rebecca's WIP

[pcomp final] Prep for final user test

This weekend was a marathon. Nick and I focused primarily on fabricating our planter box and creating our system on one plant in preparation for our final user testing on Monday. Here’s our progress from the past few days —

We created our planter box out of 3/4 inch birch plywood and in the process came to realize the extent to which wood + water + electronics don’t mix. Or at least present a very tricky combination.

The 2 x 2 foot box needs to be very sturdy, since it will be holding the weight of a lot of soil, but we also want to keep the wood somewhat unmarred on the exterior (ie. no screws). So, we used dowel pins to hold the pieces of the box together. We couldn’t find a doweling jig in the shop so I made my own out of a piece of plywood and it worked fairly well. In the end, we used dowel pins and wood glue, and added just two screws to one of the sides for some extra stability.

2014-11-29 15.55.20 2014-11-29 18.13.51 2014-11-29 18.14.05

For the legs, we used speed rail and bolted it to the bottom of the box. The planter comes up to slightly above waist height for most people (except Nick) and is surprisingly sturdy. Fingers crossed!

2014-11-30 15.03.46
2014-11-30 15.11.49

Finalizing sensors and system:
The other main goal this weekend was to set up a full system on one of our plants for our Monday user testing. This entailed making sure the sensors and serial communication were both working, and hooking everything into Max to produce responsive sound.

2014-11-29 20.19.55

On Saturday when we tested everything without the cap sensors (moisture + photocell), it all worked great. But yesterday when we included our cap sensors in the circuit, we encountered some strange behavior that we documented in the video below.

In case it’s hard to see in the video, the cap sensor gives us an error of -2 and slows everything down. The sensing pin is hooked up to the stem of the plant with an alligator clip, and the moisture sensor is embedded in the soil next to it. As soon as you remove the soil sensor, the cap sensor starts working again and reading a normal range of values.

After banging our heads against the wall for a while, we realized that the soil sensor was causing the issue. We’re not sure why (maybe something with the charged probes in close proximity?) but we need to figure it out, since all the sensors working together are an integral part of our system. We probably should have tested this earlier, but we now know what the problem is and can take steps to fix it.

For the user test, we’ll “fake it” and not use the moisture sensor (since the moisture level in the soil is not something that will respond to the user anyway). The main thing we want to get feedback on is whether the sound mapping is intuitive and responsive enough. We want to ask people: do you feel like your actions are causing the sounds? Can you tell which action corresponds to specific changes? Is touching the plant intuitive? Satisfying?

(Note: the sound itself is currently very much a work in progress and we will finalize that this week).

Testing a drainage system:
We also experimented with different ways to waterproof and create a drainage system for our box. We’re worried about drowning our plants and also ruining the wood. We tested everything out on a scrap box that we found in the shop. We lined it with plastic, drilled holes in the bottom, attached a funnel and some plastic tubing, and then put a bunch of wet soil in it to see how it would drain. We’re still undecided on whether or not we want to include a drainage system in our final iteration. It seems unnecessary, since we will no longer have users watering the plants.

2014-11-30 17.02.40 2014-11-30 17.04.57

We also started waterproofing the inside of our final box with different kinds of plastic liner and tape. Halfway through, we realized we need to give this more thought, and possibly use some more heavy-duty waterproofing techniques such as caulking and adding polyurethane to the inside.

The one thing we didn’t accomplish this weekend was experimenting with our grow light and engineering the lighting environment we want. We’ll need to do this ASAP. Below I listed out all of the remaining things we have to do on the project to get it ready for next Monday. In an attempt to prioritize, I went through and bolded everything that’s priority for next week. The other things are more cosmetic changes that can potentially wait until later on.


  • Resolve problem with capacitive and soil moisture sensors (error where we get -2, see documentation)
  • Set up system and circuits for multiple plants (also involves figuring out how exactly sound works when there are multiples, see below)
  • Get grow light and figure out environment / lighting set up with photocells; where the light is positioned and how the sensors respond to presence of a user
  • Make additional soil moisture sensors and do all soldering / waterproofing (heat shrink etc)
  • Glue things into breadboard and make sure everything is sturdy


  • Figure out if we want to do a drainage system or not
  • Waterproofing system for the inside of the box (possible solutions include caulking the inside, polyurethane coating, fixing plastic to the wood, buying thicker plastic liner)
  • Buy additional soil and plants
  • Affix hooks for headphones to the outside of the box
  • Make housing for electronics and mount to side of box
  • Sand box exterior and remove stains; finish with tung oil or wax
  • Figure out configuration for displaying the laptop alongside planter


  • Establish how sound maps to multiples of the same sensor
  • Refine and finalize soundscape


  • Finalize and submit winter show application (due Wed Dec 3)
  • Make refinements that address any remaining user experience problems / issues