Wednesday, January 15, 2014

Super Metroid SNES Cart Battery Swap (2014)

A short while ago, a great friend of mine purchased a used Super Metroid SNES cart.  However, he quickly realized the SRAM battery was shot, which meant the cartridge could no longer save his game.  I offered to see if I could swap the battery for a new one and give back the cart is functionality.

The first obstacle, I knew, would be the darn tamperproof gamebit heads used by Nintendo to keep their cartridges shut.  I had a piece of scrap aluminum around, and on the second attempt managed to file a passable makeshift gamebit driver.  (For my first attempt, I made the miniature spokes were too thin; because aluminum is so soft, the tool quickly bent out of shape.)

Makeshift hand-filed gamebit driver.
Fortunately, the cartridge screws are only very lightly tightened, and my improvised tool made short work of them.

One down.
One screw out of the cartridge.
Unfortunately, the tool lightly scratched one of the screwholes.
Then came time to have a look at the goodies!

Freshly-opened cart.
The culprit, a dead CR2032.
Look at that battery.  As Shijima said in Ninja Scroll: a strange technique!
The CR2032 is held in place using a manufacturing technique I'd never seen before.  After some research, I realized that you can purchase so-called tabbed CR2032 batteries which come with such tabs preassembled.  The tabs look like they're cold-welded to the battery casing in two spots; I presume this is done before the battery itself is even assembled, because it looks as though the welds on the bottom were made from within the battery.  The spots form a tight bond between the battery casing and the tabs, and it took quite some effort to separate them.

At this point, seeing as how it looked as though I would have to irreversibly deform the tabs when removing the battery, I decided to turn to the internets to see what others thought the best repair method was.  I found a few blogs pointing to the following page:

http://motherboard.vice.com/en_ca/blog/pictures-how-to-replace-an-snes-cartridge-save-game-battery

I thought: tape is an obvious solution, but surely a permanent battery holder would be a much better one!  And I happened to have one lying around from a discarded PC.  I removed the old battery and tabs, and tried the battery holder for fit.

Removed tabs.
Trying the CR2032 battery for fit.
Unfortunately, though it was elegant, this was not meant to work.  I quickly discovered there is not enough space inside the cartridge case:
FAIL!  Not enough space inside the cart.
Luckily I hadn't soldered the holder in place yet.  I simply removed it, soldered the tabs back in and taped the battery as suggested.  The new battery holds quite fast, and I preloaded the tabs with tension so the battery mostly holds in place even without the tape.

Tabs back in.
Battery in place.
I then closed everything up and tested the fix.  Victory!
Closing up the case.  (The soldering job looks terrible in this picture; it doesn't look nearly as bad in real life.)
Testing out the fixed cart with my furry assistant.
The galaxy is at peace!
Edit: seems there was possibly more to this cart's problems than the battery, since it just suffered another saved game reset.  To be continued...

Second edit: many thanks to reader Kincl, who pointed out that Super Metroid is apparently an ornery cart whose SRAM/battery circuit is difficult to restore to health!  It appears this is a somewhat common case:

http://www.racketboy.com/forum/viewtopic.php?f=17&t=44493&sid=02d23a1f277a5b503e4a9004b029cbe4&start=0

For what it's worth, this specific cart has held up for just about a week now.  Time will tell how durable the repair ultimately was.

Sunday, December 22, 2013

Giant Origami Crane Hallowe'en Costume (2013)

For Hallowe'en this year, my girlfriend gave me an awesome idea for a costume: to dress up as a gigantic origami crane.  I immediately fell in love with the idea and began thinking on how I could execute.

We discussed the idea for a bit and initially thought of using fancy printed paper.  However, there were two problems with this approach: large printed sheets are hard to source, and when you find them, they cost about as much as a small island.  Thus, we went to plan B and settled on making a white crane, thereafter to stamp it with a traditional-looking pattern.

I began by purchasing a few sheets of white poster board at my local dollar store, defending them from the cashier who was on his way to folding and creasing them into a small bag.  I then taped a few sheets together, cut the whole thing into a square, and folded a standard crane:



A standard, if slightly large, origami crane.
I made a much smaller model out of regular origami paper and cut it up to figure out how I was going to fit the large-size model around myself.  Then, I cut out the center of the huge crane and started working on creating a coffee stick armature to stiffen the model:

Trying to figure out how to make this thing rigid when there's a gaping hole in the middle.

Why, hello!
Adding bracing on the underside.  You can also see a brace at the base of the wings.
There is a long wooden dowel in the tail.
Another dowel braces the neck.
Top view; you can't see any of the armature.
We then selected a pattern and cut it out.


My girlfriend sculpted it out of a potato as a test:




Finished potato stamp.
I then made a second version of the stamp out of EVA foam:

The two stamps.
We promptly proceeded... not to use the stamps at all due to time constraints.  I did get the wings attached to my wrists using invisible wire though:


Flapping the wings.
A small coffee stick piece was painted white and used as a fishing line anchor inside the wing.
And that, as they say, is that!  Here is a timelapse of the whole process, from raw poster board to finished costume.

Sunday, July 14, 2013

Experiments in Dimming EL Wire using PWM (2013)

Recently, a friend of mine got into electronics.  As a first project, he wants to build Tron-themed shoes for his son.  He chose to use EL wire as a lighting accent, and would ideally like to be able to modulate the brightness of the wire using software.

However, EL wire must be powered using relatively high-voltage alternating current, and a dedicated driver circuit is usually employed to convert DC into the required format.  Because these drivers usually include some kind of oscillator/resonator to alternate the current, I was worried that using a transistor to limit the input current might screw up the driver's ability to stabilize, and initially though it might be better to throttle the oscillator output using a TRIAC or some such.  However, I realized that without special care, turning the TRIAC off would be akin to running the driver without a load, letting the oscillator run amok, which by all accounts is not a good idea either.

So... we turned to our good friend the Internet, and of course somebody else has already dived into this topic to much greater depth than I could ever hope to achieve.  Thank you ch00f!  His conclusions were that current-limiting the input stage of the resonator was a valid approach.  I treat this as great news for humanity at large because it means we can simply apply what we know about using transistors to control loads, treat the driver-and-EL-wire system as a generic load, and go on with our lives.

Here is the result:


I had some spare 2N2222s around so I adapted the idea to use those:

Manual dimming circuit.
PWM dimming circuit, with manual brightness control.
This circuit was intended for experimentation; in a finished circuit, you would probably choose some total fixed value for the sum resistance going into the transistor base and leave it at that.  You might also want to add a capacitor either in parallel with the driver or between the base of the transistor and ground to smooth out things a little, but I didn't need one in my tests.

Here are a few pictures from along the way:
The driver outside its enclosure.
Initial connection of the driver to an external PSU.
Controlling the driver circuit using a transistor.
Close-up of the transistor wiring.  (The Arduino RBBB is not used here.)
Using a potentiometer to dim the circuit.
Manually dimming the circuit.  What the heck - Picasa created this animated GIF on its own.  I never asked for it, it just showed up.  Insanity!
Final setup with the Arduino RBBB, messy desk, breakfast juice and all.  To supply a test PWM signal I simply used the circuit from my quiz game buzzers.

Wednesday, April 10, 2013

First Fold of Satoshi Kamiya's Yellow Bird (Chocobo)

As kid I always loved origami; my parents still have bags full of my models in their basement.  Every few years, I usually read up on where time has taken the art.  This time, I was quite surprised - it had been around ten years since I'd last looked into origami, and man oh man have things changed.  People like Robert J. Lang and Satoshi Kamiya have taken model complexity to new heights.

Galvanized, I looked up a few instruction patterns and settled on the "relatively simple" task of folding Satoshi Kamiya's "Yellow Bird".  (I did not fold it directly from the crease pattern, though that looks like an interesting challenge for next time.)  I'm far from completely satisfied with the way the model turned out, especially the toes, but hey it could be worse for somebody who hasn't folded anything other than receipts from the grocery store in ten years.

(The model is on a Lego "frame" so I could take pictures, because I did not wish to glue the underside of the model together to allow it to stand up.)



Yes, those are suppoed to be reverse inside folds on the back of the neck, not mountain folds.  I noticed a long time afterwards and haven't fixed it.






Working on the open sinks for the toes.



Before forming the toes.