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.

TMP117EVM: TMP117EVM Data rate and limited input

Part Number: TMP117EVM
Other Parts Discussed in Thread: TMP117

Hi

I am kind of struggling with the TMP117EVM. The GUI starts with no problems, however after this there are some issues:

1. When I go to the Settings tab, if i want to select a contonuous data mode with 1 sec sampling rate, is it correct to select CC from the uppermost left scroll and then the corresponding averaging method that yields 1 sec CC? Because when I do this and then save the cvs I get a series of identical data points with slightly different timestamps, like

0,29.796875,13:45:34:865 
1,29.796875,13:45:34:963 
2,29.796875,13:45:35:245 
3,29.796875,13:45:35:386 
4,29.8203125,13:45:35:521 
5,29.8203125,13:45:35:654 
6,29.8203125,13:45:35:792 
7,29.8203125,13:45:35:929 
8,29.8203125,13:45:36:062 

it is like the sensor does not care and gives me just the average values for all the timestamps used for averaging. Is this counted in the max samples point that can be saved? Because if even if I say sampling rate 1Hz and then the sensor keeps sending at 100-130 ms than the max data points to be saved are too few for my application

2. In the plot tab it seems the logging does not start until I change the read delay from 100ms to 1s. No lines are plotted until that moment. When I change it I see some kind of graph rolling out but then maybe the graphs stops (I suppose it is because it has reached max number). It is not clear at all when the logging starts and if changing the read delay I influence the sampling rate at all. In the csv file I always see the same 100-130 ms interval between timestamps no matter what. And then what happens when the graph stops? Is it still logging but overwriting previous data?

3. If I want to do another measure is there some way to restart the logging without having to quit the EVM GUI and then restarting from terminal? Because now every time the plot gets stuck I can hit "Erase data" or "reset" until the end of time and still no answer apparently

Overall the system seems not responsive

  • Hi Beatrice,

    Welcome to the E2E Forums.

    What version of the TMP117EVM GUI are you using and are you running it in a web browser or have you downloaded a local copy? I want to see if I can recreate the issues you are seeing on my end. 

    Best Regards,
    Brandon Fisher

  • Hi Brandon,
    I downloaded from the website a standalone application for Mac OSX. The version should be gcruntime.v7. My machine in a Mac OSX with 10.15.5 Catalina

    Beatrice Pazzucconi
    JUNIOR DATA SCIENTIST 
    bp@empatica.com 


  • So I tried to run the GUI on the Windows virtual machine on my computer, since apparently the platform can cause issues. I am afraid to say that the GUI on Windows suffers the same issues as the one for Mac OSX. The Plot tab keep behaving weirdly and moreover it seems that it started logging while I was still looking at the Settings tab.

    When looking at the csv file I find:

    35,32.046875,16:42:26:968 
    36,32.046875,16:42:28:09 
    37,32.046875,16:42:29:051 
    38,32.046875,16:42:30:089 
    39,32.046875,16:42:31:129 
    40,32.046875,16:42:32:176 
    41,32.046875,NaN:aN:aN:NaN 
    42,32.046875,NaN:aN:aN:NaN 
    43,32.046875,NaN:aN:aN:NaN 
    44,32.046875,NaN:aN:aN:NaN 
    45,32.046875,NaN:aN:aN:NaN 
    46,32.046875,NaN:aN:aN:NaN 
    47,32.046875,NaN:aN:aN:NaN 
    48,32.046875,NaN:aN:aN:NaN 

    As if something went wrong during the measurement (probably it was me setting the settings). Upon repeating the measurement and closing a re-opening the GUI I noticed that this time the Plot tab is actually running as soon as I open it, but that the settings for the conversion have been maintained (1s, 64 avg) and that after the 100 samples to save the plot does not get stuck but keeps scrolling (allegedly discarding oldest data) The issue reported above seems to happen when I change the number of samples to save from 100 to something else. The plot seems to get stuck and then I find NaN:aN and other junk in the log

  • Hi Beatrice,

    Please verify you are using version 1.2.1 from the Gallery: https://dev.ti.com/gallery/info/THSApps/TMP117EVM_GUI. I would try to run it directly in your browser first (Safari, Firefox, and chrome should all work).

     

    Beatrice Pazzucconi said:
    1. When I go to the Settings tab, if i want to select a contonuous data mode with 1 sec sampling rate, is it correct to select CC from the uppermost left scroll and then the corresponding averaging method that yields 1 sec CC? Because when I do this and then save the cvs I get a series of identical data points with slightly different timestamps, like

    If you would like a 1Hz data rate for logging then i would recommend setting both the register auto-read delay (how often the GUI reads from the device), and the conversion cycle (how often the TMP117 takes a conversion) both to 1 second. The actual auto-read time is system dependent, so it wont be completely consistent, but it is typically a bit longer than the auto-read delay setting. Putting 1 second auto-read delay is essentially just adding 1000ms delay to the polling of the device. 

    You may still see some faster-time stamps if the auto-read delay was at 100ms for example, and was changed to 1second without clearing the chart. So if you changed it after 20 samples, samples 1-20 would be at 100+ms apart, and samples 20+ would be at 1+sec apart. If the chart was allowed to run until it overflowed (100 samples by default), then the 100ms samples should begin to be shifted out. 

    Beatrice Pazzucconi said:
    2. In the plot tab it seems the logging does not start until I change the read delay from 100ms to 1s. No lines are plotted until that moment. When I change it I see some kind of graph rolling out but then maybe the graphs stops (I suppose it is because it has reached max number). It is not clear at all when the logging starts and if changing the read delay I influence the sampling rate at all. In the csv file I always see the same 100-130 ms interval between timestamps no matter what. And then what happens when the graph stops? Is it still logging but overwriting previous data?

    By default, the graph should start sampling automatically once the EVM connects. The EVM connection status is this icon in the bottom left hand corner of the GUI:  .

    After starting, the graph should roll until it hits the "Samples to Save" size (default 100), and then it will shift out old data from that point. Old data is overwritten at that point.

    Beatrice Pazzucconi said:
    3. If I want to do another measure is there some way to restart the logging without having to quit the EVM GUI and then restarting from terminal? Because now every time the plot gets stuck I can hit "Erase data" or "reset" until the end of time and still no answer apparently

    The "Activate Reset Bit" will reset the TMP117 on the EVM to default configuration, but wont change the chart data. The "Clear Chart Data" button will erase all data for fresh logging. Please note that the EVM must be connected (as shown by the same connection status button above) for the clear chart data button to update the chart.

    Beatrice Pazzucconi said:

    So I tried to run the GUI on the Windows virtual machine on my computer, since apparently the platform can cause issues. I am afraid to say that the GUI on Windows suffers the same issues as the one for Mac OSX. The Plot tab keep behaving weirdly and moreover it seems that it started logging while I was still looking at the Settings tab.

    When looking at the csv file I find:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    35,32.046875,16:42:26:968
    36,32.046875,16:42:28:09
    37,32.046875,16:42:29:051
    38,32.046875,16:42:30:089
    39,32.046875,16:42:31:129
    40,32.046875,16:42:32:176
    41,32.046875,NaN:aN:aN:NaN
    42,32.046875,NaN:aN:aN:NaN
    43,32.046875,NaN:aN:aN:NaN
    44,32.046875,NaN:aN:aN:NaN
    45,32.046875,NaN:aN:aN:NaN
    46,32.046875,NaN:aN:aN:NaN
    47,32.046875,NaN:aN:aN:NaN
    48,32.046875,NaN:aN:aN:NaN

    This is unintended behavior, if you are able to recreate this in version 1.2.1 that i linked to please let me know. I have tried to recreate it in my setup by changing the samples to save number but i haven't been able to. 

    Best Regards,
    Brandon Fisher

  • I downloaded the version you suggested.

    Apparently did not solve my issue on mac (the plot chart keeps getting stuck after very few samples (10) and does not refresh even erasing chart data. Plus also in this case when I changed the number of samples to be saved the timestamp data are all messed up afterwards).

    I tried using the online TMP117 EVM GUI (FYI Safari is not supported)

    Here is what I noticed:

    1. The graph keeps sliding after the first 100 samples as expected if if doesn't get stuck

    2. The chart test sometimes gets stuck after the samples to save reach the max number. Clear chart data does not help. Plugging and unplugging the device does not help. The only way the chart unblocks is by quitting the GUI and re-opening it

    3. The device sometimes arbitrarily disconnects (which I did not observe before) that causes the chart to get stuck without any hope for fix. But it does not always happen that if the chart gets stuck is because of the device disconnecting. Same as point 2

    4. when I change the Samples to save from default 100 I actually see the counter going above 99 in the csv. But then usually the device gets disconnected

    5. Funny data in the logs appear when the device disconnects

    6. I managed to change the Register Auto Delay while recording and you can actually see the timestamps changing accordingly


    On my windows machine it seems that changing the GUI version actually did the trick: I can see the chart being responsive (when I cancel the data or change the Delay), and also when I change the samples to save number it appears not to get stuck

  • Hi Beatrice,

    One of my colleagues was able to verify the issues you are seeing on Mac. Thank you for reporting them.

    I will be updating the GUI in an attempt to correct this. I expect this to take at least a few business days so I thank you for your patience. 

    I will notify you via this thread as soon as the updated version is live. 

    Best Regards,
    Brandon Fisher

  • Hi Brandon,

    Thank you for your thorough support. For the moment being I can use the windows application my VM. I will let you know how the new GUI behaves when it is time

    Best
    Beatrice 

  • Hi Beatrice,

    I've released an update to the TMP117EVM GUI to version 1.3.0. It does not yet fix the issues you've reported, but this version corrects the timestamp formatting for excel compatibility and allows you the option to stream data directly to file, which you may find useful. 

    I'm still pursuing a fix to the Mac version issues with our development team on my end. I will close this thread for the time being and update as soon as I have news on that.

    Best Regards,
    Brandon Fisher