Post Go back to editing

LTC2945 Sampling time

Hello,

I am looking forward using LTC2945 to measure a noisy power supply. The current to be measured will be noisy and will have high frequency step (1MHz+).

I need to be able to measure the average value of the current with precision.
What is the LTC2945 sampling time? Should i implement a differential mode R-R-C filter at the Sense+/- input?


Thanks.

Parents
  • Greetings Cyril,

       Thank you for your interest in the LTC2945.  The sampling rate of the ΔΣ ADC is approximately 30.7kHz.  Given the noisy environment in which you plan on using the part, a single-pole RCR filter with R= 10Ω and C=2.2uF would probably be a good idea.  Pin currents on SENSE+ and SENSE- are not well matched, 11Ω will put you right at the edge of the TUE.  

    Thank you and best regards,

    Brad

  • Hi, Thanks for your quick answer.
    Are the 30.72kHz sample being averaged inside the LTC2945? Because we can only read one value every 7.5Hz using I2C.

    Thanks.

  • Hi Cyril1383,

       The samples are indeed being averaged inside the part, and the registers are updated at a rate of 7.5Hz.  Between each of these updates the part samples and averages 4096 times.

    Best Regards,

    Brad

  • Hi blovell,

    we are also looking for this LTC2945 device and I was happy to find this information in this chat. This info is missing in the data sheet in such a clear way. Regarding what you mentioned here on the averaging I have another question regarding the max/min capture. Is it really so that only after a full conversion, and this is quite long, the sampled value is latched into the corresponding min/max register, after all the checks greater/less? In this case it can not be used for peak detection, right?

Reply
  • Hi blovell,

    we are also looking for this LTC2945 device and I was happy to find this information in this chat. This info is missing in the data sheet in such a clear way. Regarding what you mentioned here on the averaging I have another question regarding the max/min capture. Is it really so that only after a full conversion, and this is quite long, the sampled value is latched into the corresponding min/max register, after all the checks greater/less? In this case it can not be used for peak detection, right?

Children