carvey. I had this board set to cut a depth of 0.1 mm but it turns out the depth per pass was actually 0.5mm which is too deep. I broke a bit shortly after this board. It looks good. Though, you can see it is a deep cut. That being said, I only learned later that it is better to work in inches when using Easel. I also learned how to shim the smart clamp with paper to trick the Carvey.


The process for making this second board was different. I wanted to design one from scratch without referencing another project directly. For this board, I used the machine at ├ęchoFab with Mods. The goal of this design was simply to try and have the serial pins heading out the front of the board as opposed to the side. I later realized that I made it so that the serial cable would have to be plugged in upside down to avoid shorting. Another flaw was that 5v and GND are too close to the button.

I selected 1/64 bit traces as my process. We used a vBit with this setting and it seems to work properly as long as you don't carve too deep. With a v-bit, the deeper you carve the thinner your traces become. I chose to do 5 passes in the Mods in the prcoess setting.

I used Universal Gcode Sender and sent the job to the Mill.

One of the things I regret doing is using Flux. This wasn't the cleanest milling job. I should have cleaned the board a bit too. I think the flux attached to some of the thinner debris betweent the traces and burnt easily. This board is not looking so good but I kept at it.

I'm not going to brag about this one. I don't have high expectations but it passed all of my continuity tests. I burnt off the trace to the LED and could not fix it so I used a through hole LED and resistor. I look forward to the smoke test on this one!