Collecting (with automation) continuous data from multiple channels at high sample rates

I have a Tektronix DPO 70404C Digital Phosphor Oscilloscope, which is being used for a research project. I have been asked to find a way to automate the process of collecting continuous data on multiple channels, at high sample rates (say 625 MS/s), for multiple seconds at least. Unfortunately, the record length of this model (12.5 million) is far too small for these observation values. Despite googling for weeks on end (months, really), I haven't discovered an implementation that addresses my needs.&#160;<br> <br> Is anyone able to provide me any pointers as to what I would need to accomplish this goal? If it's not possible on this model, why is that the case and what models are capable?&#160;<br> <br> Thank you for any help in advance.&#160;

I have been able to use the PyVISA library to do some automation within the bounds of the record length, but getting high quality data is a priority and is simply not possible with lower sample rates.&#160;

Can you describe what you're looking to find?<br> <br> Generally speaking you have sample rate and record length to work with and while various purchasable options can be used to boost that, max record lengths can be limited to 1GB or so of data (how thats divided up can depend on model and number of active channels).<br> <br> Usually it's useful to use a trigger pattern to sub divide the data to localize analysis then generate statistical analysis over multiple acquisitions.<br> <br> Fast acq can work as well but they generate bitmaps with hit counts in them.&#160; (Which usually implies a statistical analysis of that data form).<br> <br> Long story short, which needle in what kind of haystack do you have?

I have an electromagnetic probe observing an Arduino system. The Arduino is running some C++ code, and we want to observe the EM data when it is and isn't executing certain instructions. This is all connected to one channel. The C++ code is setting an output port to HIGH and LOW depending on when the desired instructions are running. I've connected another channel on the oscilloscope to this output port, so we can compare the EM data with when the channel is reporting high or not.&#160;<br> <br> I'd like to create a 'heat map' of the Arduino by moving the probe to different locations on the board and observing multiple seconds of high sample rate data. Each position would save csv files of the data from each channel. I have automated the processor traversal component already, however I cannot find a way to implement this data collection due to the record length constraint. The code does work if I choose to observe at lower sample rates, but having high quality observations is a must.&#160;