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.

TSW14J56EVM: Selection of a right set of devices

Part Number: TSW14J56EVM

Hello all,

I would like to create a 16 channel DAQ (12 bit or 10 bit are both fine) with a high sampling frequency. I learned that if I use TSW14J56EVM in conjunction with  ADS52J90EVM, that will do the job (in fact I learned that I should use FMC boards). Actually, I may want a 32 channel DAQ in the future and I thought this combination will do that for me. 

Here are my questions:

1. Will this combination be able to acquire 10 MHZ signal for the duration of 100 us every 10 ms (I have a trigger)? This event will continue for about 3 minutes. 

2. Will I have to store the whole data on the internal memory of the card and once my acquisition finished, I can transfer it to a computer? 

3. What other devices do I need to buy so I could make my DAQ (power supply, signal generator, etc)? 

4. Is there any better pair that you could recommend?

Thank you very much

 

  • Hi,

    Here are some answers for the TSW14J56 - for the other ADS52J90 questions I will defer to the medical devices apps team.

    2. The acquisition only occurs for 1 trigger/capture at a time. The firmware is not designed at this time to capture multiple triggers into the memory and then pull them off for analysis. This feature may be included in a future release.

    Ken

  • Thank you Ken for the reply.

    When you say, 1 triget at a time, does that mean it is not capable to recive trigger every (for example) 1ms?

    Thank you very much
  • The system can capture 1 trigger at a time - it can re-arm automatically and wait for the next trigger, but you would need to off load the stored capture within 1 ms. This may not be possible with the USB/software overhead. If another re-arm/trigger/capture occurs it will over write the data stored in memory with a new capture.

    This firmware and software was not optimized to do this quickly. There are some limitations at this point with the existing firmware and software. There are plans to incorporate a burst mode capture which will handle multiple captures in memory before offloading to the PC, but that feature is not yet implemented in the current system.

    Ken.
  • I see. Is there any other similar board avaialbe in TI that doesnt have such problem?
  • No - there are no other ready solutions to capture data from a JESD204B ADC EVM in the manner as you describe. The best solution would be to custom build firmware and software tailored to your specific needs. Trying to fit an existing solution into your current needs will almost always suffer some limitation that was not considered in the initial definition.

    As I mentioned this burst mode capture feature is in works, but may be several weeks before it is ready for release.

    Ken.
  • Hi Ken

    I did a little bit of research and realised that there is a board that does the busrt mode.

    AFE58JD18/16 16-Channel Analog Front End Evaluation Module + TSW1400 EVM

    In my application, I have aquisition 16 channels with 10Hz, and at each aquisition, I aquire 100 us. If I set the sampling frequency of the ADC on 50 MSpS, then in the burst mode I will have 10 x (100 us /20ns) x 16 x (60 s) x 2 samples which will resulst 960 MB data in one minute.

    Does that make sence ?

    Thank you


    ps: here is the link for these two boards:
    ftp.ti.com/.../AFE58JD18_16 EVM rev A UG v1p8.pdf
  • Hi,

    I will have to defer to the AFE58JD18 device support team to respond to this.

    Ken
  • Hi Ken, is theer any chance that i can call you and talk to you for a few minutes please?
  • Hi,

    I have pinged the medical support team to see if they can get back with you regarding the AFE58JD18 device.

    Ken.
  • I really apprecite that!
  • Hi Mohammad,

    I am not familiar with the burst mode feature of HSDC Pro+TSW1400--Ken can probably speak more to that.

    The AFE5818 can handle 50MSPS just fine across all 16 channels. The question is whether there is any support for multiple captures before data collection.

    The way our EVMs are setup, the AFE5818 converts the analog data and transfers this data to the TSW1400 when "Capture" is pressed in HSDC Pro.

    i.e. there is no real limitation on the AFE5818 side but more on the transfer/collection/data-display side (TSW1400 memory size, HSDC Pro capture options).

    Sincerely,

    Olu

  • Hi Ken

    i have decided to go a head with these two devices:

    www.ti.com/.../ads52j90evm
    www.ti.com/.../TSW14J56EVM

    to perform the data aquisition. here is what i need to do:

    every 20 ms (with an external trigger) , i need to aqcuire 100 microns data simoltaneously from 16 channels. The aqcuisition is non-stop. It means that for example with 1000 trigger sigfnals, i want to recive 1000x 100 microns data with that.

    could you please let me know:

    1. can this combination do the job?
    2. can i control the capture bottom in the software to create my own trigger?
    3. can the aqcuisition be done simoltaneously?

    Thanks,
    Mohammad
  • Hello Ken,

    About your answer, what do you mean by "USB/software overhead"?

    Let me explain what we're planning to do to solve the problem Mohammadreza described. We're going to simulate the burst mode but on the PC software side, not the FPGA firmware. We're going to develop a PC software using the HSDC Pro's API. This is going to be the workflow of the software:

    1) Every 1/5 second, the software is going to send a command to the capture board to wait for an external trigger.

    2) When the trigger is received, store 5000 samples from each of the 32 (or 16) channels in the RAM (5K*2*16 = 160KB in total), and immediately transfer the data to PC before the next trigger comes, in 1/5 second.

    3) To simulate the Burst Mode, when the data is received, go to step one and send another command for sampling using an external trigger.

    Since there is going to be only 160KB of data in each 1/5 sampling period, I think the USB3.0 on TSW14J56 can handle the transmission of the data in less than 1/5 second. Unless there is a special "USB/software overhead" that we're not aware of.

    Now, assuming that these are going to be the steps that the FPGA Capture Board has to take in less than 1/5 second:

    1) Receive the command and transfer the information to the ADC board

    2) Wait for the trigger

    3) receive and store the data in RAM

    4) Transfer the data to the computer

    these are my questions:

    1) Is it possible to do the task I described? Are we missing something?

    2) Does the Software, DLLs, or FPGA board respond as fast as 1/5 second? 

    We really appreciate your input on this.

    Hossein

  • Hi Hossein,

    The system can do what you want, however we have not done any timing tests to see how long it takes for PC software to issue a control command through the API (using either matlab or labview), capture the data and then transfer the data to the PC.

    The USB3 transfer speed is very fast and the data transfer itself will be much faster than 20ms, however there may be some overhead from matlab/labview in processing the command to the FPGA which may consume the 20ms.  The only suggestion I can make is to try it and see if it works within your time frame.

    Can you tell me what software you intend to use to automate this process?  Labview or Matlab?  I can check with our software team if there is a quicker way to do this.

    Ken.

  • Ken,


    Thank you for your answer. I would really appreciate it if you could talk to your software team.


    About the software, I was planning on programming in C# (because I already have a platform for data acquisition written in C#). However, if we decide to go with something else, we will probably work with MATLAB. However, I'm pretty sure that C# is faster than MATLAB.


    About your suggestion on trying our method in practice, if it doesn't work, I assume that we have to implement the burst mode on the FPGA firmware. Am I right? If so, considering that we don't have an FPGA engineer in our team, can you give us an estimate of the engineering time it will take for a newbie? I assume that it's more of a tweak to the existing code than a major change, right?


    Thank you in advance,
    Hossein

  • I also have a critical questions, I tried looking up the answers in User's Guides and other documents but I had no luck:
    Does the TSW14j56 (or any other capture board that supports ADS52J90), support capturing 16+ simultaneous channels?
    What about the HSDC Pro software/API, do they support 16+ channels?

    Thank you again
  • Hossein,

    HSDC Pro is capable of capturing and processing 16 channels. I believe this is shown in the ADS52J90EVM users guide figure 17/18 using the TSW1400.
    www.ti.com/.../slau632a.pdf


    For ADS52J90 specifics I will let the device apps team comment.

    Ken

  • Hossein,

    I cannot estimate how long it would take to make the modifications as I don't know the extend of the modifications needed.

    Ken
  • Ken,
    Thank you,

    Can you confirm that the TSW14J56 also supports 16 channel capturing? You only mentioned TSW1400.
    Can we have access to the VHDL codes of the TSW14J56?

    Bests
    Hossein
  • Hossein,

    The TSW1400 and TSW14J56 can support as many channels as the device can support.

    The Quartus project for the TSW14J56 is on the TSW14J56EVM web page.

    Ken

  • Ken,

    Thank you very much for your help.

    My final question is, what components do we need to purchase to make the system work?

    We are going to work with the TSW14J56EVM and ADS52J90EVM. I assume that we will need a clock source (I don't know what frequency), a power source, and an anti-aliasing low pass (or band pass) filter.

    I realized that TSW4806 and TSW2200 are possible solutions for clocking and power. Can you confirm this? Is there a solution for the anti-aliasing filter or should design it ourselves?

    We will order as soon as we get your answer.

    Looking forward to hearing from you,

    Hossein

  • Hossein,

    I will leave the ADS52J90EVM evaluation questions to the apps team that supports that device. I am not the right person to suggest things for that device EVM. Typically you would start by looking at the users guide to see what the recommended evaluation setup should be. You should work with the device apps team if you want to do something else.

    The TSW4806 is a possible clock source and TSW2200 could be a good power source, however you need to be aware of the limitations of the VCO/PLL on the LMK04806 as to what frequencies you can generate on it, as well as the voltage and current limits on the TSW2200.

    I may have C based solution to speed up your capture needs with the TSW14J56. It is not officially released or supported, but if you want to take a stab at it I can provide you a C example that may help speed up your captures as compared to the HSDC Pro API. Please message me directly.

    Ken.
  • Hi Hossein,

    I would suggest you get just the ADS52J90 and TSW14J56 EVMs first and get familiar with the boards and software environment before you make definitive system decisions. For example, the ADS52J90 EVM comes with an onboard clock distribution IC and multiple LDOs.

    Sincerely,
    Olu