PGA970: Noise in phase measurement

Hello Alessandro,

I don't believe this is a known problem with the the PGA970 itself. This seems like it is stemming mainly from the input to the ADCs. Are you able to measure the input signals from the LVDT to see how they are responding when the sensor is near the center position? 

Can you share a schematic and some details on the sensor you are using?

Regards,

Hi Scott,

I used the PGA970 in the 4-wire configuratio. In this case, the excitation output (P1-2) drives the primary coil of the LVDT, while the secondary coil il read-out using ADC1 with common mode activated at 1V.

In this configuartion, when the position of the core of the LVDT is at full-scale in a direction (let's assume left), the signal has maximum amplitude and phase equal to about 0. instead, when the core is moved completely in the other direction (right) the amplitude is again maximum, but with opposite phase (about 180°).

I verified that data coming from the PGA970 correspond with this situation. To do it I calculated atan(DEMOD1_PH1_DATA/DEMOD1_PH2_DATA) when  DAC_SIN_NDS1 = 0 and DAC_SIN_NDS2 = WAVEFORM_TABLE_LEN and everithing works fine.

The problem arises when the core approaches the middle position. In that case the aplitude at the output of the LVDT is minimum, so I suppose that it is more difficult for the phase calculation circuit to extract a stable phase value. I downloaded a steam of values from PGA970 keepng the core in the middle and I observed that values of DEMOD1_PH1_DATA and DEMOD1_PH2_DATA continuously jump below and above 0 due to a superposed noise. As a result I get a very high phase noise, since the phase continuously jump between 0° and 180°. The same behavior is not observed at different core positions, since there is no misunderstanding between positive and negative values.

Do you have any suggestion to solve the problem?

I thank you for your time.

Best regards,

Alessandro Cominelli

Part Number: PGA970

Hi,

I recently asked another question about noise in the determination of the phase of the LVDT output. I have few other questions about phase demodulation and noise associated to it.

In principle phase demodulation gives both in-phase and quadrature components of the input signal with respect to the excitation, namely A*cos(phase) and A*sin(phase), so it is possible to get the amplitude measurement using those values, by simpling adding the values quadratically and extracting the square root of the result: A=sqrt((A*sin(phase)^2+(A*cos(phase)^2).

1) I tried this operation, but the result differs from the one I get at the output of the amplitude demodulator. Is there a simple scale fator between the two measurements? Which is its value?

2) Moreover, I obseved that the noise superposed on the measurement of amplitude carried out starting from the phase demodulator is much higher than noise at the output of the amplitude demodulator. To do it I fixed the position of the LVDT nearto the center and I extracted measurements in free running, then I observed the percentage rms value.
I expected similar noise values, since the analog input signal is the same one and the ADC is shared between amplitude and phase demodulator, but it seems that phase demodulator is much more noisy. Do you know the reason of this effect? Is there a way to keep noise low?

The high noise in phase measurements reflects in high noise superposed on the phase. This is really a critical issue in my application, since I need to determine if the LVDT is moved left or right with respect to the center and this opearation is done observing when phase is higher or lower than 90°. Since phase noise is much higher than amplitude noise it is really an issue to clearly determine the LVDT position.

I thank you in advance for your help.

Best regards,

Alessandro Cominelli

Hi Alessandro,

I will look into the scaling factor of the amplitude demodulator and get back to you. I will also have to do some investigating on the phase demodulator and why it may show more noise than the amplitude demodulator. This may take some time, but I'll give you an update by the end of this week.

Regards,

Thank you, Scott.
I look forward to your updates.
Best regards,
Alessandro Cominelli

Inviato dal mio dispositivo Huawei

-------- Messaggio originale --------
Oggetto: Sensors forum: PGA970: Phase noise
Da: Scott Cummins
A: mmwave_sensors_forum@mail.e2e.ti.com
CC:

Hi,

Is there any news about noise and scale factor in the phase modulator?

Thank you,

Alessandro

Hi Alessandro,

I'm working with a designer to better understand the scaling internally, and it is taking some extra time. I will get back to you by Tuesday with more information. Sorry for the delay.

Regards,

Hi Alessandro,

There should not be a scaling factor difference for the phase demodulator, so this behavior is a little strange. Is there any way you can share some of the data that you are seeing which show the differences in the noise between the amplitude and the phase demodulators? How large is the difference?

Regards,

Hi Alessandro,

I also joined your other thread regarding the phase noise to this one, since the topic is the same, and this way all of the information will be located in the same place.

Regards,

Hi Scott,

please find a little report attached, where I give you some insight about results of my measurements.

I wait for news.

thak you in advance for your time,

Best regards,

Alessandro

Report Phase measurements with PGA970.pdf

Hi Alessandro,

Thank you for the detailed report. I am reviewing this with some colleagues that have more intimate knowledge with the internal design of the PGA970 and I will get back to you as soon as possible.

Regards,

Thank you, Scott.
I wait for news.
Regards,
Alessandro



-------- Messaggio originale --------
Oggetto: Sensors forum: PGA970: Noise in phase measurement
Da: Scott Cummins
A: mmwave_sensors_forum@mail.e2e.ti.com
CC:

Hi Alessandro,

Thank you for your patience. I have some more investigation to do and I will be getting back to you this week with some more information.

Regards,

Thank you.
I wait for you answer.
Regards,
Alessandro

Inviato dal mio dispositivo Huawei

-------- Messaggio originale --------
Oggetto: Sensors forum: PGA970: Noise in phase measurement
Da: Scott Cummins
A: mmwave_sensors_forum@mail.e2e.ti.com
CC:

Hi Alessandro,

To answer some of the questions from your report. The phase demodulation output rate is just set to be output at the same rate as the amplitude modulation.

There isn't an analog or hard-coded scaling factor that is different for the phase demodulator vs the amplitude demodulator, although the additional filtering in the amplitude demodulation change may account for this difference at some point. The phase demodulator output will also need to be normalized anyway (in your firmware) to be able to refer to the waveform generation output and properly determine the phase

Regards,