Automatic down sampling within Cleverscope

Automatic down sampling within Cleverscope, a forum discussion on Cleverscope. Join us for more discussions on Automatic down sampling within Cleverscope on our Questions forum.

Back to Forum Index : Back to Questions   RSS

27 Mar 2008

Is it possible to down sample the 100 MHz sampled signal to say 10 kHz within Cleverscope and the down sampled signal stored on board allowing say 200 seconds of data to be recorded? Thanks

28 Mar 2008
Posts: 73

Hello dwlee,
We have roll graphing at 10Hz available in the Signal Information window and we plan to extend that to 50kHz this year. That would allow you to record indefinitely.

At present, the only way to do it is to set your timebase to 20 Sec (the maximum for acquiring) and in Settings|Acquisition Settings set frames to 11 and frames per capture to 10. Do a single capture and you will acquire ten frames of data with a small gap between each frame. Sampling rate will be at about 10kHz.

The current frame window will show the last frame captured. If that is frame 11 then the first frame captured is # 2. Select each frame and then you can see the captured signal. You are able to then ""Get Frame"" and get all the data points into your PC. Save it as a text file and you could then concantenate the files to see the complete 200 seconds in an external programme such as excel.

I hope this meets your needs.

Roger Carter

1 Apr 2008

Hi Roger,

The roll graphing extend to 50kHz sounds just what I'm looking for. How is this going to be done i.e. average or straight decimation? Also, I assume that the frequency will be selectable as long as it is a multiply of the scope's sampling frequency? Finally, do you have a date when this feature will be available?

Thanks Dave

12 May 2008

Any update on this?

20 May 2008
Posts: 412

Hello Dave and Peter,
it comes next after protocol display. We will be showing protocol display at Austronics in Sydney (SPI, I2C and Uart to start) next week. Streaming will be done when it's done! But we are starting on it! We have already implemented considerable chunks of the streaming capture system - especially in the hardware. Currently we use a presettable digital CIC filter operating at 100 MSPS to down sample to data rates of 100 kHz and up. This ensures that the signals streamed are band limited, and alias-free. It also gives us increased SNR and yielding more bits, something which quite a few people would like.

The problem so far has been the user interface required to deal with what can be very large data sets (eg 2GSamples - its only 400 secs at 5 Msamples/sec, which we can achieve in transfer). Perhaps we will bite off a smaller chunk, and go for a lower interim spec in teh meantime! What do you think?

20 Feb 2009
Posts: 9

I just got my Cleverscope today and I was trying to get the roll graphing to work but it never seemed to do anything. Maybe I misunderstood what it does. Can you give me a better explanation.


22 Feb 2009
Posts: 412

Hello Eric,
Thanks for buying a Cleverscope. To use the Signal Information window roll graphing display the signal information window (View/Display Signal Info), and then click 'Show Logging'. If you want to log every trigger that comes in, set the update interval to 0. Choose a file to save to, and either Excel or Text. Choose the derived value you want to log (eg DC, RMS etc). Choose the channels to log. Choose the number of values to save. Click Start Logging. The logged values will be saved to the file or spreadsheet, and at the same time displayed in the Log graph. The graph will autosize to fit the signal. If you want to log a Maths result, set the Information Source to Maths. If you want to log spectral information, set the source to Spectrum. If you want to log the derived values across just a portion of the captured signal, use the Tracker Graph to isolate the chunk of signal you want, and set the source to Tracker.

I hope this helps!

26 Feb 2009

Hi there, has there been any-update on the downsampling issue? I too am interested in using Cleverscope at a lower sampling rate with the data logger. 10Hz just isn't good enough for my application at the moment, and 100MS/s doesn't allow me to capture enough data. Ideally I would like to record a minute at 1kHz.

27 Feb 2009
Posts: 412

We are getting closer. We have now got the acquisition unit streaming to the PC at full rate (whatever it can manage) and have the rolling graph working. We are now working on storage and retrieval. (As you can imagine file sizes get pretty big quite quickly and we want fast navigation and display). And by the way, the update rate has now been 20 Hz for quite some time, though I know this still does not cut your 1 kHz need.

27 Feb 2009

Thanks Bart. Once I downloaded the software update I was indeed able to get 20Hz response. I look forward to the increased functionality of this scope.

28 Feb 2009

I am just wondering if there was any reason in the data logging file to split the time into one column for hours and mins and a seperate column for seconds and fractions of a second. If you were to plot time versus voltage for longer than one minute the data can end up non-concurrent or even overlap unless the data is post-processed. If the data file was to include an initial time stamp and then count up in seconds after that no postprocessing of the data would be required in order to plot it. Just a suggestion.

2 Mar 2009
Posts: 412

I kind of agree. There is the tradeoff between showing the actual time of an event (useful if the event occurrences are sporadic) and sequential samples - useful for graphing. It is not so hard to make another column with continuous time in it.
The standard microsoft time display does not show fractions of seconds (someone show us how?), and that is why we split it up.

6 Apr 2009

This is a useful thread, very informative indeed. It is great that you were posting updates like this one. I will be expecting more interesting features soon.

6 Apr 2009
Posts: 412

Ok, we have changed the format of our text files (both logging and text save). Now we give the trigger time in an Excel date format (days since 1 Jan 1900). We give the time of subsequent events in seconds, with a resolution of 10ns. Using this approach we can handle all the expected durations (up to a year at least) following the trigger point. We have also improved the initial configuration part so that the text file opens directly into excel with tabs in teh right places to make it easy to parse. Finally we have made the latest application save and open text files much faster. As an example it now takes 26 seconds to open a 92 Mbyte text file (containing 2 Msamples) and autimatically fit the display to the sample set. Opening the same file in Word takes 128 secs.
Back to Forum Index : Back to Questions   RSS
You must be logged in to post a reply

You need to Register or Log In before posting on these forums.


Your shopping cart is empty.