Whine Noise

Status
Not open for further replies.

Smokedparts

New member
I have built an SDR amateur radio transceiver based on the work of a ham down in New Zealand. I have not included the code as it belongs to the chap in New Zealand

The initial build was with a Teensy 3.6 and the radio produces an awful whine noise. Switch to a Teensy 3.5 and most of the whine noise disappears. The Arduino sketch is the same for both and the only hardware change is the Teensy. There was a concerted effort to address grounding and avoiding ground loops. To further assure nothing was slipping through the common power supply I installed an active power supply decoupler in the leads to the Teensy/Codec Board

Is there a difference in the structure that prevents the use of the 3.6. I have included a short video link so you can hear the whine.

The sketch was loaded using a Windows 7 machine and the Arduino IDE was 1.8.3. The hardware used was the Teensy, your Codec board, the Adafruit Si5351 PLL Clock generator and a 128X64 Black and White OLED (both using the I2C).

The front end hardware was built by me and was tested using Power SDR installed on a Windows XP machine and an M Audio Delta 44 Sound card. There is no evidence of any whine being generated by the front end board. Again most of the whine is gone with the Teensy 3.5 installed.

OLED's can be noisy and I removed the OLED display and there was no change in the Whine sound.

Any suggestions would be most appreciated.

Smokedparts

;)

 
I also have a radio using a Teensy 3.6. I noticed that there is some EMI coming from the Teensy or the connecting wires at specific frequencies. Also, the frequencies where I get noise changes when I use 48KHz audio vs 44.1KHz. When I moved to a PCB for my design, the noise level was reduced, and was not affecting my antenna gain especially when the I2S clocks are "ON".

Basically, any wire that carries I2S or SPI or I2C data will create this interference, as they are in the Mhz range. If you can output your SNR levels to a PC using a serial connection, check the SNR levels as you have them now, and then disconnect your OLED(removing the wires totally from the board) and turn off the I2S clocks, and then get fresh SNR levels.
 
It's more than likely that your circuit with the T3.6 is radiating some signal at the "right" frequency, and some other part of your system picks it up. I don't know the emission levels of the Teensys, but it's probably your wiring, not the T3.6. The simplest thing you could try is to change the clock frequency of the T3.6 to see if that makes any difference.

If that doesn't help you could try so some shielding / RF filters at the right places. If you have access to a spectrum analyzer that can be a big help. RF design can be tricky.
 
Thanks for your reply and suggestions. I did try some of your suggestions such as completely removing the OLED and its wiring -- I have a 4 pin header that provides that connection so it was just unplugging a cable. The leads to the header are about 1 inch from the codec board.

Quite by accident it was discovered that there is a tremendous reduction in whine noise if you power the Teensy 3.6 from a computer or one of these 5 VDC battery packs. It was astonishing.

So then reading the placard that comes with the Teensy there is a note that says that the jumper connecting Vin to Vusb must be separated if using a battery charger or external power. I missed that note and that somehow seems connected to the issue as there is whine noise with an external supply but very little whine when connected to the USB port on a computer or a battery pack.

I guess I just don't have a full grasp of the Teensy wiring --the note begs the question why have it jumpered in the 1st place.

I will keep working on it. For now partially solved with at least two power sources. Guess I should cut the jumper and see what happens. Beofre I do that I want to know the why?

smokedparts

I also have a radio using a Teensy 3.6. I noticed that there is some EMI coming from the Teensy or the connecting wires at specific frequencies. Also, the frequencies where I get noise changes when I use 48KHz audio vs 44.1KHz. When I moved to a PCB for my design, the noise level was reduced, and was not affecting my antenna gain especially when the I2S clocks are "ON".

Basically, any wire that carries I2S or SPI or I2C data will create this interference, as they are in the Mhz range. If you can output your SNR levels to a PC using a serial connection, check the SNR levels as you have them now, and then disconnect your OLED(removing the wires totally from the board) and turn off the I2S clocks, and then get fresh SNR levels.
 
Wow, I just ran across this. Hooked up the audio adaptor, have the trace cut so I can power by battery with the USB plugged in, and a nice horrible high pitched whine in the headphones.

Is there a solution to this?

I'm using an SPI LCD display, and will have a bunch of sensors hooked up. If I unhook the SPI LCD, the whine goes away. I'm not sure exactly how to deal with this. I'll try moving the pins for the SPI interface to the far end of the board. But I'm going to have most of the pins hooked up.
 
Wow, I just ran across this. Hooked up the audio adaptor, have the trace cut so I can power by battery with the USB plugged in, and a nice horrible high pitched whine in the headphones.

Is there a solution to this?

I'm using an SPI LCD display, and will have a bunch of sensors hooked up. If I unhook the SPI LCD, the whine goes away. I'm not sure exactly how to deal with this. I'll try moving the pins for the SPI interface to the far end of the board. But I'm going to have most of the pins hooked up.

The solution is classic RF design: Shielding, ground planes, careful PCB layout, and so on...
 
The solution is classic RF design: Shielding, ground planes, careful PCB layout, and so on...

:)rolleyes: The software guy looks around nervously, to see if there's any actual engineers hiding somewhere in the room that might be able to help.)

Kidding, but ... any pointers? I've got the audio shield on the bottom of a Teensy 3.6. I'm going to be attaching it to a cut-to-fit breadboard, and putting the whole mess in a 3D printed case (that is shaped like a recorder / bagpipe chanter). Things will be packed in, but I'm happy to put whatever I need to in amongst it all. What might one use for shielding in a one-off thing like this? :p
 
Don’t be shy, show us photos of how everything is actually wired up.

Often these sorts of problems involve ground connections...
 
Thanks Paul and Theremingenieur. That helps me understand that a lot of it may have to do with the crappy wiring on the protoboard.

I will return to this when I get it on the actual protoboards and off the breadboards! After the replies here, I'm pretty sure things will change as I change up the wiring; even touching and wiggling the wiring changes the noise. And moving the DC pin for the SPI from pin 1 (right under the headphone jack) to 28 changed the noise noticeably.
 
(Embarrassed) Update: I hadn't soldered AGRND properly.

While trying to figure out why the volume knob wasn't working, I found this thread (https://forum.pjrc.com/threads/44762-Audio-Shield-volume-controller-ground-problem ) which pointed to AGRND not being connnected. This also trigger an "oh I wonder ..." thought related to the whine.

Soldering that reduced the whined dramatically, so the original poster should check that.

Note that on https://www.pjrc.com/store/teensy3_audio.html the picture of the board that shows which pins are used doesn't include AGRND, which is probably why I didn't solder it, but the circuit diagram shows it.
 
Status
Not open for further replies.
Back
Top