• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 49 subscribers
  • Views 393 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Original question:

AMC1311: Temperature Error ~3% instead 0.29%

AMC3330: AMC3330

christophe duchemin
Prodigy 50 points
Part Number: AMC3330
Other Parts Discussed in Thread: ISO7740, ISO7820, , AMC1350, AMC1302-Q1, THS4561, AMC1302, AMC1311, ISO224

Dear Ray Vincent ,

On AMC familly and regarding the excel tools 'Isolated Amplifier Voltage Sensing Excel Calculator Rev B' Im questionning about the ibias compensation resistor r3' that should be equals to (r1+r2)//r3 in order to compensate the created offset. From my calculation and understanding of this problem I think that positive current should be closed to negative current in order to remove this offset error. But, when i read the datasheet, I understand that Inb could be 10nA while Inn is -10nA (and vice et versa) and this can compromise the ibias compensation offset. I'm right?

Another subject we discuss internally (on also on iso7740, iso7820...) is the lost of hvdc (high voltage) isolation.

I guess that the lost of isolation is very rare but right now I have no data to quantify that (FIT rate and FDM). Is there any failure distribution mode value on lose of isolation on these devices?

I found FIT rates of these devices on your web site.

Best regards,

Christophe Duchemin

  • 0
    TI__Mastermind 47625 points

    Hi Christophe,

    I support the AMC family of devices and can help clarify the purpose of the R3' resistor. The main reason this additional resistor is necessary is if a low-input impedance isolated amplifier is being used for sensing a high impedance node, such as the AMC1300B with 22kohm RinDiff sensing across a 10kohm sense resistor. As the AMC3330 has a high input impedance, this additional compensation resistor should not be necessary. Additionally, the input bias current is negative both both the positive and negative input pins. 

    This application note explains in further detail: https://www.ti.com/lit/an/sbaa350a/sbaa350a.pdf

    Please let me know if you have additional questions in regards to the AMC family of devices.

    For assistance with the ISO7xxx family of devices, I recommend posting to their forum. 

  • 0 in reply to Alexander Smith
    Prodigy 50 points

    Hi Alexander,

    It's clear that R3' is fully neccessary for AMC1300B with -30µA. R3' could also be interesting for AMC3330 https://www.ti.com/lit/ds/symlink/amc3330-q1.pdf?ts=1675665858164&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FAMC3330-Q1%253FkeyMatch%253DAMC3330-Q1%2526tisearch%253Dsearch-everything%2526usecase%253DGPN, but in this datasheet the bias current (Iib) could be positive or negative and the offset could be also negative or positive. With your last comment it could perhaps make sense to add R3' to improve a little the accuracy only if the input offset current is smaller than specified (Iio). Is Iio really +/-10nA?

    For the second part of my message, it is also for AMC familly.  I guess that the lost of isolation on AMC is very rare but right now I have no data to quantify that (FIT rate and FDM). Is there any failure distribution mode value on lose of isolation on these devices?

    Thank you for your support and have a good day

  • 0 in reply to christophe duchemin
    Prodigy 50 points

    I forget...

    On AMC1350 datasheet, I do not see input bias current specification. Is this normal?

  • 0 in reply to christophe duchemin
    TI__Mastermind 47625 points

    Hi Christophe,

    Apologies for the miscommunication - I did not realize we spec +/-10nA for AMC3330 Ib as opposed to only -10nA. According to the systems engineer the bias current from the AMC3330 should be virtually zero. R3' with AMC3330 is indeed interesting, but it's a very small error (38uV) compared to gain error (4mV), gain drift (9mV) or offset drift (800uV). Bias current error will also be no longer considered after an offset calibration and I would be more concerned with sense resistor drift over temperature of both R3 and R3'. https://www.ti.com/lit/zip/sbar013

    Comparing input structures as shown in the datasheets (I cannot go deeper on the public forum and/or without an NDA) we've made some changes that warranted the removal of the input bias current spec from the AMC1350 datasheet as it's no longer significant. 

    As for FIT, FMD, FMA, we have this document for the AMC1302-Q1, which is the +/-50mV input device similar to the AMC1300Q-1: https://www.ti.com/lit/fs/sbaa447/sbaa447.pdf

    We do not currently have a similar document for the AMC33xx-Q1 family of devices. To the best of my knowledge, we have yet to experience a loss of isolation in any of our devices - if significant over-voltage is applied they are designed to fail-open: https://www.ti.com/lit/wp/slyy081a/slyy081a.pdf

  • 0 in reply to Alexander Smith
    Prodigy 50 points

    Hi Alexander,

    I'm tryind to find the smallest error of my analog inputs with AMC1350 (as it seems very hard to find AMC3330 parts on the market). 

    The issue with AMC1350, is a bigger offset and lowest Rin impedance. Could I put a differential amplifier like THS4561 with 125k/50 gain and VOCM=0=HGND at input of AMC. My concern is to verify the input common mode issue as described in the datasheet: 'A DC current path between INN and HGND is required to define the input common-mode voltage'). (input voltage range is +/-230mV).

    Best regards,

  • 0 in reply to christophe duchemin
    TI__Mastermind 47625 points

    Hi Christophe,

    As offset error is input referred in the datasheet, the output error is the same for both devices.

    Additionally, offset errors can be easily removed with a 0V offset calibration in the MCU. Is this an option for your design? 

    What level of accuracy are you trying to reach?

    AMC3330, gain of 2 = +/-0.3mV * 2 = +/- 0.6mV. 

    AMC1350, gain of 0.4 = +/-1.5mV * 0.4 = +/-0.6mV. 

    If not, you could consider adopting a circuit configuration similar to this and use the AMC1311 instead of the AMC1302:

  • 0 in reply to Alexander Smith
    Prodigy 50 points

    Hi Alexander,

    This is the schematic I think about with AMC1350:

    AMC1311 may be a good candidat for low ranges as I won't  need to change the gain for low ranges. Thank you for this reference.

    I target an accuracy of 0.5% FS and offset calibration is not possible.

    BR

    Christophe

  • 0 in reply to christophe duchemin
    TI__Mastermind 47625 points

    Hi Christophe,

    I see, thank you for the clarification. This configuration would provide the highest accuracy but require a charge pump supply.

    The ISO224 may be an interesting device as well. The front end is similar to the AMC1350 but with a +/-12V input and is our highest precision device.