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.

ADS1282 Modulator output

Other Parts Discussed in Thread: DAC1280, ADS1282, AMC1301, DAC1282

Hello,

We develop a application a high precision voltage measurement application. Goal is to measure voltage on an isolate side, to convert this value to a digital and transmit data over optical fiber. Then, on non isoled side, after reception we re-construct the analogue signal again. Target is to re-construct the a signal with the minimum phase and amplitude error. Signal Bandwidth is 10 to 10KHz.

On another project we use ADS1282 converter. An idea was to use directly the modulator output (3 wires) and connect it to the DAC1280. Dac1280 has only one bit stream input. Is direct connection (ADS1282 to DAC1280) possible or need we obligatory a µP between ?

Have you some application note about using the ADS1282 3 wires modulator output ?

What is the phase( or time setting) of a delta sigma converter ?

Thanks for your help

Michel

  • Hi Michael,

    Have you considered using an Isolated Amplifier, for example something like the AMC1301?
    That could be a much more cost effective and single-chip solution to your problem...

    Regarding the ADS1282 modulator outputs and DAC1282 bitsteam input... 

    The ADS1282 modulator outputs and DAC1280 bitsteam input are NOT compatible. Even if you were to combine these bitsteams into a single bitstream (using equation 5 in the ADS1282 datasheet), these signals still wouldn't be compatible, due to...

    • The ADS1282 modulator output rate being much faster than the DAC1282 sampling rate.
    • The 1's density of the ADS1282 modulator would exceed the maximum one's density of the DAC1282 bitstream input.
    • Without a digital filter, the ADS1282 modulator output spectrum gets fairly noisy above a few kHz (see figure 32).

    Therefore, an MCU would need to do a lot of extra work to interface these devices. Unfortunately, we don't have any applications notes about the ADS1282 modulator output bitsteams. There is an application note on filtering PWM signals with an analog filter to achieve approximately 10-bits of resolution, found at the link below, but my guess is that you're after a much more precise solution.  

    Perhaps a single bit delta-sigma modulator with a PWM output signal could work in place of the ADS1282. I would need to do a bit searching myself to help find you a recommended device. Off the top of my head, I'm not familiar with any ADC/DAC pairings that would do what you're asking.

    Best Regards,
    Chris

  • Hello Christopher,

    thanks for your answer. I will look for other ADC solution. AMC1301 is interresting, but for this application we need much more as 1500V isolation.

    For my enlightenment can you give more explication for the 2 parametres "Group delay" and "Settling time" of ads1282 ?

    Group delay is partialy due to Decimator FIR filter, does Sinc filter always add a delay?

    Best regards

    Michel

  • Hi Michel,

    In that case, you'll need to look for optical or inductive types of isolation, since the AMC1301 provides reinforced capacitive isolation, which is the highest voltage rated capacitive isolation solution today.

    Regarding your questions:

    Michel ROHR said:
    For my enlightenment can you give more explication for the 2 parametres "Group delay" and "Settling time" of ads1282 ?

    Certainly!...

    "Settling time" is a parameter common to all digital filters, it refers to how long it takes for the digital filter to provide a stable output result after a change in the input voltage. For example, a SINC filter is a type of moving average filter and the SINC filter order (SINC1, SINC2, SINC3, etc) typically tells you how many averages are taken. So a SINC5 filter would need to accumulate 5 conversion results in order for the moving average calculation to "settle" to its final value.

    Whenever you START or RESTART a conversion, the ADC's FIRST conversion result is usually delayed until the digital filter has settled. So a SINC5 filter configured for a data rate of 100 SPS, would provide the first conversion result after about 50 ms (5/100), which gives the first conversion result the appearance of having a much slower data rate (~1/5th the normal data rate). However, every conversion result thereafter would appear at the nominal data rate since the digital filter's accumulator is full after 5 conversions. This is especially important to pay attention to when MUXing the ADC inputs, as it significantly slows down the effective data rate in a MUXing application.

    Here is some additional information on SINC filters: 

     

    "Group Delay"  is very similar to "settling time"; however, group delay is the time it takes for a step input to propagate through the digital filter while it is continuously converting. The difference is that "group delay" is measured after the digital filter's accumulator is already full and it only compares the 50% final value on the input to the 50% value on the output. Therefore, "group delay" is always shorter than the settling time as the signal is NOT fully settled after the group delay.

    You can think of group delay as how long it would take a feedback control loop to recognize that the input has changed and for the output control signal to begin correcting for this change. Under certain conditions if the feedback response is too slow, the control loop could become unstable.

    Here is more information on wideband FIR filters here:  

    Michel ROHR said:
    Group delay is partialy due to Decimator FIR filter, does Sinc filter always add a delay?

    Yes, both filters will always have some additional delay time. In the case of SINC1 filters (or SINC1-like FIR filters), these filters will be settled after one conversion period; however, there may still be some small digital delay seen in the first conversion result. This delay is usally due to digital progation and some extra overhead required to calculate the calibrated output result.

    Best Regards,
    Chris