Hi,
On the power supply lines of my Teensy 3.6 I noticed a strong switching noise which is related to the sampling frequency of the attached audio shield. There are a lot of peaks in the spectrum of the power supply lines. The lines in the spectrum are harmonics, a distance of (samplingrate / 512) or (samplingrate / 128) apart. The strongest peaks are about 45dB above the noise floor.
Fig. 1. samplerate = 44117, 5V
Fig. 2. samplerate = 44117, 5V
dF = 348 Hz =~ 44117 / 128
Fig. 3. samplerate = 44117, 5V
dF = 88 Hz =~ 44117 / 512
Fig. 4. samplerate = 22050, 5V
Fig. 5. samplerate = 22050, 5V
dF = 174 Hz =~ 22050 / 128
Fig. 6. samplerate = 22050, 5V
dF = 44 Hz =~ 22050 / 512
Fig. 7. samplerate = 88200, 5V
Fig. 8. samplerate = 88200, 5V
dF = 174 Hz =~ 88200 / 512
All figures are on 5V, similar on 3.3V. Oscilloscope samplingrate = 10MSa/s. Attached to the Teensy 3.6 are a ST7735 display, and a Si5351 clock generator.
On my Teensy 3.5 I did not notice this, although I must add that I did not check the power supply lines with my Teensy 3.5, which I do not have anymore. This switching noise was discovered after I switched from 3.5 to 3.6 and found that this switching noise is also present on the Line In input of the audioshield, although not as strong.
My Teensy runs at 180 MHz, compiled with fast code. Power supply is external, power supply pads between USB and external on the underside of Teensy have been cut. The external power is supplied through pins GND and Vin. At first I noticed it in my own sketch, but it is also clearly visible using one of the example sketches (I used MemoryAndCpuUsage). Of course in the example sketch I only tested it with a sampling rate of 44117 Hz.
I am very interested to know if other people have also noticed this, and if it is also present on Teensy 3.5?
Have fun,
Koen
fig. 1
fig. 2
fig. 3
fig. 4
fig. 5
fig. 6
fig. 7
fig. 8
On the power supply lines of my Teensy 3.6 I noticed a strong switching noise which is related to the sampling frequency of the attached audio shield. There are a lot of peaks in the spectrum of the power supply lines. The lines in the spectrum are harmonics, a distance of (samplingrate / 512) or (samplingrate / 128) apart. The strongest peaks are about 45dB above the noise floor.
Fig. 1. samplerate = 44117, 5V
Fig. 2. samplerate = 44117, 5V
dF = 348 Hz =~ 44117 / 128
Fig. 3. samplerate = 44117, 5V
dF = 88 Hz =~ 44117 / 512
Fig. 4. samplerate = 22050, 5V
Fig. 5. samplerate = 22050, 5V
dF = 174 Hz =~ 22050 / 128
Fig. 6. samplerate = 22050, 5V
dF = 44 Hz =~ 22050 / 512
Fig. 7. samplerate = 88200, 5V
Fig. 8. samplerate = 88200, 5V
dF = 174 Hz =~ 88200 / 512
All figures are on 5V, similar on 3.3V. Oscilloscope samplingrate = 10MSa/s. Attached to the Teensy 3.6 are a ST7735 display, and a Si5351 clock generator.
On my Teensy 3.5 I did not notice this, although I must add that I did not check the power supply lines with my Teensy 3.5, which I do not have anymore. This switching noise was discovered after I switched from 3.5 to 3.6 and found that this switching noise is also present on the Line In input of the audioshield, although not as strong.
My Teensy runs at 180 MHz, compiled with fast code. Power supply is external, power supply pads between USB and external on the underside of Teensy have been cut. The external power is supplied through pins GND and Vin. At first I noticed it in my own sketch, but it is also clearly visible using one of the example sketches (I used MemoryAndCpuUsage). Of course in the example sketch I only tested it with a sampling rate of 44117 Hz.
I am very interested to know if other people have also noticed this, and if it is also present on Teensy 3.5?
Have fun,
Koen
fig. 1
fig. 2
fig. 3
fig. 4
fig. 5
fig. 6
fig. 7
fig. 8
