OPA1678: single end to difference circuit matched with ADC5140

Hi Amber,

Amber Si said:

When I input 2Vrms sine waveform, output waveform will be distortion (THD is about 1.6%), if I reduce input to 1.6Vrms, distortion will disappear,

The circuit gain is approx. 1.36V/V in U24, where input of 2Vrms is approx. 2.83Vpk will generate 3.85Vpk output swing. It seems to be in the operating range. 

One suggestion, I'd open up your differential LPFs, which was 2kΩ/2.7nF or 29.5kHz range (1kΩ and 4.7kΩ at the output will also attenuate the output swing). what value is C313 = 272, 2.7nF? The output seems to be stable based on the AC response (though I can check for the phase margin later). 

OPA1678 E2E 03062025.TSC

In addition, you did not mention what load you are driving. If the output current exceed the rating below, the distortion is possible. Please let me know. 

Best,

Raymond

Hi Raymond，

Raymond Zhang1 said:

value is C313 = 272, 2.7nF?

Correct,C313 =2.7nF

I use opa1678 to drive TLV320ADC5140(I mean the audio data will be sent into ADC5140) I also found ADC5140 AC input impedance is 2.5kohm/10kohm/20kohm in datasheet(this is TI device, you can check datasheet)

do you think ADC5140 input impedance is too low and cause the distortion? 

BR

Amber

Hi Amber, 

What types of ceramic capacitors are these used in the circuit?

For audio application, it is recommended to use NP0 (C0G) capacitors. X7R and Y5V ceramic capacitors are not recommended. If wrong capacitors are selected, there may be a distortion in the circuit. 

Amber Si said:

When I input 2Vrms sine waveform, output waveform will be distortion (THD is about 1.6%), if I reduce input to 1.6Vrms, distortion will disappear, can you help check if  there is something wrong with OPA1678 circuit?

THD of 1.6% is huge. What is the THD at 1.6Vrms?

There may be something you have not described in the system. The system is using +/-5Vdc for the power. What types of power power LDO and current ratings on these? What are your load impedance? I need to have additional information to figure out what it is going on. 

Amber Si said:

I also found ADC5140 AC input impedance is 2.5kohm/10kohm/20kohm in datasheet(this is TI device, you can check datasheet)

There are 22uF ceramic capacitor in front of ADC5140 and it should be ok for the low end (HPF). What did you configure to from the ADC setting. 

Best,

Raymond

Hi Raymond,

Raymond Zhang1 said:

For audio application, it is recommended to use NP0 (C0G) capacitors. X7R and Y5V ceramic capacitors are not recommended. If wrong capacitors are selected, there may be a distortion in the circuit. 

For NP0 capacitor, customer already use it on C313 and C314 which is low capacitance, but for other capacitor(22uF in circuit), they don't have NP0 which is a little high capacitance

Raymond Zhang1 said:

THD of 1.6% is huge. What is the THD at 1.6Vrms?

Yes, THD is 1.6% at 2Vrms, very huge, and THD go down to 0.05% at 1.6Vrms.

Raymond Zhang1 said:

There may be something you have not described in the system. The system is using +/-5Vdc for the power. What types of power power LDO and current ratings on these? What are your load impedance? I need to have additional information to figure out what it is going on. 

They use our Buck TPS62933 (24V to 5V)and our DCDC TPS65131(5V to +-5.5V), I think power current for OPA1678 is very small, about 10mA? I attach schematic about power part, you can find TPS62933 at page1 and TPS65131 at page2, do you think there is something wrong with power supply? 

Partyband(TI)_MAIN_SHC V0.1.pdf

Raymond Zhang1 said:

There are 22uF ceramic capacitor in front of ADC5140 and it should be ok for the low end (HPF). What did you configure to from the ADC setting. 

Now is 2.5Kohm AC impedance of ADC, and I'm trying to increase impedance to see if it can improve THD.

Hi Amber,

Amber Si said:

TPS62933 at page1 and TPS65131 at page2, do you think there is something wrong with power supply? 

Yes, something is wrong with your switching power supply likely. 

I do not know if the switcher is designed for DCM or CCM mode. In other words, I need to know what is the minimum and max. current that the switcher is designed for. If the switcher is designed for CCM, then the poor THD is likely your issue. 

Amber Si said:

There are 22uF ceramic capacitor in front of ADC5140 and it should be ok for the low end (HPF). What did you configure to from the ADC setting. 

In addition, I need to know what is the switcher's cutoff frequency. This also determined what high frequency components that the system has on the THD noises at the Audio output. 

For precision audio application, you should always use LDO for the application in general. Again, you never mentioned what is speaker or headphone's output impedance. I need to know  what is the OPA1678's output current. 

Best,

Raymond

Hi Raymond,

Raymond Zhang1 said:

I do not know if the switcher is designed for DCM or CCM mode. In other words, I need to know what is the minimum and max. current that the switcher is designed for. If the switcher is designed for CCM, then the poor THD is likely your issue

I think TPS62933 is in DCM mode because load current should be <100mA, do you mean TPS62933 in CCM mode will lead to poor THD?

In addition, customer add LDO to reduce supply power ripple, now opa1678 supply power tree as below:

Raymond Zhang1 said:

Again, you never mentioned what is speaker or headphone's output impedance. I need to know  what is the OPA1678's output current.

Let me clarify opa1678 circuit, input is wired microphone, and output will be ADC5140+BT SOC+ amplifier +speaker(3ohm/4ohm), you can see below circuit, do you want to which impedance?

For efficiency, I'll book your calendar to have a meeting. Thanks.

BR

Amber

Hi Amber, 

Amber Si said:

In addition, customer add LDO to reduce supply power ripple, now opa1678 supply power tree as below:

Please check the following in +/-5Vdc supply rails. The supplies' ripple voltage has to be reduced to approx. 1mVpp for precision audio application. For instance, 5.000Vdc ± < 1mVpp and limit the high frequency noise <5kHz (prefer < 1kHz). 

I do not know what LDO that customer has selected. If the 24V and 5V switcher are not able to reduce its switching PWM frequency below below 5kHz or lower (cutoff frequency poles in switcher), LDOs are likely unable to attenuate further via its PSRR. In other words, the customer's supply rails are too noisy and full of harmonics when load current is increased. 

I see that the customer is using OPA1678 to drive the ADC, which it does not have significant load variation, even with input of 2Vrms. 

To proof my points, you may attach scope probe at +5Vdc and -5Vdc in LDO's output, and compare the DC supply voltage rails at 1.6Vrms vs. 2Vrms, you should observe significant increases in DC supplies' ripple voltages. If not, the issues are at the post stages, shown in the diagram below, since the speaker is low impedance and I am not sure that the driver is able to handle the dynamic load when input audio amplitudes are increased. 

Which point did you measure the THD distortion at with THD about 1.6% or 0.05%?

If you have other questions, please let me know. 

Best,

Raymond

Hi Raymond,

Thanks for the meeting time, as we talked on webex, customer remove circuit of interface as below, THD can be improved. Pls help dig out why that circuit affect THD and correct customer if there is something wrong with the circuit.

I'll work with customer to check power supply noise issue and get back to you also, thanks.

Attach SCH

3554.Partyband M_EV1(ADC5140) SCH.pdf

BR

Amber