• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 46 subscribers
  • Views 6195 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.

Bipolar precision DAC recommendation

Gergo Santha
Genius 3585 points
Other Parts Discussed in Thread: DAC8581, DAC8806, DAC8801, TLC7524

Hi,

 

I'd like to make a simple ECG simulator and my initial idea is to use precision bipolar DACs. The output is needed to be in the -2mV...+2mV range and with low noise to articulate the P wave which is about 100 uVpp. If I'm not wrong the DAC8581 can do the job - or is there any more suitable  part for my purposes? Thank you.

  • 0
    TI__Mastermind 43915 points

    Hi JustGreg,

    A lot of ECG simulator designs use a low resolution parallel MDAC in combination with a fast op amp to generate the ECG signals. The reason a MDAC is used is so that you can achieve very fast settling time without having the linearity suffer. The parallel interface is used in order to achieve faster update rates in order to create the ECG transient response. A faster trans-impedance amplifier is needed in order to create a voltage from the current. If you have the option to use a parallel interface, the DACs that I would consider looking at are the TLC7524, or maybe the DAC8806 if your looking for higher resolution. If you are looking SPI, the DAC8801 may be an option. Using a small reference voltage with these parts is recommended in order to keep the LSB step size minimal while using lower resolution DACs.

    Regards,

    Tony Calabria

  • 0 in reply to Tony Calabria
    Genius 3585 points

    Hi Tony,

     

    thank you very much for your detailed answer, I really appreciate it. I went through the datasheets and I've got more question if you don't mind:

    1. We are talking about fast settling DACs but do we really need them here? I thought a data rate of maximum 2kHz is more than enough to output an ECG signal. (That means approx 1000 values for the waveform from a lookup table which is huge anyway). 500 us is easy for all the DACs which have been mentioned above since they are in the sub-us range. Or am I missing some important points here?

    2. My initial choice, the DAC8581 has almost the same fast settling time (~0.5 us) and DNL value compared to DAC8801. It also has the advantage that - for the first glance - it requires less additional circuitry. I've no experience whatsoever in using transimpedance amplifiers and circuitry so before jumping in I'd like to ask what are the advantages of DAC8801 over DAC8581 (or what are the main points which prevent the use of DAC8581)?

     

    Thank you.

  • 0 in reply to Gergo Santha
    TI__Mastermind 43915 points

    Hi JustGreg,

    It really depends on what kind of resolution in your ECG signal that you are looking for. I agree that 1000 points may be good enough. The concern that I have is that you are going to need a fast updating DAC that has a descent slew rate to be able to update and settle quickly around the ECG pulse itself. You do not want the settling/slew rate of the DAC to limit what you can do when it comes to reconstructing a ECG pulse. The settling time of the DAC8581 is fairly well at 0.65us but that is only to 0.1% or roughly 65 LSBs. You are going to need to wait roughly 1us for settled to data closer to a LSB. That may still be good enough though. For the DAC8581, the DNL is less than 1LSB indicating that their are no missing codes, however, the INL (relative accuracy) is 20 codes typical. The relative accuracy is what you are going to care about when it comes to reconstructing a signal otherwise, you may have as much as 20 code (typicaly) deviation in some places between when you theoretically calculate and what the output actually shows. That is where the R-2R DACs (like the DAC8801) stand out. Their INL is under 1LSB as well as the DNL. Depending on the accuracy of the system, you can make a decision on the DAC. It may for you that the DAC8581 is good enough for your signal reconstruction.

    Regards,

    Tony Calabria

     

  • 0 in reply to Tony Calabria
    Genius 3585 points

    Hi Tony,

     

    my apologies for the late reply. Thank you very much for your detailed answer. You were right in the specification, 20 code deviation is a complete QRS complex so I think I try the DAC8801.

     

    Regards,

     

    Greg