As some of you may already be aware, a painting done by Banksy, a well-known (not to mention still anonymous despite being in the public eye for several decades now) British street artist, activist, and director, partially shredded itself immediately after receiving a winning bid for more than $1.3M U.S. dollars (1 million Pound sterling) at Sotheby's auction on Friday night, October 5.

Here's a video of both the preparation and delivery of the prank involving the Balloon Girl (aka Girl with Balloon), posted on Banksy’s YouTube channel and website (as well as his Instagram account):

So how did Banksy pull off this widespread attention-getting stunt, squeezing all the necessary electrical and mechanical components inside the frame … and how can you replicate (if not one-up) it on your own? Details on Banksy's setup are sparse at best, but in combination with my own engineering chops, I'm going to take a stab at answering these questions. And of course, please sound off with your own thoughts in the comments!

The shredding mechanism

Let's start with the most noticeable aspect of the design, the hardware that sliced (a carefully chosen word … read on for the details) up the painting. I'm guessing that the first thing you think of here would be a paper shredder. For mental calibration's sake, therefore, I'll begin with a few photos from the upper portion of my InfoGuard NX80P, which contains the motor and rollers and fits into a 2.5-gallon trash can-reminiscent plastic receptacle below it:

As you can see, I've already pierced through the egregious warranty sticker to get to the screw below it, in the process of an unsuccessful past attempt to resurrect its balky auto-stop mechanism.

The upside of this approach is multi-fold: for one thing, the InfoGuard NX80P will (supposedly, at least) shred up to eight sheets of paper at once, along with making short work of even more resilient plastic credit cards, optical discs, and the like. Also, by virtue of its robust design (along with gravity), it will grab onto and pull material to be shredded into the mechanism.

But here's the mental calibration bit: it weighs more than six pounds, according to my kitchen scale, and has dimensions of approximately 13″ long by 7″ wide by 4″ deep. Not exactly easily stowed within a painting frame, eh? Not to mention the fact that it is AC-powered (again, not exactly nondescript). DC-fueled paper shredders do exist; check out, for example, dual-AA-powered Shreddie or any of its similar looking clone-competitors on Amazon. But Shreddie only handles paper up to 4.5″ in width and, to put it kindly, online reviews of reliability and battery life for such products are unkind .

And here's the thing: cross-cut shredding , which is intended for dealing with confidential information and thereby turns a piece of paper into something akin to a pile of confetti, is overkill here. In fact, if you've already watched Banksy's video at the beginning of this article, you may have noted that what was done to Balloon Girl was more akin to slicing the painting into a bunch of strips. Conceptually, at least, you can accomplish this more sensible goal by passing the paper through a row of fixed-position, non-moving blades; a much more compact, not to mention battery-friendly, approach. And if you watched the video, you've also already noted that this is exactly what he did, using a row of 38 cutting elements reminiscent of the X-Acto knife blades I used to create my plastic and wood scale-model masterpieces as a wee lad.

There's only one potential hiccup here. By pushing paper into a row of blades below it, versus pulling it into rollers as with a shredder, you've incurred a non-zero probability that the all-important initial pierces of the paper won't actually succeed (especially with thicker-than-average-paper painting canvas material), leading to only crumpled paper jams ahead of the slicing mechanism. Blades can only be so sharp, after all, and atmospheric moisture-induced rust and other long-term environmental effects also don't work on your favor; recall that Banksy's video claims that he assembled the frame “a few years” prior to the auction.

Your likelihood of success is far greater, however, if the initial pierces are already in place and the paper has already been partially fed through the slicer. And, I'd wager, that's exactly what happened in this case. Look closely at the post-sliced state, and you'll see a distinct “bend” in the canvas partway through. Note, too, that the bottom part of the sliced portion of the painting wasn't visible when it was in the frame, only after the slicing was complete. That's because, I'm postulating, it was pre-sliced and loaded up ready to continue the procedure, once the stunt began.

The battery pack

Even if the cutting mechanism is passive, there's still plenty of other electricity-fueled equipment still necessary in the design; the motor (and connected roller) that pushes the painting down and onto the blades, for example, along with the remote control circuitry that activates the assemblage (both of which I'll discuss next). With an AC power cord an obvious non-starter, that leaves an integrated battery pack as the only feasible alternative (wireless long-distance power transmission and solar cells aren't credible options, sorry).

The top two (IMHO, at least) battery technology alternative candidates are alkaline and lithium-metal. Alkaline is less expensive on a per-cell basis, but in a one-off design like this, that's not a particularly critical selection criterion. More important here is lithium's significantly higher comparative stored charge energy for a given-sized cell, which is beneficial in several key respects; higher instantaneous current output (key when overcoming initial static coefficients of friction in the motor, when initially cutting through the painting's canvas, etc.) and longer shelf life. And, if Banksy had decided to tear apart a smartphone, tablet, or laptop in order to obtain the battery pack, for example, he'd also have benefitted from lithium polymer’s form factor-moldable characteristics; its rechargeability is, conversely, a don't-care in this particular application.

Speaking of shelf life (I'll remind you again that Banksy's video claims that he assembled the frame “a few years” prior to the auction), I'm going to take this opportunity to toss out an admittedly controversial theory (full disclosure: my wife thinks I'm off-base). Watch that video again. Notice how the Sotheby's auctioneer and other staff remain unnaturally calm when the ear-piercing beeping noise (easily accomplished via a diminutive and inexpensive piezo transducer, located either at the remote control transmitter or, more likely, within the picture frame) kicks off and the sliced-up painting begins slipping out the bottom of the frame? Doesn't that seem odd to you, too? Perhaps it's just a reflection of the perpetually serene British mentality, but I'm skeptical.

My theory, public disavowals from Sotheby's aside, is that the auction firm was in on the stunt in advance. Specifically, someone on staff flipped on a power switch on the back of the picture frame right before this particular auction began. Think about it; Banksy got tons of publicity out of this, as did Sotheby's. And unsurprisingly, the winning bidder decided to keep the sliced-up painting, rightly guessing that it was worth much more now than before. Everyone wins!

Even without the motor(s) operating, and regardless of the battery technology chosen, motor leakage current would still slowly sap the connected battery pack, not to mention the steady electron drain induced by the regularly “sniffing” remote control receiver circuitry. The only reliable way to ensure you had a fully charged battery pack available when you needed it would be to cut power completely until you needed it. In this respect, Sotheby's would be a convenient co-conspirator.

Remote control

After watching the video, you'll likely (and accurately) conclude that the slicing operation was remotely kicked off by someone in the audience. The task to be performed here is as elementary as they get; a one-shot “on” function. And if cost was a key criteria, infrared (IR) would certainly fit the bill; IR receivers at Arrow (full disclosure: Arrow is the parent company of EDN's publisher, Aspencore) cost only around 50 cents in volume and are still less than a buck in single-unit quantities.

IR comes with multiple issues, however. You wouldn't want the IR portion of the output spectrum of a camera flash, for example, to prematurely turn on the motor circuitry. On the other hand, as anyone who's operated an IR-based TV remote control already knows, the natural background IR “radiation” of any environment (especially outdoors, although that wouldn't apply in this particular case) can destructively interfere with the desired emissions of an IR remote control transmitter. And then there's the fact that IR operates only in line-of sight fashion; I doubt someone in the audience striving to act surreptitious would be enthralled by the need to stand up and point a handheld controller at the painting!

Therefore, in a case like this, I'd suggest that a Bluetooth or other RF wireless approach would be best. And indeed that seems to be what was used in this particular case. Here's the money quote from the New York Times coverage :

A photo posted on the private Instagram account of Caroline Lang, the chairman of Sotheby’s Switzerland, showed a man in the salesroom operating an electronic device hidden inside a bag. Ms. Long said that she later saw a man being removed from the building by Sotheby’s security staff.

The motor

Last, but not least, what about the motor that pushed the canvas through the row o'blades? Compact, powerful, cost-effective DC brushless motors are a dime a dozen nowadays; in the latest twist on the longstanding scandal-plagued bicycle racing industry, they've even been found (accompanied by batteries in the seat tube and/or down tube, presumably) built into bikes' bottom brackets, where they connect to the crank arms and supplement the rider's own pedaling efforts.

As an example of what an integrated motor-plus-RF receiver assembly might look like, one that's conveniently also available off-the-shelf in case Banksy (or a wanna-be) didn't want to bother doing his or her own design, I hereby present Somfy’s model CL32 RTS (reference #: 1002422), six examples of which currently power the wireless remote-controllable motorized blinds in the sunroom of my house. Here's an outside shot of one of them tethered to two parallel-wired Somfy 12V lithium battery packs (sold at Amazon and elsewhere) and mounted inside at the top of one of my sunroom windows:

And here's the Somfy Telis 4 RTS remote control that zone-manages all six motors, with a penny for size comparison:

Back to the motor assembly. The spec sheet (PDF) includes (among other things) the following high-level information:

  • Voltage: 12V DC
  • Torque: 0.34 Nm
  • Thermal protection: 2.5 minutes
  • Over current protection: 1.2 amps
  • Sound level: 53 dBA
  • Speed: 68 rpm
  • Radio frequency for RTS: 433.42 MHz

On that last point, by the way, RTS stands for Radio Technology Somfy, a Somfy-proprietary (at least in intention, I presume, although the hacking community once again comes through) protocol.

Ironically, per the discussion in this particular article, the retailer I bought the blinds from several years ago had to come back and replace all six CL32 RTSs shortly after doing the initial installation, because the battery packs were already drained (due, I was told, to faulty motors that never completely turned “off”). I convinced the installer to let me keep the six original Somfy units, with visions of future teardowns dancing in my head. Here's one of them, again with a penny for comparison:

The all-white wire is the antenna, while the dual-color dual-wire connects to the battery pack. That black metal rod coming out one end of the CL32 RTS goes into the blinds' roller mechanism. Removing it required the use of my iFixit 64-bit driver kit. Here are resulting views of all sides-and-ends of the now rod-less CL32 RTS:

Time to dive in. Removing the plastic mounting brackets on either side was easy, involving only a flat-head screwdriver acting as a lever arm. Taking off the top was even more straightforward, involving the removal of two Phillips head screws. The motor assembly pulled right out.

In the image above you can see the motor, the RF and voltage regulation circuit board (I'm guessing at its functions at this point, since both the antenna and battery pack wires go straight to it), and the motor control circuit board (again I'm guessing at this point, based on wire assemblies connecting the two together) mounted along the side.

The image below is a closeup perspective of the RF and voltage board and its connection to the mated motor control PCB along the side. The boards are swathed in mysterious amber translucent tape that I've never encountered before (readers, what's its use aside from likely non-conductive protection?)

You might not have noticed a Phillips screw also underneath the mysterious amber translucent tape. Or maybe it was just me … I torqued around with a flat head screw driver for a few minutes, trying to pry up the RF board incorrectly guessing I might be dealing with stubborn glue. Then I realized what was actually holding the PCB in place. Prepare for liftoff.

The dominant IC on this side of the RF-and-voltage PCB is Maxim’s MAX1473, a 315MHz/433MHz ASK (amplitude-shift-keyed) data s uperheterodyne receiver. I was right! I'm guessing the IC in the lower right corner is a voltage regulator, but I can't discern any markings on it.

Following a tedious yellow-tape-removal process (that's some seriously stubborn stuff!) I was able to see the motor control PCB clearly. Atmel's (now Microchip Technology's) ATmega644PA dominates the landscape on the front side. Given the 8-bit AVR microcontroller's peripheral mix and feature set, I'm feeling pretty confident in my initial “motor control” forecast for this particular PCB's primary function.

Also freed from their tape bondage, the motor shows some really strong magnets (they kept grabbing the frame of my camera lens), which were a bear to peel away from the motor below them. They immediately stuck together afterwards. But finally, exposed to view, is the part number of the motor, along with a few high-level specs.

I'm not a motor expert by any remote means, so I'm going to provide these last disassembly shots sans commentary. Knowledgeable readers, feel free to chime in below to describe what we're seeing.


There you have it, folks. A media attention-grabbing stunt orchestrated by a creative artist, based on a design that probably wasn't as complicated as it might have appeared at first glance. This is my take on it. Did I overlook, under- or over-analyze, or otherwise mess up anything? Let me know in the comments below. I look forward to hearing from all of you!

Update : Banksy has since said on Instagram, “Some people think the auction house were in on it, they weren't,” and posted this video showing more of the preparation and a rehearsal where the picture shreds completely. Does it provide any new design clues?

Brian Dipert is Editor-in-Chief of the Embedded Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company's online newsletter .