At the Robot & Makers fair

HaSGO at Robot & Makers

This weekend we are showcasing EarthBeat at the Robot & Makers fair in Milano, Italy. For everyone that that saw EarthBeat at the fair, a more complete description can be found in our wiki.

Python code and schematics will be available on GitHub soon!

 

Earthbeat v2 Part II

We already explained about Earthbeat’s future, but a mystery still stand: what external sensor are we going to use? Fear not, for I’ll tell you soon enough!

Let’s talk about CD-ROMs

The CD-ROM drive is a sophisticated piece of electronics, mechanics and optics. The tracks on a CD-ROM are just \(0.6\ \mu{}m\) wide and \(1.6\ \mu{}m\) apart, so the optics must compensate for any eccentricity and imperfections in the CD-ROM it reads from. Using small but fast actuators, it can move the lens to match the track position quickly and accurately.

The lens’ focus needs to be constantly adjusted, too. This is the task of another small actuator that moves the main lens towards or away from the disk. The exact movement is calculated using the error signal originating from the PDIC (Photodiode Integrated Circuit). It reads the laser bouncing back from the CD-ROM surface and converts it to an electrical signal. There are two major ways of measuring the error signal: the astigmatic and single-Foucault method.

Astigmatic or 4-detector method

This method uses an astigmatic lens positioned before the detector. The detector itself composed of four square photodiodes forming a bigger square.

Astigmatic detector method

Astigmatic detector method (taken from Sorin G. Stan: The CD-ROM Drive: A Brief System Description)

When everything is in focus, the laser beam is round and hitting all the photodiodes with the same intensity. When the CD-ROM moves away from the lens, the beam deforms into an ellipse along one diagonal, if the focus is too far, and the other one if the focus is too close. If we label the four photodiodes, we can define the error signal as: $$V_o = A + C – (B + D)$$

Single-Foucault method

A “knife edge” is inserted along the optical path, obscuring half of the laser beam. An out-of-focus situation will generate a mismatch between the light intensities perceived by two detectors. We can define the focus error signal to be: $$V_o = B – A$$

Single-Foucault detection method

Single-Foucault detection method (taken from Sorin G. Stan: The CD-ROM Drive: A Brief System Description)

The idea

So, we have a signal that is proportional to the “displacement” of the CD-ROM if the lens are stationary. This type of sensor is:

  • Nearly instantaneous
  • Fully analog: no need to fiddle with ADC’s, at least not in the feedback loop
  • μm-precise: originally used to measure displacement on micrometer scale

So, I hope you’re all thinking: “That would be perfect to measure the displacement of the actuator lever in Earthbeat!”. So we examined the optical assembly of some broken drives we had laying around and, after destroying one “for science”, we found out that it’s fairly hard to find any information (or datasheet) for the PDIC. It was time for some reverse engineering.

We printed a small breakout board and started reversing. After a lot of searching, we found out the model and datasheet of our PDIC: Panasonic HUL7208. The happiness of discovery wore quickly away as the module heated up and burned. We interpreted the (scarce) datasheet we found wrongly and applied \(5\ V\) to the laser diode, where the absolute maximum was \(2.4\ V\)!

We’ll be back next week with a new optical module, more information, and hopefully less burned components!

 

Test etchings and latest news

Bromograph testing

Last Sunday we tested the big bromograph we repaired not so long ago. We did some exposure tests and noticed that it took more than ten minutes for the boards to develop properly! We were more happy than surprised though, as slow exposure is good as you don’t have to be as precise with timings.

open bromograph

Come get a free tan! Or melanoma, I forgot which one was it. Anyway, don’t stare directly into the lamps, kids!

Our bromograph

Our bromograph, happily creating UV rays.

As we submerged the boards in a sodium hydroxide solution, we had some bad news: the solution we prepared beforehand has gone “flat”! The sodium hydroxide bonded with other ions, probably forming sodium bicarbonate, as there was a white precipitate at the bottom of the bottle. The culprit was probably the water we used, not distilled but just tap water. Lesson learned, we’ll use distilled water next time.

We quickly did another batch and developed the boards. Our first try was pretty bad (isn’t it always?), but the second one was spot on!

boards etching

Our test board submerged in ferric chloride. That stuff is nasty!

 

EarthBeat v2

Some time ago we started thinking about building a better EarthBeat based on the input we received at the Maker Faire Rome. The first thing we wanted to change was the damping mechanism. Then, some geologists we met at the faire suggested we could do an active measurement instead. Let’s see what these things are:

Damping

In our first prototype, damping was done with rubber bands. This is a big source of trouble as their behaviour is difficult to model, and their properties change over time. Remember that old rubber band you forgot about for months, and it just disintegrated as soon as you touched it? You don’t want that to happen to a precise sensor as a bromograph! Not all is lost, though, because this problem can be solved (read: removed altogether) by using an active measurement system.

Active measurements

Right now, we estimate seismic activity be measuring the displacement of an inertial mass, i.e. how much does the disk head actuator move. This approach has some issues with non-linearity and repeatability of the measurements. A better way to measure the “displacement” would be to have another sensor, capable of precisely measuring short distances, and a feedback loop controlling the disk head actuator arm (in the way it was originally intended to behave) to keep the mass still. In this way, the non-linearities are a non-issue because the mass does not move appreciably. Since damping is electronic, no rubber bands are needed, and measurements are more repeatable

Key component

So, the key component will be the “sensor capable of precisely measure short distances”, and that was where we got a struck of genius. We found a part that would fit our needs perfectly! Moreover, it is common and readily available from obsolete electronics. What is it you say? We’re gonna tell you that next week, so stay tuned! :)

 

RadioAmatore Fair

At the Fair

This weekend we were invited by the PNLUG to show our EarthBeat at the LinuxArena 2 booth in the RadioAmatore fair in Pordenone. We gladly accepted, of course, and did a last check before loading everything into Aljaž’s car. After an hour drive we arrived at the fair and quickly set everything up: people were already coming in and asking questions!

EarthBeat at RadioAmatore fair

Our mini-booth at the RadioAmatore 2 fair in Pordenone.

The public was quite different form the Maker Faire Rome: a lot of people were here just to buy stuff. A typical exchange was:

“So, you guys sell Arduinos?”
“Nope, sorry.”
“Where can I find them?”
“In the store over there!”
“Thanks.”

But some people were genuinely interested: we met with a teacher who wanted to do a similar seismometer with his students, and a guy that was building a weather station and needed a little help. We invited both to visit us on our weekend openings.

Of course, we weren’t in the booth alone! To our left was a hackerspace named Crunchlab showing off their quadcopters. They are opening to the public on January, so best of luck to them! To our right there was the ICTP 3D printing lab. We met them some time ago and asked them some info on 3D printers. They’ll probably print us the parts for the RepRap we intend to build.

GOWeb at E-center

Over the weekend we also received an interesting email from e-hiša, which has now changed name to e-center: they are organizing an introduction to HTML5, CSS3 and jQuery for beginners called GOWeb. Unfortunately, the course will be held in slovenian, so if you don’t know that language it will be more difficult to follow.

 

Server solved, bromograph building

We finally solved the server issues: it looks like the linux kernel doesn’t support the DVD drive we used! After exchanging it with a supported one, the install went on without any problems.

We worked on the bromograph and there were awful smells as Aljaž cut the scanner’s plastic enclosure. We fitted the LCD display, encoder and reed switch to the enclosure, so all we need to do now is finish the code.

There are new pictures of EarthBeat at the Maker Faire Rome! Come check them out!

 

Back from Maker Faire Rome

Maker Faire Rome: the Weekend

EarthBeat awarded!

We are so proud.

While we were amazed and excited by the people’s reaction to EarthBeat on Friday, we were absolutely overwhelmed by it during the weekend. In the first hours of Saturday we already gave away all of the project flyers and most of the business cards. Everyone was staring at each other, wide-eyed, but with a big smile as we realized that our project was going to get some attention. Later, we managed to print some more flyers, thanks to the help of our booth neighbour (thanks again, Pier!), but they disappeared quickly, too!

Hacking: the other side (of the monitor)

You can see the LCD display here.

HACK IN PROGRESS

Aljaž was trying to use a touchscreen LCD to display the waveform.


signs

Some people were more interested at the “hack in progress” sign…

Everyone was busy explaining EarthBeat’s details to anyone interested, we even took turns for lunch, some people eating and the others talking (on a side note, the ones eating attracted a lot of people by wolfing slices right there in the booth, and yes, phrases like oh look, they invented pizza! and are you 3D printing pizza now, tee hee were heard).

Simone and Aljaž explaining

The fish was THIS big! I swear! No, we were really talking about something else…

Maker Faire Rome is the perfect place to meet makers, so Aljaž and Arturo were walking around to hunt for interesting projects and new connections. There was a group of people launching rockets, radiation detection using PIN photodiodes, lots of 3D printers and robots.

Pneumatic robot

This thing was huge. Like three meter huge.

Restart Project

Not very far from us there was a booth titled Restart Project, with a white wrench on black logo. We met Ugo, one of the founders, who told us about their work and the events known as Restart Parties. There, people bring all their broken stuff and together, they try to fix everything, and learn how to do it from others. This is a brillant idea, and we were thinking of doing something similar for quite a while!

Feedback

We met some geophysicists that worked at the National Institute for Geophysics and Volcanology, who were very interested in our project and shared tips and methods on how to improve on our sensor. One of them invited us to their workplace, and so we even managed to take a look at the National Seismic Observatory control room!

Intel Galileo

We were one of the lucky makers that got a sample of Intel’s newest creation, their Galileo board. The CEO in person gave it to us after its presentation! The Galileo has a Pentium processor that runs Yocto Linux, but it can be programmed directly from the Arduino IDE, and has a shield-compatible pinout.

Lessons learned

On Thursday we had a lessons learned meeting where we discussed everything that we did good or could do better, and some good food from the famous “da Gianni” restaurant helped to keep us focused until late in the night.

 

Maker Faire Rome

Maker Faire Rome started today, and the project we worked for the past couple of months, Earthbeat, is finally ready.

Last days

There was a lot of unfortunate events in the last days before our departure. Sunday, when the boards were finished, we tested them and no board worked properly. There was a little panicking, but fortunately we managed to get them working on Monday, and we also built two spares, just to be sure. Then we had to edit the EarthBeat poster and the flyers and print them. This was all done on Tuesday, when we also tested a computer to bring to Rome with us, but we found out it was too slow for our needs. Wednesday was the day to start packing and put Earthbeat in the car, and this morning we started our trip to Rome.

Our team stuffed in the car, ready for departure.

Our team stuffed in the car, ready for departure.

The journey was uneventful, and we arrived at the Maker Faire just in time to set up the sensor, do some tests and meet new makers!

Maker Faire Rome

This is our booth! The one missing is Arturo, taking the photo.

This is our booth! The one missing is Arturo, taking the photo.

At closing time the faire still looked like an anthill! There were people everywhere, preparing booths, testing robots, printing 3D objects, or just talking with each other. After setting up the sensor relatively quickly, we divided into two groups and one started looking around, while the other was at the booth. There are a lot of interesting things around us, from a wi-fi controlled balloon to a musical instrument called Guitoulele. A lot of makers stopped at our booth, too, and we explained them how does the sensor work. Surprisingly (in a pleasant way!) quite some of them were interested, and asked more questions.

Simone explaining how our sensor works.

Simone explaining how our sensor works.

Weekend plans

The Maker Faire will we open all weekend from 10 AM to 7 PM, but tomorrow only schools are allowed. If you happen to come by, meet us in booth E19!

 

Hot caddies!

Our team was waiting for the EarthBeat components to arrive, so we all did other stuff except for Aljaž and Philip, who were soldering the vias in our freshly etched circuit boards. In the meantime, Riccardo was working on the EarthBeat sensor, adding weight and a better dampening system for the x and y sensors, which increased our sensibility!

Aljaž and Arturo were thinking about the server, which lacked some nice SCSI drives. We have some, but their caddies were incompatible! So, with a little hacking (literally) they convinced some blank caddies to accept the disks. Then, the next step was to test them. One had a troublesome-looking “Bad” writing on it. Maybe some defective sectors? Anyway, the two 73 GB drives (one of them was the “Bad” one) tested OK, but the twin 146GB ones are still untested.
The thing that lowered their spirits happened when they took the drives out: scalding hot! It looks like the plastic blank caddies were unable to convey enough heat away from the disks. We’ll have to buy real ones!

We plan on hacking some more as soon as the EarthBeat components arrive, somewhere around Wednesday.

 

EarthBeat tuning and server servicing

I think this is the busiest week we ever had at HaSGO. Today we worked on EarthBeat and our server rack, adding our main development server.

The server

Our main server at HaSGO is a Fujitsu Siemens TX150. It has a Pentium 4 processor and 4 GB RAM. The case has support for two ATA hard disks and four SCSI. We had some SCSI disks laying around, but the caddies were for a HP! After much thought (and looking up how much does a caddy cost) we decided to use the plastic placeholders as makeshift HD trays. Temperature is going to be a problem probably, but we resolved to test that before even thinking about buying new caddies.
So, we have two 250GB ATA with two 150GB and two 73 GB SCSI. Our server is looking good, and you can almost see the spark in Arturo’s eyes!

EarthBeat

When Riccardo brought the sensor back, we started poking and prodding it to see how it responds. The z axis is picking up some signals from the x-y plane, and that’s not good. We’ll probably revert to the old inertial mass, fixed with some string.
The x and y axes are still lacking an adequate mass, but the signal looks good! We’ll conduct more thorough tests and try come calibration procedures on Sunday.
Here you can see a detail of the shield we use to shape the signal, AnalogBoard:

20130920-001423.jpg

20130920-001605.jpg