Team Vision for System-Level Design PhaseOur team vision for this phase is to come up with multiple good concepts for both transmissive and reflective sensors. To do so, we will be prototyping various concepts to prove the feasibility of said concepts. Along with these prototypes, David Malanga's V2 PCB will be completely populated and prepared for testing. This will all be completed by the design review.
Two techniques were used for the functional decomposition of our project: the Function Tree and the Transformation Diagram. Each method has their pros and cons. For example, the Function Tree follows a logical process of uncovering functions by asking the questions "How?" and "Why?", but it does not account for interactions between the main function and its end goal. The Transformation Diagram, however, does show a clear interaction between functions, but it is harder to show levels of detail.
The main function of our project is to reproduce string vibrations as an electrical signal and transmit it to sound. Through our Function Tree, we identified 4 main functions that we would need to complete this:
- Visualize String Vibrations
- Filter Noise/Create Effects
- Support the Components
- Supply Power
Supporting the components specifically refers to how the phototransistors and infrared transmitters will be mounted on the guitar.
The Transformation Diagram goes into more detail on how these functions interact, such as the effect and importance of sensors that can be adjusted by the user and what systems/functions the 9V battery will be powering.
The traditional guitar pickup is the magnetic pickup. In recent years, an active magnetic pickup has been used to lower the dampening effect on the strings. These pickups were chosen for benchmarks in the first phase, as we were determining what to base our development off of.
A more accurate benchmark of what we are working off of would be the latest RIT revision of the optical guitar pickup, or the Opto column in the figure above. This pickup used a transmissive setup where the photo-detector and infrared transmitter were on top and bottom of the strings. The board was designed to sit over the bridge of the guitar and have the sensors directly attached to the board. This non-embedded PCB design and sensor location led the guitarist to need to hover over the pickup to not hit it, which hindered play-ability.
The previous design contained the following specs, which we look to improve upon for the next revision.
- Transmissive Sensor Type
- Non-Embedded PCB
- Bridge Pickup Location
- Battery Not Attached to guitar
- No effects
- Surface Mounted sensors
- Analog and Digital paths
- No casing
- No adjustments
The two tables above also serve to compare multiple infrared emitters and photodetectors. Only through hole parts were considered since they were evaluated to be the easiest mounting type to work with when compared to surface mount parts and light pipes. This was decided through the use of the Pugh Chart.
While most of the electrical characteristics are essentially the same, the main issue will be the physical size of the infrared emitters and photodetectors. Smaller parts will allow for smaller casings around to parts to mount them to the guitar, which should ultimately result in the best possible playability of the guitar for the end user.
Feasibility: Prototyping, Analysis, Simulation
Question of Feasibility - Reflectivity of StringsWill the reflective pickup have enough reflection to replicate the string vibration into sound?
- Nickel Wound/Plated Strings
- DSP will be able to process sound if signal can be taken from this design
- Guitar String and Guitar
- Power Source
- Infrared (IR) LED and Photodetector
- Align IR LED and photodetector under guitar string based on setup from morphological chart
- Hook up power source to IR LED and Photodetector, hook up photodetector with oscilloscope to track what the photodetector sees
- See if photodetector is detecting IR through Oscilloscope
- Pluck string. See if the oscilloscope shows drops in voltage, as the string should be reflecting the IR to the photodetector
- Find a clear threshold from the oscilloscope capture to determine '1s' and '0s'
- TBD. Test needs to be run
Question of Feasibility - Tuning FrequenciesWill specific string frequency filtering work for multiple tunings?
- Filtering should allow the string frequencies from multiple tunings to pass through
- Could cause problems if the pass band allows other strings to be read by the wrong sensor
- According to David Malanga's Graduate Paper "The Optical Guitar Pickup: Converting Light to Sound", the existing Version 2 PCB and its components should allow for multiple tunings. Below are pictures of the low E string's fundamental frequency shown on a spectrum analyzer as well as the Bode Plot of the pass band implemented by the analog components on the board. According to the Bode Plot, lower tunings should be feasible since the pass band extends from just above 0Hz to about 1kHz. As for the higher strings, this pass band would simply shift to the right to accommodate the higher frequencies.
- Also, string interference should not be an issue since the digital signal processing algorithms should properly filter each string frequency regardless of which sensor picks up the vibration. Thus, no testing needs to be done for this part according to the previously collected data.
Question of Feasibility - Adjustment Method PrecisionWill the adjustment methods be precise enough such that the play in the adjustment does not interfere with the working of the sensors? Per previous designs, a movement as small as 1mm can cause the sensors to become misaligned. Assumptions
- Purchased parts will conform to specified tolerances
- Printed parts will not suffer from serious imperfections
- Fabricated parts will conform to ±.002 in
- Parts will not greatly wear down over time, adding more play to the system
- CAD software (PTC Creo 5)
- Draft up adjustment method with needed tolerances for motion and manufacturing
- Add up clearance in each direction, make sure that each direction adds up to less than ±.5mm (±.019685 in)
- If clearance is too much, make adjustments to method to tighten clearances (ex: add dowel pins to locate from instead of only using screws)
- Create CAD model of adjustment method and verify results
- If needed, 3D print/fabricate parts
- Currently, two adjustment designs have been proven to be well under the required clearance. More designs will be analyzed
Question of Feasibility - DSP ImplementationHow will digital distortion DSP algorithms be implemented?
- Input signal will be band limited before it is sampled
- Sampling is fast enough to prevent aliasing to the band limited
- Basic Clipping distortion will be simulated
- Sound clip to run through the algorithm
- Digital amp simulations use distortion algorithms, so by bench marking, this is possible
- Simulating a real time distortion algorithm should be possible using matlab
- Use the bilinear transform to make a discrete transfer function
- Pending simulations
- Matlab simulations should be a similar process to programming the pic micro-controller
Concept Selection Criteria:
- Design Feasibility
- Playability, Ease of Use
- Standard Pickup Size
- Ease of Installation
Seven concepts were created by putting together the sub function concepts from the morphological chart. The sensor type/position concepts will be tested at a later date to determine their feasibility.
- Design Feasibility
- Playability, Ease of Use
- Standard Pickup Size
- Design Time
- Ease of Installation
These criteria were created based on the time provided to complete the project, as well as improvements from the last version of the pickup. The concepts chosen for design will show these improvements.
The concepts we have created each use various design ideas from the morphological chart. See the Concept Development section to see that many of the ideas developed during this phase are interchanged between concepts.
Each concept is explored in a pugh chart. The first few concepts relate to the sensors themselves, and both their electrical and mechanical configurations. Unfortunately, our IR LEDs and Phototransistors did not come in on time to perform a feasibility test for this, so the reflective designs remain untested.
The next few concepts are related to the playability and ease of use for our design.
The last few concepts relate to the overall electrical system and the DSP that will be creating sound effects for our design.
Below is the Pugh chart comparing the concepts by our concept criteria. This is one version where the datum is the original design by David Malanga. The document with the datum rotated to each concept is located Since the feasibility is not finished, our concept selection is based on our current knowledge, and will change as feasibility is completed. here
The overall system will have the following basic architecture:
Furthermore, the signal path of David Malanga's v2 PCB can be seen below:
Designs and FlowchartsDesign and flowcharts are going to be completed in the next phase. This cannot be done until proper feasibility analysis is completed and configurations are chosen.
Design Review MaterialsOur Systems Design review will be held October 18th at 12:00pm in 09-4425 at RIT. Include links to:
Plans for next phaseTeam Goals:
- Complete feasibility
- Sensor configuration and parts selected
- Finish populating board
- Preliminary mechanical designs for pickup casings
- Preliminary software designs for DSP
Individual 3 week plan can be found here
To see more details on the team's plan, see this excel sheet.