Pocketwatch of Goo

6 Mar 2025
Progress: Concept

One of the problems in having a large backlog of projects not yet published, and a very very long list of half-baked ideas, is that people often suggest to me things which I've already considered, or even already built.

Following the Fluid Sim Pendant video, a multitude of people have asked me to turn it into a timepiece. I acknowledge that it's a pleasing idea, but it's probably not one I'm going to implement any time soon – I have my hands full with several other fluid sim spinoffs. However, the concept is worthy of a few illustrations.

A digital clock face could be illuminated inversely to the liquid.

Nonchalantly reaching into one's pocket, producing a small gold pebble on a chain, flipping up the lid of the pocketwatch only to inspect the glowing goo that dribbles around inside. It's certainly a swish way to tell the time.

The pendant as it stands is far too low resolution to do this. At only 16 pixels in diameter, we'd have to crush the font almost beyond recognition. Omitting the colon and running edge to edge, playing with the pixels I was just able to fit four characters in a 3x5 font.

The addition of goo makes this quite hard to read, and there's nothing worse that a timepiece that's bad at its primary function. Increasing the resolution is possible, but this needs more power (both electrical, and processing).

In lieu of moving to a higher resolution, we could possibly scroll the message horizontally in a bigger font. I'm uncertain about how this would feel.

Even this, when we add back in the fluid, becomes pretty incomprehensible. I suppose I should concoct a proper demo. That might happen at a later point.

But naturally, there are other features a watch requires so a redesign is inevitable anyway. We need a way to set the time. While we're at it, we could do with a dial for viscosity, and an adjustment for brightness.

Taking inspiration from mechanical stopwatches, or perhaps the alethiometer, a triplet of knurled knobs couldn't hurt. Brace for a particularly low-effort illustration:

Building these knobs and connecting them to internal potentiometers or encoders is a major task. I can't immediately see any way to do it without machining lots of complex parts, or making the case a whole lot thicker. The best case scenario would be to find a small right-angle board mounted trimpot, to which we could extend the shaft and attach a knurled knob. Space is at such a premium. Perhaps we could mount the potentiometer externally, with wires to the inside, and fit a knurled cap over the top of it.

One method of coping with low resolution is anti-aliasing. This depends on a display that can render brightness, instead of the monochrome creation I made for the pendant. For the fluid, instead of plotting cells where liquid is or isn't, we can plot the density, which leads to some captivating visuals, particularly as it lets you follow the various vortices that form after a splash.

But plotting antialiased text on a density plot of fluid ends up even less legible.

Instead of merely inverting the pixels, perhaps we should snap to min or max brightness wherever the font is present.

Or maybe the inverted colours method is just doomed to be unclear. More complex arrangements are always possible though. Perhaps, after some gesture (such as shaking it) the liquid could coagulate into the time, and after a couple of seconds, the viscosity would subside and the goo would slip, spread, and gradually return to sloshing about. There is probably a further idea in here somewhere. I think the coagulation could be an interesting challenge, some kind of magnetic attraction to the desired shape. It would look odd when the digits change.

Another idea lots of people have suggested is some kind of invisible maze, that the liquid interacts with as it flows around. That might be difficult to use to convey the time though.

The above diagram is at a diameter of 58, which is well beyond anything we were considering building out of LEDs. I don't own a smartwatch, but their existence has brought about a number of small circular OLED displays. If we were heading in that direction, a simpler option would be to produce an app for a smartwatch. I assume such devices support apps, or at least customised displays. It could be quite cool.

One final avenue, because I can't seem to let go of it, is a real liquid timepiece. An ordinary LCD watch display uses polarised light. If we could somehow come up with a liquid, along the lines of an oil/water mix, that affected the polarisation of light, it could naturally invert the brightness of the display.

From a physics perspective, I'm not quite convinced on how we could do this. Generally, outside of solid crystals, polarising filters are things which exist in thin planes. About the only way I could imagine it working, and still blobbing around like goo, would be to have some kind of interaction at the surface of the glass. Giving the glass a static charge (like an electret) might work, so that as the lipids in one phase of the liquid approach the surface they line up. But, this might just end up with a thin layer covering the whole surface.

Even if it worked, the end result might just look like some goo got trapped in your watch.