follvalsch

follvalsch:

Project INFRA #4: GPN14

bronsen and I went to GPN14 a week ago and learned a few very valuable design-lessons.

1) The coils using ~10 loops of 2.5 mm² NYM wire don’t cut it. We can’t feed enough current through them to produce any serious magnetic field.

2) Thinner wire but also more loops results in way better magnetic performance.

3) Using pieces of threaded rod as ammo we did science (a bunch of measurements and some math) and determined there’s an optimum distance between coil-center and projectile-center paired with an optimal trigger-pulse duration to fire a single stage coilgun. It doesn’t seem to matter much how long the coil and/or projectile is.

4) Record muzzle velocity was 3.8 m/s (not really much)

5) We need better coils (more loops, way more loops)

We did some preliminary trials with a double-stage coilgun. Results weren’t really what you could call a break through, we’ve only gained around 0.4 m/s in muzzle velocity. Probably the hand-tuned timing is off.

Timing will be replaced by proper real-time calculations in the next iteration. As a preparation for better speed measurements, we’ve already soldered simple light barrier breakouts that come with the required resistors and can be simply hooked up to the Arduino board which we’re using currently for prototyping.

In other news, we’ve decided to name the project INFRA. Short for infuanfu rapid accelerator. There’s also some code on Github which is almost certainly no use to anyone at this point.

follvalsch
follvalsch:

Project INFRA #3
The borosilicate glass tubes arrived! So did the light barriers.
I quickly soldered a prototype on a piece of perfboard and hooked it up to my oscilloscope to measure the falling speed of anything I drop through the glass tube.
A rough calculation tells me I’m in the ball park and there should be enough accuracy to scale up to higher speeds which will be required for the actual coilgun setup.
Timing will later be done using an Atmel Atmega32U4 (probably) with options to go nuclear using a CPLD or maybe even an FPGA. No reason not to go insane ;-)After all, I’ve got those components still around form a previous project.
Updates will follow.

follvalsch:

Project INFRA #3

The borosilicate glass tubes arrived! So did the light barriers.

I quickly soldered a prototype on a piece of perfboard and hooked it up to my oscilloscope to measure the falling speed of anything I drop through the glass tube.

A rough calculation tells me I’m in the ball park and there should be enough accuracy to scale up to higher speeds which will be required for the actual coilgun setup.

Timing will later be done using an Atmel Atmega32U4 (probably) with options to go nuclear using a CPLD or maybe even an FPGA. No reason not to go insane ;-)
After all, I’ve got those components still around form a previous project.

Updates will follow.

follvalsch

follvalsch:

Project Fronius Repair #3

The first batch of replacement components arrived and I started replacing a few here and there.

In the process of doing that I removed one of the larger circuit boards and noticed huge black smoke stain on the back. Turns out that one of several APT5010JVR power mosfets died.

On Digikey those go for 36 EUR. On Ebay however you can get them for as cheap as 14 EUR. Well, you do pay with longer shipping time (3 to 4 weeks).

Time to wait again…

follvalsch

follvalsch:

Project INFRA #2

The second iteration of the coilgun project yielded in one awesome decision and one bad decision.

First the awesome decision: bronsen (who joined me in scaling up the coil gun) and I decided to get a proper high power capacitor. Specifically the Maxwell BCAP3000 that comes with a whopping 3000 F (note: no milli or micro here, proper Farads) capacitance.

It takes a “while” to charge it up (read: long minutes to few hours) depending on how much current you can provide.

However, the short circuit current of this beast is at around 10 kA thanks to a mere 0.28 mOhm ESR. Enough to use it for spot welding and melting of nails.

The bad decision we made was using 16 mm PVC tubing as a barrel.
The extra power we got from the huge capacitor was not enough to produce a magnetic field strong enough to get anywhere close the power we had with the small capacitor bank in the first iteration.

Also, with a capacitor of this size, we would not be able to simply shut down the magnetic field by waiting for the cap to discharge. It would simply take too long and produce way too much heat in the process.

So a means to actually switch such a high current was called for. The most powerful FET we found at Digikey was the IXYS IXTN600N04T2. It’s able to switch 40 V and up to 600 A. So we got a few :3

To measure the projectile speed and time the switching of the coil we decided to go for the bare bones solution using IR LEDs and suitable photo darlington detectors.

To attach the homebrew light barriers to the PVC tube, I quickly designed a clip-on holder in Inkscape and used the shackspace laser cutter to cut a prototype from birch plywood.

The light barrier turned out to be sub-par for the task. Especially the holes that had to be drilled into the tube were a bit tricky to manufacture. The most difficult part was sanding down the edges of the holes inside the PVC tube so that the projectile wont get stuck.

The next iteration will change two things:

  1. Replace the PVC tube with a borosilicate glas tube. It’s transparent so no holes have to be drilled.
  2. Replace the 16 mm sized tube with an 8 mm sized tube. This will result in a stronger magnetic field.
  3. Replace the homebrew light barrier with off the shelf 8 mm light barriers. No tricky alignment or mounting issues.
follvalsch

follvalsch:

Practical Data Projection

Easterhegg is over and we had some printouts with personal data left. Mostly lists of names and a few notes taken during the event.

To make sure no personal data ends up in the wrong hands we decided to take the most fun solution and employed the process of rapid oxidization paired with mechanical agitation of the carbon fallout to ensure proper information disposal.

follvalsch

follvalsch:

Project Fronius Repair #1

Just a day after I finished the iron flowerpot stand I ran into one of our shackspace-neighbors, Andi, who operates a huge metal workshop just downstairs of the hackerspace.

I asked him if he could teach me how to weld properly and he happily agreed. However, when I told him that I want to learn how to use the electrode welding machine we have at the space he was quick to tell me that he a) would be willing to teach me WIG/TIG welding and b) has no clue about electrode welding and really dislikes it.

After some back and forth talking about the cost of new WIG/TIG machines he remembered the Fronius MagicWave 2000 fuzzy that died a sudden heat death two years earlier. He offered to give the machine to shackspace for free so we could try and repair it.

I quickly agreed and went to work with fellow hacker derdritte to find out what was broken.

A first check resulted in a list of possibly and definitely broken parts, mostly high-power resistors.

A few clicks through the reichelt electronic store website resulted in shopping list of around 10 EUR for a first iteration to try fixing the machine. Way better than the 2000 EUR the official repair (without guarantee) would have cost.

Components were ordered and it was time to wait.

follvalsch

follvalsch:

Iron Flowerpot Stand

The balcony garden at kombinat23 needed to stack some large flower pots to plant more herbs and veggies.

So I went on an excursion to the local hardware store and bought some metal. The stand is made of a ring that holds the actual flowerpot and four legs. Everything is welded together using the trusty old electrode welding machine and lots of frustration (did I mention that I have no clue about welding?).

In the end I learned enough to get a half way decent weld most of the time.

follvalsch

follvalsch:

Project INFRA #1

I decided it’s time to propel small metal pieces through space. What better to use for this than a coilgun.

The principle is simple: Use an electro magnet (the coil) to attract a magnetic projectile and turn of the coil at the right moment so let the projectile travel along the barrel.

To get a strong enough magnetic field, a bank of capacitors (each 1 mF) was quickly soldered together.

The coil is simply hand-wound wire wrapped around a drinking straw which also acts as a barrel.

Turning of the magnetic field is relatively simply: the capacitor bank discharges so quickly that you can simply adjust the system by moving the projectile (a tiny 8 mm nail) closer to or further away from the coil.

Said nail could then be shot “across” the room. That is, it dropped to the floor around 3 meters into the room when shot from table height.

A few iterations of different coil sizes were done, but in the end there simply wasn’t enough power in the capacitor bank to get anywhere more… interesting ;)

follvalsch

follvalsch:

Muscle-powered USB Charger

At shackspace we had an old manual sewing machine table sans machine standing around for years. I apparently wasn’t the first one to get the idea that “one could build a muscle powered charger” out of this, however I simply got to work on it. And lo and behold: it was done in one afternoon.

I got a bike dynamo from the bike shop across the street from shackspace. This was actually the most expensive part since it was the only thing I bought. Everything else was already there at the space.

After some prototyping I built an adapter to fit the dynamo to the sewing machine’s flywheel using some spare wood from the workshop.

The electronics is radically bare bones: no PCB used, everything is soldered together using the component’s leads and made more durable using generous globs of hot glue.

The 6 V (nominal max) AC output of the dynamo was first rectified and then buffered in a bank of 1mF capacitors. A 7805 voltage regulator without any external components is then used to convert this to a 5 V level used for USB.

Theoretically this piece of leg-muscle terrorizing equipment should provide around 500 mA of power. However, actually charging your cellphone is quite a challenge ;-) Let’s say it’s the most work-intensive 1% you’ve ever charged your phone.