INA217: Hum noise in Pre-amplifier INA217

Hi Lakshmanan,

eventually you created a hum loop. Dissapears the hum when you disconect the microphone from the XLR-Plug?

Kai

Hi Lakshmanan,

I suspect that by tying the USB GND to the signal GND from the XLR connector you've created a ground loop which is picking up a line-frequency hum. One way to resolve this would be do place a large resistance in series between your XLR pin 1 and USB_GND_CODEC, then bypassing this with a capacitance. This should help to bleed off DC charge from the signal while still maintaining AC coupling. 

Hi Kai,

  When I disconnect the microphone I hear more of a hum noise.

Thanks & Regards

Lakshmanan V

Hi Alex Davis,

I placed 1M ohm resistor parallel with 1000pF capacitor between the XLR pin 1 and USB_GND_CODEC . Still hum noise is present in audio signal. This hum noise is serious issue in the board. How to eliminate this hum noise?

Hi Lakshmanan,

if the hum still exists without having anything connected to the XLR-plug, then there's is a serious fault in your circuit.

1M is a bit high. Can you repeat with 100R? This should already heavily decrease the hum, if there is hum loop.

Kai

Hi Kai & Alex Davis

I connected 100 ohms parallel with 1000pF capacitor between XLR pin1 and USB_CODEC_GND

Now I am hearing hum noise in audio signal. After removed mic from XLR connector I am hearing more Hum noise in Audio signal.

This is the serious issue in this board

Hereby I attached Audio codec schematic pdf , PCB Top & Bottom pdf for your reference.

C17145B PCB TOP.PDFC17145B_PCB BOTTOM.PDF

C17145B_AUDIO CODEC SCHEMATIC.pdf

Thanks & Regards

Lakshmanan V

Hi Lakshmanan,
I notice that you're providing power to the INA217 and OPA137 directly from the USB connector. While both the INA217 and OPA137 have good PSRR specifications, there will be some noise from the USB supply that bleeds through into the output of these amplifiers. It's also possible, given that OPA137 has a worse PSRR specification for its negative supply, that noise from the LTC1046 is being coupled through.

Could you provide more information on the hum you're experiencing? If it's a line-frequency (50 or 60Hz) hum, then there could be noise being picked up somewhere in the design. If it's a higher-frequency hum, I'd suspect noise from the power supplies.

Sorry...did not see the -5V. I rescind my last post....

But do short the inputs and see if the hum goes away.

Hi Lakshmanan,

there are some heavy mistakes in your layout: You use a charge pump to generate the -5V supply voltages. Charge pumps are known for creating heavy noise. At the same time you omit a solid ground plane but run signal ground and power ground along narrow copper traces. As if that wasn't bad enough, you have a very small microphone signal in this circuit which you amplify particularily high. And to beat all this, the circuit is a mixed analog digital design. This cannot work.

First, you MUST tremendeously improve the signal ground routing by providing a solid ground plane. Then, you should add low pass filtering at the input and output of the LTC1046 charge pump. But also the +5V USB supply voltage can be contaminated with noise. So, you should add some RC low pass filtering to both the supply lines of INA127.

The total lack of solid ground plane also causes that the digital circuitry with its noisy ground currents inject additional noise into the analog signal chain. So, the hum could also come from the digital signalling of PCM2900. The only remedy is to heavily improve the signal ground routing by adding a solid ground plane.

Kai

Hi Lakshmanan,

To redeem myself...

There are a few things to try to pinpoint where the noise is getting in. Break the system down into stages and eliminate each section at a time.

The first one is to short the two XLR_AUDIO inputs together. Just bridge the jack pads with a screwdriver. If the hum goes away, then the input is the source and it is most likely because the inputs are floating and picking up noise. But I don't think that is your problem...

Noise on the inputs would common mode and should be mostly cancelled-out. My guess it is coming in from the reference line somewhere since noise on the reference would be directly reflected in the output.

My guess is that the noise is coming in from the servo loop integrator ("LPF Filter" U2) since there is a high impedance node there (pin 4 C7/R9). The caps in the low pass filter could also be sensitive.

Try quickly shorting the PRE_OPAMP_REF line to ground and see if the noise reduces or goes away. This will tell you if the noise is coming in through the reference input.

Next try quickly shorting the output of the INA (pin 7) to GND. If the hum is still present, then the source lives upstream from the INA.

Next would be to short the outputs of the filter amps. Remember these are at Vs/2, so don't short them for long. If the hum goes away, then the filter amps are the source.

If the hum is still there, then it is the CODEC. Check the grounding and the reference line. Do note that the 2900 CODEC is not recommended for new designs and may be obsoleted soon... I noticed you connected the analog ground of the codec to a trace that goes off to the GPIO pins. That could be a source of ground noise. It should be grounded to the same plane as the signals.

Does the hum change as you move the board around in the X, Y and Z planes? What if you put it in a grounded coffee or paint can?

Do you have any other equipment connected to the board? Oscilloscopes and line-powered test equipment are notorious ground loop sources. Does the hum go away if the board is "floated" in space with just the USB cable (and input shorted)?

Is the hum present when powering from a bench supply vs a desktop or laptop computer on batteries? Computers can also have dubious grounding.

Without touching any grounds, move your finger around the board a few mm above the components to try to find a spot where the hum gets louder. This will reveal the sensitive nodes that are susceptible to capacitive coupling.

Is this a 2 or 4 layer board? Make sure you did not inadvertently isolate a section of ground plane.

Hi all,

   We are giving the Power to USB Audio codec board through USB adapter.

Following test case are done in this board without presence of mic input.

1. XLR Connector differential inputs ( Pin No 2 and pin No 3 ) are shorted. 80% of hum noise gets canceled and I heard very less noise only.

2. Pre-amplifier output pin is shorted to ground and removed the short between XLR Connector differential inputs.In this case also 80% of hum noise gets canceled and I heard very less noise only.

3. Then Pre-amplifier REF pin is connected to ground. In this case, I am hearing hum noise

4.Then We shorted both op-amp output in LPF built by OPA2353UA. In this case, I am hearing hum noise

Right now we don't have pad to mount RC Low pass filter to filter the +5V & -5V Power supply noise

In this board XLR input section is more sensitive.

This board is Two layer PCB.

Please tell me the solution to eliminate this hum noise.

Hereby attached the USB Audio codec schematicC17145B.PDF

Thanks & Regards,

Lakshmanan V