PF3000 v2’s Touch Sensors: Making Art for Napa Lighted Arts Part 2

December 17, 2025 Doctor Cosmos

Hello friends, welcome back. As you know, we’re crunching hard on the next iteration of our award winning PF3000 and things are moving right along. Heck, we’re less than a month to the show – holy moly!

We thought what a great time to talk more about the super cool engineering that’s gone into this project. Back in part 1 of this series we discussed some of the challenges we faced with little things like winter weather and the structural modifications we made to ensure endless safety and joy. For part 2 we’re going to talk to you about our latest touch sensor and how we built it to be ready for hundreds of thousands of interactions.

Alright, let’s put on our electrical engineering hats, get a nice cup of coffee, and dive right in.

Background

PF3000’s original touch sensor: Indium-Tin-Oxide (ITO)

Touch the pixel and it lights up, touch it again to change the color. This is an in-progress photo from the PF3000 v1 build from way back.

Before we tell you about the new touch sensors I’d like to talk about the old ones. For the original PF3000 we wanted to have backlit “pixels” (backlit meaning: the light comes from behind) that you could touch directly. Each pixel would be a square and when you touch the square that square would light up or change color if already lit.

Backlighting the pixel created a unique challenge given how we wanted the touch interface to work. Most touch sensors are some form of button or conductive surface (like copper) that block light altogether or at best are translucent with some degree of light loss. We wanted all the light to shine through our pixels so that when you experience the PF3000 you’re literally playing with big beautiful colored light; only the best for you.

Note: There are various ways to inform our system that a touch has occured and then spawn the lights. For example you can use mechanical devices, like traditional buttons, though these are prone to breakage, difficult to scale, and difficult to waterproof. Alternatively you can use a material that conducts electricity for touch interfaces. As you interact with the conductive surface we can detect changes in voltage on that surface. This is nice because there’s no moving parts which removes a lot of complexity and maintenance burden. At least, in our opinion.

We knew backlit touch interfaces were a solved problem though. Just about anyone has the solution in their pockets these days. No, I’m not talking about your car keys, I’m talking about your smart phone. Like, consider this image I stole from the internet:

The cell phone is back lit – light comes from behind the phone screen through leveraging some mysterious technology. You are then able to then touch images displayed on your phone and they react to your touch. This tech is hardly mind blowing these days, but my point here is that it means there must be some material out there that you can shine light through and also use to build a touch interface. So what is that mystery material and can we use it for art?

Introducing: Indium-Tin-Oxide. It comes in little sheets like this and is surprisingly flexible.

Turns out there are a few options, but there’s one option in particular that stood out during our research: Indium-Tin-Oxide or ITO as it is often referred. This stuff is near completely transparent and it conducts electricity. What it is exactly is a thin clear composition that is applied to surfaces like plastic and glass. Because its conductive, you can use it to build touch interfaces out of; this material is commonly found in devices like phones too. It’s probably a bit beyond this article to explain ITO deeper than this but you might check out this wikipedia article if you’re interested in learning more.

The below video is also cool if you just want to see ITO in action:

This material is just fantastic, it really is. It worked so well that in the original PF3000 build we just slapped a sheet of this ITO stuff on each pixel; all 256 of them. You just touch the ITO we applied to the surface of the PF3000 and voila – touch interface magic complete!

Getting Pixel with it at the Autumn Lights show in Oakland, CA

There is however one little problem with ITO

Things worked pretty well… for a while. And then… well.. things stopped working so well. We found that after the Oakland Autumn Lights festival with a tens of thousands of touch interactions the ITO wasn’t as sensitive as it was in our shop. Specifically we noticed that the centers of our pixels weren’t that touch reactive after significant use but just the centers.

We actually figured this out by watching people try to tap-tap-tap the pixels with a single finger and noticed that the pixels were reluctant to change color. However when folks would slide their hands across multiple pixels they would change color just fine; hrmmmmmm. After witnessing this behavior a few times we started experimenting ourselves by poking different areas of the pixel with our finger, and yep, the center of the pixels weren’t working that well but the outer perimeters of them were working great. What gives?

And so over the first couple years we had the PF3000 at the Oakland Autumn Lights show we watched people interact with the pixels and tried to figure out what was going on. By and large people would use a single finger to poke the center of the pixel to change its color; this was happening over and over from almost every person. Poke-poke-poke-poke. And where were people poking exactly? Naturally right on the center of each pixel, you know, where things weren’t working that well.

Cracks in ceramic. Despite occasionally aesthetically pleasing, not particularly electrically performant.

It turns out ITO is sort of a ceramic, like that flower vase that your mom made last weekend. If you’ve ever interacted with any kind of ceramic you’ve probably across the situation where that ceramic cracks (sorry about your flower vase mom!). That’s exactly what was happening with our ITO sheets; as users would poke-poke-poke the center of the ITO sheets making their sweet pixel art, little microscopic cracks would form in the area they were poking. These cracks, while we couldn’t see them, are large enough to affect the ITO electrically. That is, the cracks in the ceramic ITO layer were preventing electricity from flowing from the center of the sheet to our wiring on the outer perimeter of the sheet. Where people weren’t poking the ITO very much these cracks would not exist and so electricity would flow fantastically which is why touching the outer perimeter of the pixels worked but not the centers. Neat! So uh, now what?

A New Sensor for Napa Lighted Arts Festival 2026

Alright recap – we’re all about this cool transparent conductive material called ITO as it makes for a fantastic touch sensor that you can shine light through. The big bummer though is when you touch a touch sensor a lot it wears out. But what’s all that mean for the Napa Lighted Arts Festival?

Napa Lights Arts Festival expects nearly a 100,000 visitors and this is a lot more people than any show we’ve done previously. Autumn Lights Festival in Oakland, CA has around 20,000 visitors which is where we normally hang out – so about 20% of the people that we’d expect in Napa. And we know that in just 3 days of interaction at Autumn Lights we’ll see touch sensitivity degradation (say that three times fast). You don’t need a PhD to see where this is going – if we take v1 touch interface design to Napa Light Arts 2026 the PF3000 is going to totally not work very well for very long.

We want everyone to have the same fantastic experience with the PF3000 no matter what night they come to the 30-day festival. We need to come up with a touch sensor that’s robust, practically indestructible. Something that can not only stand up to weather, but stand up to millions (possibly billions? million billions?) of hands and fingers mashing away as folks play and design with light. And we need it to be something that we can still work with our original backlit design cause we can’t exactly just re-do the entire art piece.

Hrm. This is quite the problem to solve, and solve it we did. Keep reading, this is where things get nerdy.

PF3000 v2 – Capacitive Touch Pixels

Through research we found that we could use the ITO in a different way. I’m going to explain exactly what this means as the article continues but if you want the TL;DR its basically this: we can turn the ITO into a sort of proximity sensor meaning you don’t have to touch the ITO directly rather you touch it indirectly. This proximity sensor works because you create a capacitor using your hand and the ITO – this is type of touch sensing is known as capacitive touch. Using this capacitive touch we can place a strong protective layer on top of the ITO to prevent it from wearing out; instead of touching the ITO directly you touch this protective layer and the art piece functions amazingly forever (or so we hope).

Capacitive touch is kind of complicated to be honest, but we’ll try to break this down as we present to you the PF3000 version 2 capacitive touch pixel.

First things first – pictures (and now you know why people don’t pay me to make pictures)

I like to call this a sammich pixel because the sensor magic is sandwiched between two thick layers of acrylic (aka plastic). Each layer matters though some of the layers are more about qualities of light whereas others are there for capacitive touch. How about a quick bulleted list of highlights?

Starting from the bottom going up:

The bottom layer is a thick acrylic diffuser. This is just a big plastic square that is kinda foggy so that when light passes through it kinda ‘fills out’ the square with light.
Surrounding that acrylic diffuser is a led strip. It actually wraps around the whole diffuser layer and shines colored light into that diffuser.
Above the diffuser is a sheet of indium-tin-oxide or “ITO” (i’ll come back to the adhesive in second cause i did totally skip that). This is that magic stuff I’ve been talking about the whole article, though we’re using it in a totally different capacitive way.
The top most layer is the Acrylic Dielectric and is the “protective layer” – this is a 3mm piece of acrylic (plastic) and basically functions as the ‘touch surface’ but also armor for the ITO. This is what you actually touch.

And then there are two layers of adhesive. The upper layer is electrically necessary and I promise, I’m going to explain why as its quite fascinating… but not yet. The lower layer of adhesive exists because it’s convenient to use an adhesive there; sometimes you just want a little glue.

Now let’s talk a bit about how this capacitive touch business actually works.

Capacitive touch – tl;dr edition

If you really want to understand like the fine electrical engineering details of capacitive touch there are resources out there much more comprehensive than I intend to go into here. Heck, I have a small handful of engineering degrees and understanding capacitive touch took me some work. Since I want this article to be available to folks with different back grounds I’m going to try to keep this brief but hopefully give you enough information to follow along. Strap in, you’re an electrical engineer now, just like me.

Capacitive touch is this idea that you form a capacitor with your hand/body and some conductive surface. The closer your hand gets to that conductive surface the properties of that capacitor change in a measurable way. Once we measure enough of a change to that capacitor we consider the sensor as touched.

Capacitor – A crude definition

Capacitors often look like this on the outside but its the inside where the magic happens.

Now if you are new to electrical engineering that explanation probably doesn’t mean much. Let’s break it down some more. Starting with: what’s is a capacitor?

Textbook definition: A capacitor is a passive electronic component that stores electrical energy in an electric field, acting like a tiny, fast-charging battery that releases energy quickly, made of two conductive plates separated by an insulating material (dielectric). It’s used in circuits to smooth voltage, filter signals, block DC, and store energy for applications like camera flashes or digital memory, measured in Farads (F). Totally makes sense right? Keep reading, it gets easier.

Usually capacitors are formed by two metal plates sandwiching an abstract concept known as a dielectric. Capacitors are a common electrical component much like resistors, transformers, diodes, and transistors.

Textbooks are all fine and good but I don’t find them particularly helpful if you’re new to this stuff. Now for my layperson definition for the purposes of capacitive touch. Capacitors themselves often look like two metal plates with some stuff in the middle (see the image above) – this ‘stuff in the middle’ is called the dielectric medium but there are a bunch of different materials suitable for dielectrics. Electrons build up (aka charge) one of the metal plates – in the picture the charged plate has all the +++++ symbols to indiciate there is a positive charge on the plate (electrons baby yea!). Then the electrons pass from this charged plate (+++++) through the dielectric medium to the other plate —–; essentially electricity is flowing from one plate to the other through the dielectric stuff. Because of the dielectric medium not all the electricity is going to flow from the +++ plate to the — plate, just some of it. With me so far?

Capacitance is a measure of how this capacitor is ultimately going to impact those electrons as they move between the two metal plates. The size of the metal plates, distance of the plates, and that dielectric ‘stuff in the middle’, all impact the capacitor’s capacitance. In other words, capacitance tells us just how this capacitor is going to affect electricity and we use a unit of Farads for this (Farad is an old dead scientist fwiw). That said, a single farad capacitor is huge (see below); the capacitors we usually work with are measured in micro-farads.

A 1F cap is about as big as an adult forearm

I’m leaving a lot out just to be super real. Capacitors are pretty complicated but to get the gist of how our sensor works you just need to know a couple of things I’ve already mentioned:

Capacitors charge up electricity and then some of that electricity goes through the capacitor from one metal plate with a positive charge to the plate with a negative charge.
Capacitors have physical properties that impact their effects on electricity. Things like the distance between the two metal plates, the size of the plates, and the type of material used as a dielectric
Fun note: we actually call them caps.
Another fun note: if you plug electrolytic capacitors backwards they explode, we used to blow them up in the lab in college. Just the little ones though

Capacitive Touch – When you’re a capacitor too

Now circling back to this whole touch sensor thing. Imagine for a second that the charged plate (+++) is that ITO I’ve been telling you about. We charge it up, basically we put a bunch of electrons on it. These electrons are just sitting there waiting for somewhere to go. They won’t just go anywhere on their own, they’ll just sit there. Electrons need both a medium to travel through as well as something to convince them to go on the journey through that medium; the negative plate (—–) is usually what convinces the electrons to get moving. Without a negative plate waiting for them they don’t have anywhere to go just yet though. But they will.

A 3mm thick dielectric medium… aka acrylic.

On top of that charged ITO we insert a dielectric, a 3mm piece of acrylic. This is a material that electrons will travel through should they have some destination. The choice of material matters, different materials have different electrical properties and the measure of these properties for our purposes is called the dielectric constant. Acrylic happens to have a pretty good dielectric constant meaning it will let some electrons through. Other materials, say tempered glass, are a lot better but we use acrylic over glass due to cost and weight as its dielectric properties are good enough for capacitive touch.

Going deeper: Air has the default dielectric constant of 1 and its what we compare all other materials against – it is worth noting that air is not a very good electrical transmission medium though also not the absolute worst. Acrylic has a dielectric constant of 4 and glass has a constant of 8; the higher the number the better it is as allowing electricity to transfer through it. This stuff matters because, among other things, the higher that dielectric constant, the more sensitive the touch sensor will be; for example a touch sensor made with tempered glass will be twice as sensitive as a sensor made with acrylic. Check out this cool chart to learn about other materials.

hrm, maybe we should have made the sensors out of water.

Now imagine that second plate, the —– one, is actually your hand. When your hand is far away from the ITO the electrons just sit there, excited but stationary; remember that while air is a dielectric, its a really terrible one and so the electrons don’t really want to go anywhere. As your hand gets closer to the ITO the electrons start to get a little motivated as you make an excellent destination for them (you just happen to be negatively charged). Finally once your hand gets all the way to touching the dielectric (in this case the acrylic) the electrons have their destination (you) and travel through that dielectric (acrylic) to get to their happy place (you) and that’s exactly where they go.

This is our left handed oscilloscope. In black you can see the nominal voltage reading. When there’s a touch, this voltage will dip; on an oscilloscope it just looks like a smaller peak which i’ve drawn in red.

What, did you really think all engineering diagrams were fancy?

We’re able to detect touch at this point. On the charged plate (+++) we’re able to read a voltage, like you would with a digital volt meter. When the electrons travel to you through the acrylic, that voltage drops for a period of time. If the voltage drop goes beyond a specified threshold, our little controlling computer calls that a touch. Cool right?

Uh Electrocution bruh? Electrons jumping to my hand? Oh my! But Don’t worry! You won’t get electrocuted because this is actually a really really really small amount of electricity, not enough for you to even feel. But not so small that we can’t measure it.

Now, as per usual, I’m leaving a lot of details out as this stuff is quite a bit more complicated. I’ve just given you the basics on how capacitive touch works. In practice there are a lot of factors we need to contend with from humidity to the general entropy of the universe and then some. Oh, and air, actually, let me tell you about that rq before we wrap this article up.

The “Air Gap”

Remember way back in this article where I said I’d explain some adhesive we used? Do you also remember where I nonchalantly mentioned the dielectric constant of air? Yea – there’s been a subplot this entire article and we’re finally gonna wrap it up. Let’s go back to our sweet diagram of the capacitive touch pixel but note where I’ve drawn a red circle:

Note the upper layer of adhesive with the circle around it. This is actually electrically necessary, read on to find out why.

In our sammich pixel we’ve got all our various layers. While I’ve talked to you a bit about the ITO and the dielectric, I haven’t really explained the adhesive layer between them and well, this is pretty interesting and not necessarily that intuitive (at least we didn’t think so).

Let me start by asking you a question — When you stack two pieces of paper on top of each other, what is between those two pieces of paper?

Think about it for a second, I’ll wait…

Okay, ready for the answer? Well, not much but not nothing. Even though those two pieces of paper are smooshed tightly together it turns out there is something between them. Something that you’re breathing right now. Yep, air exists between those two pieces of paper.

It turns out air gets in all kinds of places if you’re not careful. In our case when we placed the dielectric acyrlic on top of the ITO things didn’t work very well. I mean, they sort of worked, but very poorly. This turned out to be because of, yep, air. Air was slipping between our Dielectric Layer and our ITO, this was a whole additional material we hadn’t accounted for in our dielectric.

But why is a little air such a big deal? If you’ve been paying attention you might remember I mentioned that air has a dielectric constant of 1; as in the lowest. As in air is absolutely terrible at moving electricity around (which most of the time is a very good thing lol). When air gets into your capacitor between either of the plates and the dielectric electrons don’t successfully flow when we need them to; this messes up touch sensitivity significantly. Air is a pretty regular thing you have to compensate for in capacitive touch designs too. Us engineers even have a formal name for this phenomenon: air gap.

Careful management of an air gap substantially improves the sensitivity of a capacitive touch sensor but how do you ensure there is no air between layers? This is where the adhesive comes in. Adhesives are sticky. That is, to stick an adhesive they bond to things very tightly. This bond is so tight in fact it doesn’t leave room for anything else… like air. You have to use the right adhesive of course, an adhesive that is itself a friendly dielectric, and luckily for us those are pretty easy to get (pictured below).

You can’t just use any old glue. In our case we found 3M’s 467MP acrylic adhesive double-sided “tape” (its not like masking tape, let me tell you). Since it is acrylic with a known dielectric constant it works well for our purposes.

And now you know why we have a layer of adhesive between the acrylic and ITO. It took us a while to figure out this air gap business, it really did. I just think its absolutely fascinating myself; maybe you do too?

Air Gap In Documentation

There’s a large pile of technical documents you need to read when designing capacitive touch sensors. I’ve extracted an excerpt from one of these documents regarding the air gap so you can see what they are like. Believe it or not this is one of the easier documents we had to read, some of them are quite intense. Check out the full document here.

We actually read this page a thousand times and still didn’t quite figure out what to do. Thankfully we came across a forum post deep on the interwebs which led us to the final solution.

Putting it all Together

Phew – you made it past all the tech stuff, nice work! Did you have to grab a second cup of coffee at some point? Let me tell you, designing this stuff required a few cups of coffee on this end too especially when we were trying to figure out that air gap. All in all we probably spent about 4 weeks getting to a basic prototype and then another two months doing deep engineering work get to a repeatable quality experience.

This of course was just the beginning. Once we understood how the sensor was going to work we had to build it, well, them. The PF3000 includes 256 individual touch pixels. We made every single one of those pixels by hand (that’s a lot of sammiches). It took a small team a few hundred hours to get all the work done. Maybe I’ll share the details of the build process with you all sometime but that’s a whole different story. For today I thought I’d share some media I took while we were doing this mountain of work. Ah, Memories!

We recycled the diffusers from the original touch pixels. This meant cleaning them all up, removing any prior adhesives, and scraping any stuck bits of dirt off all 256 of them.

ITO sheets are very plasticy and this isn’t very good for attaching wires to. To solve this problem we place a perimeter of copper tape around the ITO on each pixel as we can easily solder wires to the copper. Since there are 256 pixels with four sides each we have to apply 1024 pieces of copper tape. This took DAYS… so I made a couple time lapse videos.

Indium-Tin-Oxide comes in sheets so we use a paper cutter to cut the ITO to size. You can see Robert in the back there grinding away on applying adhesive to the squares. Yea, we had some late nights. In case you’re wondering, Robert and Carl seen here in the picture are the designers of the new capacitive touch pixels.

It took a long time but eventually we finished. Things went a little faster because the extended Filma Community helped at various stages. All in all the rebuild of the touch pixels involved at least 6 people and we’re so grateful for their help.

From July to October we re-built the PF3000’s 256 touch sensitive pixels with the new capacitive touch design. We debuted the new Pixels at the 2025 Autumn Lights show in Oakland California. We felt it was important to test the new design on a small crowd before committing it to napa. I think it went successful, see for yourself.

We debuted our new touch pixels at the 2025 Autumn Lights show in Oakland california to thousands of excited families, really a small fraction of what we expect at the Napa Lights arts festival. We figured we needed to a show with 20,000 people to ensure we could live up to Napa’s massive lights festival. Everything at Autumn Lights 2025 went great and worked quite a bit better than our original design; read: no touch sensor degredation, woot! You can see for yourself in the above video, assuming I didn’t link it incorrectly (sometimes I do lol).

This was a huge undertaking for the team here at Filma Collective, especially Robert and myself who headed up the touch pixel redesign and implementation. We’re really excited to bring our new touch pixels to Napa and we think they will hold up splendidly.

Thank you for making it this far and investing your curiousity in our art. While this was a massive piece of the version 2 upgrade, there’s so many stories left to tell. If you haven’t, check out part 1 of this series and learn about the structural engineering that’s gone into PF3000 version 2 and look forward to part 3 soon.