INSTANT GRATIFICATION: Video of a hacked Santa in action! (same video - MPEG format)
(please do not link directly to video)

The problem is that I hadn't been able to find many interesting things to do with it. I created a little display toy (persistence of vision) with some full color (RGB) LEDs. That was about it. But last month, an interesting announcement hit my mailbox. It was a call for entries [Word Document] for a local non-conformist art show in Tulsa!

I've got to do something. But what can I do? LEDs? No, too boring. So I went browsing the isles of Wal*Mart and see if there were any good hackable items to be had there. Sure enough, there were a few candidates. But one item out of the entire store screamed "hack me" more than anything else. The $49.84 animatronic Santa Claus.

When I got home with my very own Santa, the first thing I did was try to get a peek at the wiring which controls the speaker and the motors. I removed the staples that connected the clothing to Santa's boots. I discovered that all of the electrical wiring goes from Santa's left boot, up his leg, and into the body. Looked easy enough to trace... and all the smarts appeared to be in Santa's left shoe. Exactly the same place where all the external user controls are located.

I noticed a curious thing, though, when I was undoing the staples on his left shoe. Hanging out the side of the shoe were two green wires connected to a variable resistor (covered with a black rubber protective cover). What was its purpose? What did it do? It didn't appear to be something that they wanted consumers to touch. I fiddled around with it, but I didn't notice any changes in its behavior. I'd have to figure that small mystery out later.

The shoe is held onto the base with four small screws. With Santa placed on his back, it was a simple matter of undoing the screws and lifting up the shoe to look at the PCB hidden below it. Each set of wires going up into Santa connected to the PCB via a mated connector. It wasn't hard to determine the purpose of the different wires.

GREEN/GREEN pair to potentiometer: It seems that this variable resistor controls the speed of Santa's side-to-side "dancing". Turn it low, and you'll get a performance worthy of a nursing home. Turn it high, and it'll be jerky and could topple Santa over given the right conditions. Seems to me that the default is good enough unless you're looking for something special. Effective voltage range to the motor (which controls the speed) seems to go from 5 to 12 volts.

ORANGE/YELLOW pair: This is a three pin connector, but the middle pin is unused. The orange wire will open and hold open Santa's mouth when 12vdc is applied. When voltage is removed, Santa's mouth automatically closes by itself (the default state for the mouth). The yellow wire is used for ground.

WHITE/BLACK pair: This is the audio connection that goes straight to the speaker inside of Santa's chest. The audio signal is on the white wire, and the black wire is ground.

RED/PURPLE pair: This controls the left/right movement of Santa's head. Unlike the other connections, the red wire will carry + or - 5 volts, depending on direction. With 0 volts, the head stays in its current position. The purple wire is used for ground.

This information is enough for someone to create their own hack for the Singing Santa. You can bypass or remove the internal PCB and drive the motors and speakers directly. Basically, you've got an analog electronic puppet you can play with. If you want digital controls, or to use the built-in volume control, motion sensor, or start/stop button, you'd have to do a little more work.

Santa's personality chip.

Remove the small board, and you have direct access to Santa's digital controls. This makes it very easy to operate from a microcontroller or even a home PC. Going through the digital controls, you take advantage of the built-in motor control electronics, volume control, motion sensor (and motion sensor enable/disable switch), and start/stop button.

Alternatively, you could use this information to remove Santa's personality PCB, and place it into a completely different device. That could be really funny.

PIN	DESCRIPTION
1	+5 Provided by IC when a routine is active (not used by PCB or required)
2	+5 Provided by IC to move Santa's body back-and-forth aka "dance"
3	--UNUSED--
4	+5 Provided by IC to open Santa's mouth (automatically closes on its own)
5	+5 Provided by IC to move Santa's head to Santa's right
6	+5 Provided by IC to move Santa's head to Santa's left
7	--UNUSED--
8	+5 Supply from board to IC
9	--UNUSED--
10	Red START/STOP button sends +5 to IC while held down
11	Motion sensor, if switched on, sends +5 to IC while motion detected
12	Clock input for IC - not used for this project (ignored/unused)
13	+5 Supply from board to IC
14	AUDIO line output from IC (before amplification/volume control)
15	GROUND connection between IC and Santa

Unfortunately, I live in a highly conservative town. Highly. (Did you read the title of the Call for Entries? "You Can't Show That in Tulsa!" It isn't empty rhetoric.) Although the message could actually be quite pro-Christian (how Santa and the holidays should not distract from the religious message of Christmas), the literal interpretation would be damning.

Santa nailed to the cross, acting like Jesus? If I tried to pull such a stunt, I'd quickly find myself on the local news and be the target of scorn of half the county. My family would never live it down. And I would forever be given a label by the media, such as "The controversial artist, Josh McCormick, has..."

This was probably one of the most difficult choices. I really had to rule that out, even though it would have been more 'artistic' and meaningful. Instead, I had to settle for second best. A drunken Santa Claus with a variety of mischevious holiday messages for the boys and girls.

You can create .WAV files in Windows and record them onto the chip. It'll record up to 240 seconds of audio (segmented however you need) with random playback. The stamp talks to the QV306M4 via simple RS232 communications, checks the playback status with a BUSY line, and resets the sound device with a RESET line. Simple.

But how would I coordinate the actions with those sounds? Writing the code to open and close the mouth and move the head and the body with the words would be just too painful. So I thought of a similar hack... the Boogie Bass Hack. How did they do it?

They had the perfect solution. Record the audio. Then, let the user record the individual movements (head, mouth, dancing) as the audio plays back. Kind of like laying down tracks. Perfect! Except we're talking about a lot more recording time here, which ends up being way more than can be stored in the EEPROM of a regular microcontroller.

From my POV LED project, I had some leftover 24LC515 EEPROMs [PDF]. These are 64KB EEPROMs which were accessable through the I2C communications protocol (a simple 2 wire bus). At 20 samples a second, it could potentially record 3000 seconds of movement. Just one of these was complete overkill for this project.

Finally, I needed a good timer to create a regular 20hz clock to synchronize the movements to. I came across an easy solution, an RC circuit, where with two resistors and a capacitor, I could get a 50ms timer by measuring the time that it took to discharge a capacitor. Thanks to the Parallax Boards for giving me an alternative to more exotic and unnecessary solutions! (And explaining it far better than the Wikipedia page did.)

Santa's secret shoe technology. Personality PCB removed, extra wires added.

I changed some of the sounds in the QV306M4 with new sounds that I downloaded from Quadravox's website. Seems to work well. (A friend asked if I reprogrammed an answering machine. HA! I can see why, with phrases like, "You have a call from... papa.") I can reprogram the chip (and therefore, Santa) to say what I want. It is just a matter of creating my own recordings with what I want it to say.

Testing out the QV306M4 sound chip on the prototype board.

Soon after, I downloaded some Christmas performance samples (Tweety Bird, The Simpsons, Weird Al Yankvoic) from the Internet and download them into the sound chip. It plays them perfectly. I write the code and put buttons on the prototype board to record the actions that go along with the sounds. A bit more developing, and it works flawlessly. Now I'm ready for the real thing. Recording Santa's voice.

Actually, this is the part I dreaded the most out of the project. I considered going to Craig's List and trying to hire someone to do the voice work. Or having a friend do it. I put it off for several days. But I spent the time developing things for Santa to say. I think I hit the gold mine when I contacted a long-time friend, Fipi Lele, for advice. She came up with some real gems this time! I ended up refining my list down to 19 different things for Santa to say.

It was getting to just a few days before the deadline. I had to get it done. I found some decent freeware software (Audacity 1.3 beta) and began recording. I wasn't eager. But once I got started, things just seemed to flow. In short order, I cranked out 19 different .WAV files, cleaned them up, and downloaded them into the QV306M4.

The next day, I started recording the actions to go along with the recordings. First laying down the lip sync. Then the head movement. Then the 'dancing'. There was so little to dance to (because this was no longer much of a singing Santa) that I decided to 'dance' when Santa laughed. It worked out very well. In almost no time, I had a fully programmed Santa! The recorded sounds seemed so much more 'alive' once they had the movements to go along with them. Everything was perfect. The Santa was working, programmed, and complete!

I started to recreate the wiring on the breadboard, and to put all the components in its place. I checked various contact points along the way to make sure that connectivity was good and that I hadn't created any accidental shorts. It was tough because my contact lenses were fuzzing out on me. I'd take a break and come back to soldering once again. But once I put the QV306MP4 onto the breadboard, I noticed a major problem. My ground signal and my +5VDC supply were shorted together. Something was wrong!

I went through unsoldering and soldering back a whole series of connections and components. I went over the PCB with a magnifying glass. I couldn't find anything wrong. I was so screwed.

Finally, it was 2am. The PCB has a short that I couldn't resolve. After a bit, an idea came. Transfer everything back to the prototype board. Call it a 'Drunken Santa Prototype' and it'll actually be a feature. Put the prototype board into a nice display case, and show it off!

Why not? It was my only choice at that point. I unsoldered all the major components. It was about 3am, so I decided to get some sleep and to pick things up at 8am when I was fresh.

The remains of a botched soldering job.
(Palm Pilots take the worst close-up pictures.)

The QV306M4 has a BUSY line which informs the BASIC stamp when it is busy playing a sound, and when it is done. The code uses this line to determine when some of the initialization has completed, and also to know when a performance has completed and it should go into wait mode for the next performance. But the BUSY line was malfunctioning. It semeed that the BUSY circuitry was damaged while it was on the breadboard. It wasn't going to work properly. And there was no chance of a replacement in time for the art show.

So I started to think of how I could get around this second major problem. The QV306M4 could no longer report when initialization is complete, and it could no longer tell me when it is done with a performance. What can I do? The answer was simple. Put in pre-programmed delays during the initilization. But for the performances (which have already been recorded), I could look for the last movement in a routine, and then a second afterward, put in a special flag (decimal 250) to indicate that the routine has ended. Then write code to recognize this as the end of the routine, and remove the code that monitors the faulty BUSY line.

It worked. The extra 1 second delay was a little much, but it worked. And that was good enough. I was so happy that everything was back to working again! I went to Hobby Lobby and picked up a special shadow box to put the prototype board in. I drilled the holes, and a friend who was experienced with a hot glue gun used it to hold the wiring in place. It looked very presentable (even though the original plan called for the board to be hidden inside of Santa's other shoe). It worked, it looked good, and that was all that counted.

Santa Claus and the prototype board are ready to go!

Once I arrived, I was directed to set Santa up in a corner. I tested him out, and everything worked great. While I was there, another artist tried him out and seemed amused. A Wal*Mart Santa is a rather unusual canvas to create art upon. I was confident that it would make the cut in this Monday's judging. I handed over the required $35 for membership in the Tulsa Artists' Coalition.

Two days later, the judging came, and I got a phone call from the art gallery. They said they had a lot of entries this year, and they apologize that the juding was "rather ruthless". My entry did not qualify for the exhibit. I was slightly disappointed, but I was a little confused. From what I was told, the judges leaned more towards mainsteam art. You know... sculpture, painting, etc. No big deal, I really didn't put my self worth on the line with this one. But they wanted it out of their gallery by the end of the day.

I arrive at the art gallery and checked out the remaing artworks. Most of them seem sexual in nature. You know. George Bush with an erection, that sort of thing. Were these the winners, or other losers like myself? I went to my digital Santa. It looked like the prototype board had been roughed up a little bit. A rival artist trying to knock out the competition? I'll never know.

But Santa had some serious problems. He'd play the sounds correctly, but he wouldn't move. Yeah, I can see how a judge would call it lame. The movement really added punch to what would otherwise be a tape recorder performance. A random art teacher who was there at the time to pick up her own work seems to have really liked it. (sigh) It seems that everyone really liked it.

I got it home and I quickly identified the problem. The EEPROM was loose. I pushed it down into the prototype board, and everything worked normally again. That was the downside of using a prototype board... it just wasn't rugged enough. What a disappointment.

I'll never know for sure if it was rejected for artistic reasons, or because it was broken before the judging. My self censorship certainly didn't help. I suppose it doesn't really matter. In the world of art, it may not have been anything special. But in the world of hacks, it was pretty cool.

Here are some more Santa videos: Santa 2 Santa 3 Santa 4 Santa 5 Santa 6
Use these MPEGs only if above video doesn't work: Santa 2 3 4 5 6
(please do not link directly to videos)

The 'jackhammer' sound at the end of "Santa 3" is Santa's lips opening and closing at an inhuman rate. Magical powers, indeed. Mrs. Claus is one very lucky woman! In "Santa 6" is a closeup of the PCB in action. The top LED is used to indicate head movement. The bottom LED is used to indicate the lips opening.

At the same time, it asks a more direct question to Wal*Mart and Gemmy. Won't you sell us a Santa that we can record our own messages with? Your Santa is boring. It does the same thing every time. Why would anyone want to buy it? We want a Santa that we can control and make it do what *we* want! That way, it wouldn't grow old after the first week. We want Santa that we can control and record! (And I'm not talking about the live performance model that you plug a microphone into.)

Are there other characters I can hack? I believe so. I haven't tested these out, but Gemmy makes a number of other characters that should be partially or totally compatible: Snowman, Rapping Skeleton, Animated Skeleton, Animated Pirate Skeleton, Halloween Butler, and Freddy Krueger.