How to reduce power supply noise

Status
Not open for further replies.
I guess I would start by disconnecting 3v3 of the teensy to the audioboard bit power the audioboard from a separate 3v3 regulator.

The audioboard is designed to connect directly to the teensy, but keep the 3v3 power pin out and use a low noise LDO with recommended capacitors and see houw much that helps. I was one of the biggest noise sources in a ultrasound circuit we made.

Edwin
 
In case anyone is examining the schematic, the Teensy actually uses an LDO to generate it's 3.3V, but since that supply is feeding all the digital circuits on the Teensy, the loads induce noise on the LDO output, so you need another 3.3V LDO off the 5V supply to create a clean 3.3V just for analog.
 
this is very interesting, can anyone recommend a good information resource on LDO usage? which part is teensy using, which LDO would be best suitable for audio application in general? supporting circtuitry needed?

i am using 2 seperate big linear +/-12v PSUs in a big hybrid digital/analog system currently, with the basic eurorack 5v transistors, 3 different MCU platforms, loads of OLED and ws2812 and a bunch of analog audio... while not too bad i am still struggling with some digital noise bleeding.
 
If you feel you hear digital noise it is not so very much important what voltage regulator you are using. It is more important to run the audio circuitry from a different voltage regulator than the digital circuits.

If you are using the teensy audio board you might be tempted to place wires on each connection.
When you do that you will connect the 3v3 of the Teensy to the 3V3 of the audio board.

Just be aware of that and create a clean 3.3 volts to power the audioboard and audio related circuits.

To advise a voltage regulator it would be nice to know the circuit or at least the current it uses and the source voltage. I am sure some people here have ideas about a very low noise LDO but since you tell you hear digital noise, I think the type of LDO wil not make a big difference.

Kind regards,

Edwin
 
Power circuit for my teensy synth..

Link: https://forum.pjrc.com/threads/63255-New-Teensy-Synth-quot-Shruthi-2-quot?highlight=Teensy+Shruthi

MCU-Jeannie.jpg

AMP-Jeannie.jpg

Fx-Jeannie.jpg

Jeannie-Contrl-JPG.jpg

Greetings from germany. Rolf
 
It looks like you have a nice project going on there, nice video's on youtube....

In the schematic I see failrly low values of capacitors on the outputs of the voltage regulators, specially the LDO's usually require larger values on the output side.
The 3v3 for the audio chip gets its input voltage form a 5V regulator which is supplying the Teensy, why not put the input on the 9V?
(I see max vin is 7V, so maybe a test with an extra 5V regulator or change to an other 3V3 regulator...)
If there is a ripple on the 5V you might see it in the 3v3..... Or try larger caps also in the input side/output of 5V

I also like to use a larger cap over the 3v3 of the Teensy. I had the feeling the DAC output of the teensy sounded a lot better when I used an additional 470uF on the 3v3 output of the teensy. Maybe that can help you too.

In an application with a sensitive microphone we have a resistor in every dataline we used. The display update noise was quite bad, after using an 150 ohm resistor in every dataline of the display and audio codec it was way much better.
Since you only have sound output I guess it will probably not help you a lot, but hey its worth the try...
The resistors limit the current rush when switching to high or low level so there will be less overshoot ans waveform will look al little less square so don not overdo this by using higher value resistors. Something in the range of 100ohms seems to work well enough.

Kind regards,

Edwin
 
Hi Edwin, thanks for the feedback. I've tried to answer all your questions comments below.

"In the schematic I see failrly low values of capacitors on the outputs of the voltage regulators, specially the LDO's usually require larger values on the output side."
- the values are the ones recommended in the datasheets for those particular devices. The 9V LDO is only supplying the opamp so ripple is minimal so the datasheet spec'd value is used. The 3.3V LDO is supplying the codec also uses parts spec'd in the datasheet, but also note the codec already has 10u bulk tantalums on all the rails.

"The 3v3 for the audio chip gets its input voltage form a 5V regulator which is supplying the Teensy, why not put the input on the 9V?"
With LDOs, the heat produced is proportional to the difference between the input voltage and output voltage, and the load current. So you always want to feed LDOs with the lowest input voltage possible for normal operation to ensure the regulator doesn't need heat-sinking.

"I also like to use a larger cap over the 3v3 of the Teensy. I had the feeling the DAC output of the teensy sounded a lot better when I used an additional 470uF on the 3v3 output of the teensy. Maybe that can help you too."
If you are using the Teensy onboard DAC, having good supply filtering would be essential and the large cap might help if the Teensy is pulling a lot of current dynamically. . However, on the TGA PRO, the Teensy is digital only. All analog circuitry is handled by the CODEC which has dedicated/isolated anlog supplies. Any ripple on the Teensy 3.3V will not have any impact on ripple on the analog supplies they they are not derived from it, they derive from the 5V USB.


"In an application with a sensitive microphone we have a resistor in every dataline we used. The display update noise was quite bad, after using an 150 ohm resistor in every dataline of the display and audio codec it was way much better."
Note on the TGA Pro schematic all digital lines between the Teensy and the Codec have 33 ohm source-series resistors to reduce EMI / reflections from fast edges.
 
Hi Blakaddr,, sorry, I was only trying to get people to think, apparently I missed some parts looking over the schematic.

The IRU LDO specifies an output capacitor with a minimum of 10uF, the text suggest to use larger values. There is a 47uF cap on the audiochip, but it is part of an LF filter so not directly on the IRU1117.

I never looked at the IRU1117 before, it seems that the AMS and LM 1117 can handle higher input voltages.

Edwin
 
Hi Blakaddr,, sorry, I was only trying to get people to think, apparently I missed some parts looking over the schematic.
Edwin

Hi Edwin, no, I think I made a mistake. My previous post was on a phone via email link and I thought the comment was directed at me (misread, my fault = apologies), but I think they were intended for Rolfdegen. Your advice is good, I designed similar aspects to your suggestions into my board as noted above. The schematic info I linked to earler can by used by anyone wanting a design reference for Edwin's good comments.
 
Status
Not open for further replies.
Back
Top