- Unraveling the JPEG
- When Measles Arrives: Breaking Down the Anatomy of Containment
- The Rube Goldberg of Skateboarding: Matt Tomasello (YouTube, via Rusty’s Electric Dreams)
- Gareth’s Tips, Tools & Shop Tales a newsletter from Gareth Branwyn
- Project Phaedra is digitizing and transcribing the notebooks of the first women astronomers at Harvard, including Cecilia Payne-Gaposchkin who was the first woman Full Professor and department Chair at Harvard (via an excellent twitter thread by Robert McNees)
- PCB Fabrication with EDM
- Stranger Things Lego Set
- Firms That Promised High-Tech Ransomware Solutions Almost Always Just Pay the Hackers
- A Lego Turing Machine
- The Macintosh Garden: A treasure trove of Mac abandonware
At the 2018 Bay Area Maker Faire, our project Uncovering the Silicon showed off a number of simple and complex integrated circuits (with rather large feature size) under the microscope. We had a great time helping visitors look at the features and get a glimpse of what’s inside those black box integrated circuit packages. To take this to the next level for this year’s Maker Faire, we decided to try and close the loop; to take one simple integrated circuit and elucidate its workings well enough that visitors to our booth will be able to see every single component of the circuit, understand their function, and relate it to the macroscopic behavior of the chip. For this, we picked what turns out to be a rather obscure chip: the Fairchild μL914, which is a dual 2-input NOR gate. This chip belongs to the resistor–transistor logic (RTL) family.
Here’s what the chip looks like. It’s in a funny old “glob-top” can package with eight leads.
Here’s the pinout; there are two NOR gates in the chip, plus power and ground.
Ken Shirriff built a circuit with the chip to demonstrate its functionality. When we push either of the two buttons for one of the gates, that LED will turn off.
Here’s the schematic diagram, adapted from the original datasheet. If you look at the left side, if either of those inputs goes high, the transistor pulls the output low.
John McMaster decapped a few of the chips and sent us a die photo. He made a video about the process — no small feat. We’ll be bringing one of these bare chips and a microscope (equipped with both eyepieces and a camera) to Maker Faire.
For the macroscopic scale, we approached visualizing this circuit from a couple of angles: the physical structure of the chip, and the electronic structure of the circuit.
Eric Schlaepfer used the die photo to model the structure of the chip in CAD.
Simultaneously, Ken designed a printed circuit board version for use with discrete components that maintained the same structure as the IC.
Working from Eric’s CAD model, we built a single NPN transistor model from layers of colored acrylic. If you lift it up, and look through the transparent middle layers, you can tell that the emitter (red) is embedded into the top of the base (yellow) and does not go all the down way through it. (Transistors like these are planar: The emitter is above the base, and the base is above the collector.)
The top layer of this little model has labels for the collector, emitter and base. It is removable so that the layers of the model can be more easily inspected.
The model of the chip die includes a transparent cover representing the oxide layer, and that supports the metal layer with the wire bond pads on the edges.
One of the reasons that this particular chip is educational to look at is that there are a few unused components on the die. There are two unused transistors: one of them is unconnected, and the other is shorted. There are also several unused resistors (resistors are the dogbone shapes). The unconnected and unused components are easier to see, and provide a visual example that is useful for understanding what the connected components look like under the metal layer.
It is also fun to imagine what other circuits could have been made with different connections.
We glued most of the layers together, but left the top two layers removable so that it is easier to see the internal structure when the top is removed.
(Aside: we left out most of the epitaxial pocket material, because even though we used transparent acrylic to represent it, the layers of the components are much more visible without it present.)
There are cutouts in the oxide layer where the metal layer connects to the circuitry below.
One of the most noticeable things you see when you look at this type of IC under the microscope is the bond wires. We’ve used silver glitter hot glue sticks to represent them.
The glob of melted glue represents where the wire is bonded to the pad.
When you look straight down on the model with its glitter bond wires, it looks very similar to what you’ll see in the microscope.
To round things out for our acrylic model, we made a physical legend to make it easier to identify all of the parts of the model.
Once Ken got his PCBs back from our friends at OSHPark, he built it up with the same example circuit.
The PCBs turned out beautifully, and it’s great to see the familiar discrete packages on the enlarged circuit. Ken has published the PCB design on Github.
We hope to see you at Maker Faire this weekend!
Bonus: Ken laid out some hypothetical alternate metal layers to use the same die to create different chips.
When we first started roasting coffee, we used the air popcorn popper method. We learned about it from Sweet Maria’s, which in addition to selling green coffee beans, has a wealth of resources for home coffee roasters. We gradually refined our method, even making a DIY coffee bean cooler. We also tried out a lot of different types of beans, buying samplers from Sweet Maria’s and making notes on which flavor profiles we enjoyed. Eventually we outgrew the batch size limitations of the popper method and we purchased what is the gold standard of home coffee roasting: a Behmor 1600.
The Behmor can roast up to a pound of coffee at a time, and does so reliably without fuss. Newer models are programmable, but this one has a few preset roasting profiles. I use the default one pound setting and normally roast about 0.9 pounds of coffee at a time. The canister I use for coffee weighs about 0.2 pounds, so when I put it on the scale (it’s handy having a shipping scale nearby) I aim for 1.1 pounds. This smaller quantity roasts a little faster than a full pound would, and I have a wider time window to stop the roast when it gets to the stage I like.
The green coffee beans (which smell grassy, a bit like fresh hay) go in the roasting cage which gets put into the motor socket so that it can be rotated to toss the beans around for even roasting. After putting in the chaff catching tray, I start the roasting process.
The chaff is the papery membrane around the bean that comes off during the roasting process. Roasting creates quite a bit of smoke, and even though the Behmor has a smoke-suppression afterburner to reduce the amount of smoke, I prefer to roast outside.
Roasting takes about 20 minutes and the cooling cycle takes another 12 minutes. Because this is basically a toaster oven and fire hazard, it should be monitored during the roast. I take advantage of this half an hour in the sun to call my dad or catch up with friends. It’s an enforced break in my usual routine when I get to listen to the local birds and enjoy the changes in the sky through the seasons. When the roaster is done, I pull out the chaff tray and it’s quite a mess.
Most of the chaff is collected in the tray, but there’s some still mixed in with the beans, and it gets pretty much everywhere when you bring the roasting cage out. This is another good reason to roast outside. The chaff will just blow away in the breeze and joins the leaf litter below the shrubs that line our little parking lot.
I shake the roasting cage repeatedly until the amount of chaff dwindles, and then I can put the roasted beans into the canister for storage until I need need them.
There’s still chaff in the roaster in spite of the chaff catching tray. I blow it out of the roaster and sweep it out with a small clean basting brush.
Once the roaster is cleared of chaff I put it away for the next time.
The coffee loses a lot of its moisture during the roasting process, and reduces in weight by about 10% or so. It also increases in volume.
We have two to three people drinking coffee and use about two to three pounds of coffee a week. The flavor of roasted coffee starts to deteriorate about six or seven days after roasting. Since I’m roasting as needed two to three times a week, it never gets past about four days old.
As our coffee bean usage increased, we started buying our coffee 20 pounds at a time, but eventually realized that even that was seeming to be a little frequent. We had settled on the flavor profiles we enjoyed most, so we started purchasing 50 pound bags from Sweet Maria’s wholesale site, the Coffee Shrub. Green coffee has a long shelf life when stored well, so this means we don’t have to think about it very often. When you buy for several months at a time, you need to be confident that you will be happy with it. When we first got started, we didn’t know what we liked well enough to commit to purchasing at that scale, but we’re pretty set in our ways now.
Occasionally someone will really enjoy a cup of coffee I make for them and will say that I should start a coffee shop. It’s a well-intended sentiment, and I take it as a compliment. However, starting a coffee roasting business or coffee shop would take away many of the things that I love about coffee roasting. I only roast the kind of beans that we like, and I only roast as much as we need. And I get to use roasting as an excuse to take a break from my other responsibilities and enjoy being outside in our beautiful weather here.
I still enjoy trying other coffees. I love going to my favorite coffee shops and having someone else make me coffee. I love trying new coffee shops. And I get a lot of satisfaction from the coffee that I roast, grind, and brew myself.
If you’re interested in home roasting, I highly recommend Sweet Maria’s resources. In addition to working with farmers around the world to source beans equitably, they share their deep knowledge freely. They have articles, tutorials, and videos. They hold workshops at their warehouse in Oakland, and will be presenting at Maker Faire as they often have in the past.
Bleeptrack is a generative artist who works in a variety of media, but we are of course particularly fond of her plotter work with AxiDraw. We were privileged to see her work displayed at the Plotter People meetup in San Francisco. One particularly cool and generous thing she has done a few times is to create and then share tools to generate artwork.
She has used overflower not just for plots but also for fabric printing and she has produced some beautiful bags for sale from the fabrics.
The label on this floppy disk (which was used with a vintage digital camera to store pictures) was made by plotting onto label stock with an AxiDraw, with a drawing from the 35c3 circuit generator, which
creates fictional electric circuits with a message inbetween the lines. The design is in the style of the event design of 35c3 (Chaos Communications Congress). The project was built to enable all participants to easily create their own decorative circuits and ended up ‘highjacking’ the original event iconography. During the 4 days of 35c3, the circuit generator was used hundrets of times and resulting images were displayed all round the event halls, stages, stickers, appearal and banners.
You can read more about the fictional circuit generator on its project page.
She has used the generator to make stickers and connected the generator to a print-on-demand t-shirt shop (in Germany) so that you can have your own unique bugs on a shirt. Her beetlesbot twitter bot generates and posts a new beetle every 8 hours.
My favorite non-generative Bleeptrack plot uses the data from her vacuum as it mapped the space in her home. The robot motion is plotted within the floor plan of the house.
This time a medium sizes rhea egg with three six petal flowers. What I love about the EggBot is that I can get 3 equal flowers onto any sized egg.
After marking the egg with the EggBot, it gets painted and any decorations such as crystals, beading and figurines are added. She also incorporates 3D printed bases into her designs.
Donald Bell of Maker Update was recently on the Cool Tools podcasts, and gave the WaterColorBot a very nice shoutout. He also mentioned Turtle Toy, which is a tool for creating your own generative art using a minimalistic Turtle graphics API. We recently found Turtle Toy as well, and have been inspired by the art people are sharing there.
Previous Cool Tools posts:
- The art of transcribing old documents
- 3D Print your own Digi-Comp II
- WRL Technical Note TN-13, April 1, 1989: Characterization of Organic Illumination Systems (139 KB PDF)
- Examples of legal and illegal LEGO assemblies with explanations of why: Stressing The Elements (1 MB PDF)
- Ken Shirriff on Reverse-engineering a hybrid op amp module
- How to make a Super-Hydrophobic Labyrinth Game (YouTube)
- maker.js: A library for creating and sharing modular line drawings for CNC and laser cutters.
- Plotter People #2 is happening on March 13 in San Francisco
- 3D print your own traditional Marble Clock
- A crimp connector compendium (via hackaday)