AMC1035: Signal impedance, input impedance and settling time

Hi Antoine,

One other big difference between the AMC1035 and the AMC1306 is the fact that there is no isolation provided with the AMC1035.  Do you need isolation in your application?

Hi Tom,
Thank you for your questions. No I do not need any isolation. Independent whether I do or do not need an isolation I would like to know more about the design of the analog side and the limits in term of impedance/settling time.

OK - Thank you!

I just wanted to point out the isolation aspect between the two parts. 

With the higher input impedance of the AMC1035, you are correct in that it maps up to voltage measurements at the bottom of a resistive divider.  There will be longer settling time on the sample/hold cap with the AMC1035, but like with the AMC1306, you could still setup two different filter paths with two different OSR's - one higher for accuracy and one lower for faster step response.  That being said, I've not come across anyone using the AMC1035 for current measurements in this way so I don't have any immediate 'this is what you could expect to see' detail to share.  On the accuracy side, are you scaling this +/-200mA to meet the +/-1V input range of the device?  What is the nature of your your original signal?

We do have two different digital filters in the processing unit. However, I am dealing with the analog part and I want to avoid signal damping/distortion due to bad impedance balance. Although the AMC1035 were not designed for current measurement and we do not have example for that, analog engineering/modelling shall make possible to understand the physical limits due to the input impedance. However, I would not know how to do it correctly. Here the input analog front end is explained Designing with delta-sigma ADCs: system design considerations to optimize performance | TI.com and the settling time is only mentioned but not more about the modelling. Do you now better know what I mean ? 

No, the scaling comes from a current sensor. It scales it down by a factor 5000 and creates the isolation barrier needed between HV and LV. So the nature of the orginial signal is a current, and the signal I get from the sensor is also a current. To meet the voltage range of the ADC I just choose the correct resistor value.

Antoine

Hi Antoine,

From that video link, go down to the 'Modulator Sampling and Analog Front End' video for a better understanding on how the AMS1035 works - this device only has the modulator portion of the full-on converter that Ryan described in the initial video.  The input impedance though is not directly related to fmod as it would be in the AMC1306.  I've asked the design staff if they have any analog modeling details that can be shared, including the useful bandwidth of the AMC1035 input.  I'll let you know what I hear back.

Hi Tom,

Thank you ! In the meanwhile I followed the design steps from [FAQ] Delta-sigma ADC anti-aliasing filter component selection - Data converters forum - Data converters - TI E2E support forums. However, I do not expect any overvoltage so I just chose Rflt = 10Ohm. I attached a screenshot from the schematic. Note the Harness output OUT delivers a current from -200mA to 200mA. Maybe I need to add a further resistor in serie to the shunt resistors to make the amplifier of the current sensor work in its linear/accurate range (waiting for manufacturer confirmation).
fMOD = ~10Mhz
OSR = 128
fDATA = ~100kHz
Sinc3 filter
Input signal bandwidth: 25kHz

Do you see an issue or a No-Go ?

Antoine

Hi Antoine,

I don't see your schematic snippet, however, from your description - I do not think you will have any issues.