Group Instrument Project

Boyce, Hong, Noon and Sears: (Sensor Work)     Instrument Project             10.10.05

The concept began as a series of fifteen acrylic disks mounted with Qprox (capacitive) sensors. The direction, based upon the standard 88 key musical range would move up and down full octaves by switching the appropriate keys, the remaining 13 keys would be the standard notes for an octave. It is also our intent to enable users to both learn and excel through the use of LEDs beneath the keys, through instruction, or simply interaction.

The board was wired with a capacitive sensor: Quantum Technologies  QT-113h

See Page 288-89 in Physical Computing
 As well as the Quantum Website (http://www.qprox.com)

Step one: We soldered on a small nickel sized piece of copper mesh, this was intended to increase the area of the sensor. Upon doing so we immediately realized that the standard insulation on the wire did not shield sufficiently. In the process of this conclusion we also tried a smaller aluminum mesh. As well as shorter wires and various shielding attempts. In all cases if someone was holding  the shield it would not work. 

Step two: We moved to a coaxial cable, however this proved to be a too extreme approach as no readings were acknowledged by the sensor save direct contact. Initially we did have a two-foot piece of coaxial cable, however the problem remained with a six-inch piece. Additionally we did work with our capacitors to insure our accuracy but with the coaxial cable we were unable to resolve the issue.

It was recommended that we:                        Switch from the QT-113h to the QT-160h
Increase our capacitor size
                                                            Increase the screen contact

                                                            Use a smaller shielded cable

 

PROTOTYPE DESIGN

Today’s meeting went exceptionally well. The immediate concerns of the group were, first, materials. After our initial design we had an idea of what we needed, so this time we made a list.

Then we began discussing the design of the actual interface. We had an idea, or at least a direction from the very start. The layout of the twelve keys would follow a resonance pattern out from the user, we felt that this would make it not only a new instrument, but one that was intuitive to anyone who has played music before. For those that have not we feel it will be equally simple. Through the discussion we also liked the idea of challenging the three dimensional plain rather than have just a flat linear interface, such as with the piano.

As we discussed different options we added and subtracted from our materials list.

The prototype design of this interface will have the keys angled at roughly 20 degrees outward from the user. We also liked the idea that these sensors would be mounted on top of plastic tubing. These tubes would then go through two layers of lexan spaced roughly two inches apart, like a test tube rack, for stability and strength, to our ground floor where the hardware would be located. This ground floor will be enclosed and removable via four bolts attached at the corners to the upper two floors.

Additionally we discussed the possibility of having our boards made for us once the design is formalized for their ease of use and obvious ready-made benefits. Through discussion it was also brought up that we might want to include a few option keys on the upper floor.

Our final stage of discussion, stemming from the idea of interface options, involved the addition of a possible LCD screen which would further enable the user, through an actual score for sight-reading of notes, as well as interface options such as key and chord dynamics. This final step we realize is beyond our next step, but the process has begun.
                                                                                               
October 18, 2005

Today we ordered and received the Plexi pieces that we initially discussed, however in the process the design changed so we need to order more. The Canal Plastics experience was good as we had an hour to survey the various options on the street in terms of materials and it was quite beneficial. Vincent and I both focused on varying sound amplification devices, which could work to our benefit.

 

October  20, 2005

PROTOTYPE

Today we finalized the Illustrator File for laser cutting the prototype. The design has changed a bit on the past four days as development ensued and new design opportunities were realized. We also today ordered out 18 RGB LEDs for the project as well as 18 Digital Potentiometers (AD5204BN10)

Today’s struggle involved the fact that our project requires parts that are not readily stocked by most online stores. After ordering with next day air the set from Analog Devices we found out that the “ Next Day” term was used loosely as the “next day” was when the parts arrived in California from the Philippines. Digikey wouldn’t have the parts for two weeks with rush shipping. Our lead time created the biggest problem.

Eventually we found that Arrow Electronics in NJ actually had ten in stock and we ordered them all. We also ordered from Analog ten more with an undetermined delivery date as they were samples.

In the Laser cutting stage adjustments had to be made because the size of our plastic tubing was less than the dowel we used for the mock-up. What has also changed is that instead of a square base we will be using a curved bent piece, viewable online, to mirror the contours of the prototype key pad.

Throughout the course of the day we met whenever possible, and received input from everyone in the community we could find that had experience with both the Qprox as well as the DigiPots. It was also brought to our attention that the RGB LEDs might vary if we use different sources so we ordered one shipment.

The laser cutting is scheduled for tomorrow, day six, and then we wait for materials. In the meantime we have each set about the task of familiarizing ourselves with the datasheets of our various hardware parts.

As we near our final design stages we have also begun the microscopic adjustments of fine design as we all agree that this project has the potential to be great. Our enthusiasm is at it’s peak and we are very excited to get to work,

 

October  23, 2005

PROTOTYPE
The goal for today was to begin wiring as much as we could with the materials that we had. We wired fifteen Qprox QT113 today to fifteen separate boards. We also made two sample keys in the fashion intended. We cut and wired the Qprox copper wire mesh field, as well as wired a copper wire ground surrounding the field to eliminate field interference. We also wired LEDs into the sample keys. This process took  six hours.

During the course of this work we determined that due to time constraints we won’t be able to realize all that we hoped in the actual presentation part of the piece. The tubing which is scheduled for the laser cutting tomorrow, will be mounted to a piece of cardboard, rather than the intended plexi-glass. It was decided that due to our deadline our concern about the materials was not as important as the actual programming and hardware of the piece.

 

October  24, 2005

PROTOTYPE

In the morning we brought the files to the laser cutter and began the process. There was a problem, as the laser was cutting our plexi tube caught on fire. We immediately began making plans to find a way to get more tube as soon as possible. As it worked out we didn’t get the new tube until five o’clock. By that time however we had the new Illustrator CS file ready to go. Because we were tied up in the lab we missed multiple shipments, which had been over-nighted to us.

This delayed the process of our implementation because we needed, and were waiting patiently for the digital potentiometers. Specifically the AD5204. We also missed two other shipments for sample Qprox. We spent most of the morning doing damage control with UPS and FedEx.

The work would have to wait until tomorrow on those items.

The afternoon consisted of a strait session of more than twelve hours in the Pcomp Lab fabricating. The boards were wired but we needed to fabricate the fifteen keys. We had tested the qprox in their eventual housing, as well as adjacent to a LED with no trouble, so we set to it.

Once the keys were made we then wired all the RGB LEDs. Additionally we fabricated the stand-in base out of cardboard. At 2:00am we retired for a few hours. 

October  25, 2005

PROTOTYPE
The morning began again with the Laser cutter, as we needed to create the key posts destroyed yesterday.

Today we received the digital potentiometers after we went to UPS to pick them up. At that point however the concentration was setting up our prototype. We didn’t really get started, due to our classes until late. The process took five hours. We did learn that in the fabrication one of our keys was shorted. Once the final positions were set we had lost a few more. However the majority worked and our prototype was a success. 

November 8, 2205

PROTOTYPE

Today we beagn discussing the possibility of a redesign of the project. It seemed like a risky move seeing how we basically had just barely enough time to finish the first, however we decided to go ahead, while simultaneously finishing the first. It was planned that over the next two weeks before Thanksgiving Break we would finalize the first project's design, and order custom circuit boards for both. The decision was reached without too much difficulty and we set out. Nick Sears took it upon himself utilizing his computer engineering skills to design the boards for both projects, while I completed the first prototype, Anne Hong set out on creating the new piece.

This divergence was due mainly to an unsettling lack of faith in regard to our interface design and the limitations found within it. It was, and remains, my belief that we can succeed, though we are branching out from here.

November 18, 2005

SONIMO

Today the first prototype design was completed. due to a lack of time, and a very busy laser cutting facility we utilized our own on floor shop for all of the Fabrication. This was not ideal, though few would question the legitimacy of the results. As a mirror to the keys themselves we have mounted the instrument in a cylindircal tube, it is our thought that this new design, instead of the proposed platform will allow for better light transmission and a clearer, more refined interface.

We have changed our sensors as per numerous reccomendations from Tom Igoe to the QT-160, use of this sensor is rather political as some believe in it firmly while others wouldn't use it at all. We unfortunately have the time to test these chips so if they don't work we are in the position of complete failure, a notion we refuse to entertain.

December 4, 2005

SONIMO

Today we soldered all of the lights, and sensors to the board, over the break, and during my adsence which occured last week, Nick and Anne mounted all the chips to the boards. Today we connected the hardware. Immediately it was apparent that we had miswired a few of the lights.

December 5, 2005

SONIMO

Once the instrument was wired we could program the chips on the board. The code was relatively the same as what we had for the mid-term project a month ago, though the process was not easy at all. In the eventual arrival of our presenation period we were able to get only three of the QProx sensors to work at all. This was only after cutting all of the ground wires from the board and frantically trouble shooting the problems. Time did not allos for us it's graces so this project has been a race against the clock for it's entire span. Once our presentation was over, due up again next week, we could spend more time going through and over each of the 72 seterate wires which may or may not be correct.

December 7, 2005

SONIMO

Today I actually reviewed all of the wires, and save the cut ground wires, and the many that were dislodged during our trouble, none of them seemed to be cross-wired, which is both a good and a bad problem. In the event that we have to replace any of the QProx sensors the entire board has to be detached from the instrument due to their very small size. What is left to do is to go through each of the leads to make sure they are behaving correctly. It is our hope, though we have so far been unable to confirm this, that we have overlooked some small detail in the process and we don't have to start again at square one, or at least the start.

On the other front, the Qprox sliders which are the entire functionality of the second design have not functioned at all in our tests, this is a very large problem as without that functionality we basically have wasted a large amount of resources and time in this endevour. We remain confident though that confidence is waivering as our final deadline approaches.

December 10, 2005

SONIMO

Today further reviewed the wiring. Then as this avenue was exhausted and fullfilled we moved into the Lab for work in the code. It became immediately apparent we had drastically mis-labelled most of our circuitry. After mutilple hours of exhaustive reserach we postponed this torture until later.

December 12, 2005

SONIMO

Today we discovered and defined our label mess which drastically had disrupted our entire process. By testing the circuits with the ocilliscope we determined that none on the clocks on the QProx sensors were working at all. Nick then, through his trouble shooting experience determined we needed to remove the Flux from the board. At the time we had only 70% Alcohol Solution, we let it soak for two hours. Once we had dried out the board with the heat gun we resumed testing and low and behold it was working! We went from a single key to ten, the remaining two we determined, weren't working because the flux remained. As the alcohol evaporated away, we guessed, the keys faded away. It was at this point that we decided upon an Acetone bath.

(THIS PROCEDURE IS NOT ALLOWED IN THE FACILITY AT ITP)

So I brought it home for the night and put it on my fire escape in the West Village. It is our hope that this will at least eliminate this variable from our equation. Our problem at this stage is that we are having trouble assigning the specific keys. Only one key actually lines up with it;s LED RGB command code, the others do not. This is our goal for the morning. 24Hrs to go. . . .

 

December 13, 2005

SONIMO

Today we continued with our washing of the flux, after a twelve hour bath in the acetone flux remained on the board. At this point I went out and got a paint brush to finely wipe off each chip with alcohol. After a few hours of that I decided to go for a pressure wash with water. This seemed to do the trick, though it did take awhile for the board to dry even with the heat gun. Once it was dry I then rewired all of the grounds to the board and capped the connections with hot glue. It seems that we have ten keys functioning but it is hard to tell as we have lost connection with a number of wires during this process.

VIBRATO

While this was going on Nick and Anne were beginning to assemble the second prototype. The QProx sliders which we are using have worked wonderfully well and I am forced to admit that our evolution within this project was for the best. The fabrication process is deathly slow but thatis to be expected with a perfectionist designer. We logged nine hours on the housing along, before we got to the programming stage. Despite a rather rocky middle I think our group has preformed quite well. It is difficult to spend so very much intensive time on a project and feathers were ruffled here and there along the way. I myself believe I took the entire project too seriously and too personally where I should have had the very element I thought was lacking which was faith in the project. It has been a tremendous learning experience from start to finish.

 

December 14, 2005

Sonimo and Vibrato

Today with two instruments to de-code and a presentation in only a few short hours we attempted to right the wrongs we have encountered along the way. I will say that we could have managed our time a bit better but that is not always possible, as it wasn't in this case. The projects both have come a long way, and luckily they weren't entered in the show to save us the embarrassment of hardware manlfunctions that would be nearly immpossible to control.

In retrospect the process could have taken a different timeline as we did not get to test properly and did not get sufficient user feedback before we started what are now our final designs.