LPCXpresso LPC1769 board: why DAC is so noisy?

[lpcware]

Content originally posted in LPCWare by giusloq on Tue May 19 08:46:11 MST 2015
I'm using LPC1769 development board (OM13000) without a base board. I'm trying to use the DAC peripheral, but I have a problem.

When I set a 0 value, no problem. When I set 0x3FF (max value), the output value changes accordingly, but I see a noise signal with a scope. It seems it has bursts at about 20ms with a peak-to-peak level about 40mV.

I noticed the same noise signal is present on 3.3V rail too. I'm supplying the board through USB.

Is it normal? I'd like to use DAC for audio signal and this type of noise will be audible on the speaker.

I tried to change the clock divisor and DMA timeout for DAC peripheral, but the noise is the same.

[lpcware]

Content originally posted in LPCWare by giusloq on Thu Apr 28 05:37:43 MST 2016
It is very strange that the analog supply (generated from digital supply filtered with a RF choke) is more noisy than digital supply. I think I have missed something on the hardware connection.

This is how I connected the DAC output to LM386.

[lpcware]

Content originally posted in LPCWare by giusloq on Fri Apr 22 09:07:25 MST 2016
I finally created my own custom board, with a LM386-based audio amplifier that drives a small magnetic speaker.

When no audio is generated, I set a constant value to DAC (medium value), but I can hear some noise from the speaker.
If I configure the DAC output pin as GPIO output, the noise disappears at all.

What could be the problem? The analow power supply and VREF pins are connected to VDD through an RF choke.

[lpcware]

Content originally posted in LPCWare by LabRat on Tue May 26 01:00:09 MST 2015

---

Quote: giusloq
Yes? I'm not using PHY at all... how can I swith it off by software?

---

You have to init it and then set the Power Down bit in Basic Control Register.
By default it's not powered down and consuming 30-40 mA  :O

---

Quote: giusloq
In order to disconnect the PHY, I think I can remove FB2, right?

---

Yes  :)

[lpcware]

Content originally posted in LPCWare by giusloq on Tue May 26 00:19:05 MST 2015

---

Quote:
This 20ms signal is a burst signal created by Ethernet PHY  Puzzled

So switch off PHY via software or disconnect it...

---

Yes? I'm not using PHY at all... how can I swith it off by software?

In order to disconnect the PHY, I think I can remove FB2, right?

[lpcware]

Content originally posted in LPCWare by LabRat on Mon May 25 10:07:15 MST 2015

---

Quote: giusloq

---

Quote: starblue
20ms corresponds to 50Hz, so you probably have a problem with noise from the power grid. (Maybe a ground loop?)

---

I already thought about this and I used DC battery without USB. The problem is always present.

---

:D

Nice idea, but unfortunately complete nonsense in this case...

This 20ms signal is a burst signal created by Ethernet PHY  :~

So switch off PHY via software or disconnect it...

[lpcware]

Content originally posted in LPCWare by giusloq on Mon May 25 07:54:57 MST 2015

---

Quote: starblue
20ms corresponds to 50Hz, so you probably have a problem with noise from the power grid. (Maybe a ground loop?)

---

I already thought about this and I used DC battery without USB. The problem is always present.

[lpcware]

Content originally posted in LPCWare by giusloq on Mon May 25 07:54:10 MST 2015

---

Quote: uLipe
So, you said the 3.3V rail supply is noisy too? Are checked with a scope the stage before your 3.3V supply? The issue can be there.

---

Battery 12V - 7805 - LPCXpresso board that generates 3.3V rail with another linear regulator. If I unplug the LPC board, the power supply rails (12V and 5V) seem good. When I plug the LPC board, I see the noise on 5V and even on 12V rails! How is this possible?

---

Quote: uLipe
Another tip is to load the DAC properly with the value based on datasheet (if I am correct is about to 1KOhm). Feed you DAC signal to a well designed low pass filter in order to remove the sampling spectral component, a butterworth with order about 2 can be sufficiente depending on your sample rate.

Can you post the DAC firmware section here? let-me remember you the DAC peripheral has a settling time which must be satisfied too, if you change the DAC value too fast (even if you writing the same value) some oscillations can ocurr until DAC enters in steady state.

---

I'm not sure, but I think the problem is with power supply noise and not on the DAC output itself.

[lpcware]

Content originally posted in LPCWare by starblue on Tue May 19 10:05:36 MST 2015
20ms corresponds to 50Hz, so you probably have a problem with noise from the power grid. (Maybe a ground loop?)

[lpcware]

Content originally posted in LPCWare by uLipe on Tue May 19 09:31:02 MST 2015
Hello giusloq, let me see if I can help you.

So, you said the 3.3V rail supply is noisy too? Are checked with a scope the stage before your 3.3V supply? The issue can be there.

Another tip is to load the DAC properly with the value based on datasheet (if I am correct is about to 1KOhm). Feed you DAC signal to a well designed low pass filter in order to remove the sampling spectral component, a butterworth with order about 2 can be sufficiente depending on your sample rate.

Can you post the DAC firmware section here? let-me remember you the DAC peripheral has a settling time which must be satisfied too, if you change the DAC value too fast (even if you writing the same value) some oscillations can ocurr until DAC enters in steady state.

I hope that helps.

Felipe.