Rigol MSO5000 unrecorded memory garbage

[LiquidLogik]

Hi,

I recently purchased the LA active probe for my MSO5000 series Rigol.  I've been collecting some data with 8 of the digital inputs and i've noticed the end of the memory/unrecorded memory exported to CSV is full of garbage.  I tried pressing the clear button multiple times before and after collecting each data sample but it's still there.  I am aware the nonsense data is just semiconductors being semiconductors and is normal in any IC.  Ssome people intentionally clear the memory in their programs to ensure all variables have a known starting point.

I've programmed a couple of Python scripts to automatically process the data but the garbage has caused me some hassle.  It takes an inordinate amount of time to graph the data and the data output is about 3GB instead of about 100MB when the garbage is not present.  I've thought about attempting to detect the garbage but i know i'll either get false positives and/or allow some garbage regardless how i attempt it. The data i'm collecting is not a protocol or structured - it's random and i might have some garbage data signals, which i'm keen to record.


Is there something i can do to actually clear the memory so i don't get this garbage?
I could allow the oscilloscope to record empty/0V samples until all of the memory has been written to at least once, but if i set the memory depth to 200M and set the sample rate to the lowest then this is not feasible.


Below are a couple of examples.
X is the sample number
Y is the integer of the recorded hex value in the CSV file
The tests ended around the 5 million data point. Between the ~5.2-5.5 million data point until the end is the unrecorded garbage.

Thanks for any help in advance.

[thm_w]

Are you manually downloading to a USB key? Can you do this via SCPI instead, then you can ask for specific sections of data.

Depending on the bandwidth required you can also get a quality USB logic analyzer for <$200.

[LiquidLogik]

Are you manually downloading to a USB key? Can you do this via SCPI instead, then you can ask for specific sections of data.

Depending on the bandwidth required you can also get a quality USB logic analyzer for <$200.

I am exporting the data to a CSV file on a USB device.  I could try via SCPI but isn't it just shifting the problem?  It also requires a computer, which is not ideal since my oscilloscope doesn't have a permanent home moves around the workshop and occasionally out on site.

I already have a USB logic analyzer but the signals i'm capturing are 'industrial voltages', 24V switching signals.  Even though the absolute voltage rating for the device is 30V, the switching threshold is only adjustable up to 5V.  In the past i've stepped down the voltage down to 5V so i could set the threshold to some where in the middle of the signal but i frequently picked up noise since the environment is quite noisy, ground bounce, etc.  The selling point on the active logic analyzer cable for the Rigol is that the switching threshold is adjustable up to 15V, which works beautifully for my application.

[thm_w]

Ah yes if you are using it as a portable industrial recorder, then not really a way around exporting via USB stick.

Is it possible to save as .bin too? That might speed up things slightly.

[tv84]

2 suggestions:

- If you don't want to go around with a PC, choose something like a RPi and do the math there.

- To get rid of the shown garbage, I would start from the end of the captured file and I would cut it until the first non-garbage ocurrence. Process the remaining. I think it is the best solution possible.

BTW, think about this: if there was no "garbage" at the end, how could you tell that it was  a "garbage" or "non-garbage" zone?