This week the team met again on West Campus to conduct some more testing on the CODEC and DAC output circuit.
Unfortunately the team suffered their first casualty on Tuesday. While preparing for the team meeting, John noticed a funny smell and tiny wisps of smoke coming from the breadboard.
One dead CODEC
The CIRRUS CODEC he was working on had begun to cook. Luckily the team had anticipated such tragedies and purchased 2 extra chips. John had already prepared one back in January, so the team wan able to conduct the planned testing.
Before replacing the fried IC with a brand new one, John checked every point and took voltage and current measurements. The board was pulling around 600mA of current after it failed. This is about 10X the stated maximum!
After about 30 minutes of observation, John noticed a tiny strand of wire sitting neatly between pins 8 and 9 on the board. These pins correspond with the digital power and digital ground. This was likely the cause of failure.
Once the CODEC was replaced with a fresh chip, John took extreme caution and monitored the currents of each supply to make certain nothing was exceeding the stated limits. After around 20 minutes of this he decided that the issue was indeed the tiny short and that it was safe to move on.
Using Hunter's thermocouple, temperature measurements of the chip were also taken for extra verification of proper operation.
MCLK from the DSP
LRCLK and SCLK
Closeup of SCLK
LRCLK and SDATAOUT
Noise from the DSP infects all other signals. LRCLK (yellow) and static preamp output (blue). 20nS horizontal division
Preamp output without DSP attached
Preamp output with DSP attached to circuit
ADC input (yellow) to positive DAC output (blue)
ADC input (yellow) to inverted DAC output (blue)
Testing Throughput of the entire system from Preamp to ADC to DAC to differential amp
+ADC output to differential amplifier output
-ADC output to differential amplifier output
The differential amplifier output successfully filters the DC offset and combine the differential DAC output signals.
As a quick test, peak-to-peak voltage measurements were taken at the differential amplifier output. From 50 - 300Hz the output remained at 3Vpp. Further testing will be conducted to verify an accurate frequency response.
Agenda Moving Forward:
troubleshoot DSP generated noise issue
test throughput from MEMs microphone to DAC output circuit
test power amplifier
establish DSP throughput and test with basic DSP programs
test DSP with ANC algorithm using small-signal analogue setup
Update and edit report
Update Website
John
After the team meeting on Tuesday, John began researching solutions for the DSP-induced noise that was observed.
Texas instruments has an application note that details methods of EMI elimination in mixed-signal circuits. After a bit of reading, he decided that adding a separate 5V supply for the digital signals would be a possible solution for the noise issue.
5V supply for digital devices
Noise on Preamp output signal is significantly reduced.
John also began learning to program the DSP using sigma studio. He successfully compiled a simple program that controls an onboard LED.
This simple program serves as verification of computer-to-dsp communication using the sigma studio software.
John Also began constructing the final device enclosure. This wills serve as the prototype for conducting tests.
A metal enclosure was chosen for shieling/grounding purposes. This fuse box will house the power supply, preamp, adc/dac, dsp, power amp, and arduino. The transducers and microphones will be separate. A 5-pin XLR connector will be used to carry the mic and transducer signals.
Hunter
This week, Hunter updated the “weekly minutes with the professor” page.
Next weeks meeting has been canceled so our team can focus on the project during spring break.
Hunter also updated the third lesson learned that the team discovered when working on the power supply, where we noticed the difference a shielded vs. unshielded inductor can make when EMI can effect a units performance, more information can be found on the lessons learned page.
Worked on the Arduino display that starts the program by displaying a starting screen that shows the name of the project. Some troubleshooting was required, as the original program was not working well with the opening screen that was added to the program. The problem when switching to the next screen can be seen in the short clip below.
After some troubleshooting, the issue turned out to be in the loop of the program. The updated version, shown below, now has the opening screen with project name that is delayed for a few seconds before the screen changes and starts reading the input from the assigned pin.
Added capacitors with better power and voltage ratings to the power amp, shown on the right side of breadboard (the left side is the DAC output connections). This will lessen the strain on these components and some of the surrounding components making the circuit safer and help with heat reduction, which also prolongs the life of the circuit.
Set up master clock on DSP and started practicing with SigmaStudio to get more familiar with the software for GPIO mapping and programming the DSP. [1]
AnalogDevices ADAU1462 datasheet:
GPIO Mapping in SigmaStudio, AnalogDevices:
References:
[1] Daumemo, “How to program an analog devices DSP?” Daumemo, 08-Mar-2021.
[Online]. Available: https://daumemo.com/how-to-program-an-analog-devices-dsp/. [Accessed: 19-Nov-2021].
[2] AnalogDevices, “SigmaDSP Digital Audio Processor,” ADAU1462/ADAU1466 datasheet, 2017-2018
Comments