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.

AMC1311-Q1: DC Voltage Sensing for energy metering

Part Number: AMC1311-Q1
Other Parts Discussed in Thread: AMC1311, AMC1301, , AMC1300B-Q1, AMC1301-Q1, AMC1302-Q1, AMC3302, AMC3330, ADS131M04

Hello Everyone,

We are building a high precision DC energy meter for EV Fast charging stations.

DC voltage range:- 150 V to 1000 V DC

DC current range:- 100 A(Maximum)

For Voltage Sensing:- I am planning to use AMC1311 B along with resistor divider with 0.5% tolerance resistor. Is it possible to get less than 0.1 % error in voltage measurement with this method even after calibration???

For Current Sensing:- I am planning to use AMC1301 along with high precision shunts . Is it possible to get less than 0.1% error in current measurement with this method after calibration???

Thanks & Regards

  • Hi Biswajit,

    yes, this is possible.

    Use the full linear input voltage range of the chips which is -0.1V...+2.0V for the AMC1311B and -250mV...+250mV for the AMC1301.

    Keep the lower resistor of the voltage divider in front of the AMC1311B smaller than 1G / 10,000 = 100k and keep the shunt resistor in front of the AMC1301 smaller than 18k / 10,000 = 1.8R.

    Why a factor of 10,000?

    A factor of 1,000 for achieving less than 0.1% deviation and another factor of 10 to provide some headroom because the specified input resistance values of the chips in the datasheets are only typicals.

    Use 0.1% / 15ppm resistors instead of your planned 0.5% resistors because of enhanced temperature and long term drifts. Keep the surface temperature of these resistors below 40...50°C in order to minimize the long term drifts. Higher surface temperatures will cause way higher long term drifts! Remember Mr. Arrhenius Relaxed

    Kai

  • And take a look at the input bias currents of AMC1311 and AMC1301. To keep the impact on precision minimal furtherly reduce the resistor values by a factor of 10 finally resulting in <10k and <180mR.

    Kai

  • Hi Biswajit,

    Yes, both of these measurement accuracy targets are possible, however there are several factors to consider. Here are some measured results for a +/-85A sweep that meet these targets. Please note that only offset calibration was performed and the shunt was not subjected to changes in ambient temperature: https://www.ti.com/lit/an/sbaa464/sbaa464.pdf

    Here is an excel calculator that will help you calculate the typical and worst-case error expected for the devices. 

    https://www.ti.com/lit/zip/sbar013

    Three important questions to clarify are:

    1. What is your temperature range?

    2. Are your error targets worst-case or typical? 

    3. What type of calibration will be performed? 

    For your voltage sensing circuit, I recommend using AMC1311B-Q1 over AMC1311-Q1 as it has slightly higher performance. 

    For your current sensing circuit, the amperage is quite high and you must consider the tradeoff between maximizing the FSR of the isolated amplifier and the power dissipation in the shunt resistor. As Kai pointed out, I would recommend keeping the resistance as low as possible. This will help minimize the power dissipation in the shunt resistor. If the power dissipated in the shunt resistor is too high, it will begin to drift rapidly. 

    4. What is the nominal (typical) current range to be measured? 

    5. Do you require 0.1% accuracy at maximum current as well? 

    I would not recommend using AMC1301-Q1 as it is an older device. I recommend either the AMC3302 or AMC1302-Q1 as they have a smaller input range of +/-50mV.  Another possibility would be to use a device with a larger input voltage range such as AMC1300B-Q1 (+/-250mV, newer generation P2P with AMC1301-Q1) and an amplifier on the front end between the shunt and the isolated amplifier. This would allow you to use an even smaller shunt, reducing the power dissipation even further. An example of this configuration is shown in this TIDA: 

    https://www.ti.com/tool/TIDA-00445

  • Thanks Alexander for your reply!

    I am using the AMC excel sheet you have shared for voltage sensing! But the problem is the sense resistor keeps on varying as the DC voltage varies. I have a wide DC voltage range. 150 Vdc to 1000 Vdc. I was wondering if i will be able to meet 0.1% error criteria in this range.

    See the thing is we are targeting a 0.5% error on the over all DC energy meter so essentially all errors will add up.(Up to first approximation) i.e. Voltage sense error + Current sense error + Energy meter IC error(0.1% already known) so we are targeting the other two error (Voltage sensing & current sensing errors)to be less than 0.1%/0.15%

    To answer your queries

    1)-40 to 85 degree total range...Operating could be between -20 to 45 degree depending upon ambient

    2) It must meet 0.1%  to 0.15% between 10 percent to 90 percent load. 

    3)Regarding calibration you could suggest .

  • Hi Biswajit,

    Happy to help! 

    You should set the DC voltage to maximum, 1000V. When calculating error, you divide by the full-scale range and yes your measurement can be capable of 0.1% error. However, is this a typical or worst-case error target? 

    You can use the same sheet for current sensing calculations if you reduce the R1/R2 values to be near 0. 

    The difference in typical vs total degree range will have a large impact as drift becomes an increasingly dominant error source. Clarifying what your actual temperature target is will help in your calculations. The device operation is specified at 25C, so when calculating the error from your temperature coefficients, use whatever temperature is farthest from 25C as your worst case drift value.   

    Gain & Offset calibration would be necessary given that the shunt resistor will have a 0.5% tolerance. It is also possible to calibrate over temperature or do local temperature measurements near the shunt when in operation to help combat this error. If you do the gain and offset calibration, you could consider using a shunt resistor with a higher tolerance, since the error will be calibrated out anyway.  

  • Hii Alexander,

    I am simulating AMC1311B in PSPICE for TI software. I will vary the temperature to observe the impact on the measurements. I guess AMC3302 will be a better alternative to AMC1301 for current measurement. I am seeing we can get <0.1 % error with it.

    I have two queries

    1) Could you suggest any IC for DC energy measurement (From TI)that can be interfaced with above sensors. Currently I am using MCP39F511A (From Microchip) that has a range of +/- 1V .To interface the AMC1311B to MCP39F511A, I have to put a differential amplifier to bring down the o/p below 1V. Will it impact the over all accuracy ??

  • Hi Biswajit,

    The AMC3302 is a great device and features the internal DC/DC converter which greatly simplifies design. You may want to consider looking at the AMC3330 as well as it is intended for voltage measurement and has an internal DC/DC converter as well. However, this device has a +/-1V input range, so you would only be able to use 0-1V for a DC measurement which may impact your accuracy requirement. 

    As for replacing MCP39F511A, I would recommend taking a look at the ADS131M04 family of devices. This device has a similar input range of +/-1.2V at an internal gain of 1, so a differential amplifier would still be necessary. This amplifier would of course introduce some error, mainly the devices inherent offset and the gain error introduced by the tolerance mismatch of the feedback (Rf) and gain resistor (Rg). 

    This TI Design shows an alternative way to perform a measurement such as this: https://www.ti.com.cn/lit/ug/tiduem7a/tiduem7a.pdf

    Please take a look and let me know if an architecture like this is of interest to you and I can connect you with someone more specialized in the ADS131M04. 

  • Hii,

    I want to remain with MCP39F511A for energy metering as of now. It has a maximum differential gain of +/- 600mV/GAIN at the input  of voltage & current sensing. GAIN can be increased from 1 to 32.

    For Voltage Sensing:- I am planning to use AMC1311B along with a differential amplifier. I have attached the schematic here. I have done the simulation & I am able to reduce the maximum differential voltage below 0.6V. This seems to be a workable solution.

    Will i be able to achieve less than 0.1 % error with this method across the 150 V-1000 V range.

    For Current Sensing:-  Do you have TI PSPICE model for AMC3302 ??? I didn't get it online. Though I have simulated in TiNA using the TiNA Model. Yet I want to confirm the simulation using TI PSPICE.

  • Hi Biswajit,

    Please refer to Table 8-1. Full-Scale Range in the ADS131M04 datasheet. This device is capable of +/-600mV input too and it's gain settings can vary from 1 to 128. 

    Glad to hear that you have a workable solution though! 

    I do not currently have the TI PSPICE model for AMC3302, this is still in the works.