we are considering to by a Pico Scope with USB 3.0 for our research group. After reading your article on the usb bottleneck most of my questions have been answered, however I am still not sure in one point, since the article seems to focus on the live visualisation of the data via USB 3.0:
In our case, we are measuring for approximately 10 ms the voltage at a photo diode. Within this 10 ms time span a chain of hundreds of optical pulses will arrive equidistantly at the diode. While each pulse is 30ns long, the spacing is about 20ns between two pulses. My goal is just to extract the average height of each pulse via Matlab and store it in a matrix.
Our current setup consists of a premium oscilloscope with 200 MHz band width, 2.5Gs/s, which acquires 5MS within the measurement time. To improve the signal quality, the scope is set to averaging mode. Unfortunately, due to the high number of samples the oscilloscope only manages to acquire about 10 to 20 waveforms within 3s of averaging. Afterwards, the acquired waveform is send via LAN to the computer and processed by Matlab.
Since the overall process of data acquistion and transfer is too slow, I thought about speeding it up with a USB 3.0 oscilloscope, since in future, it would be good to increase the total measurement length from 10 ms to a higher value. During the measurement no visualisation is needed.
Could you please comment on which way I may expect an improvement? I would expect that the data transfer via USB 3.0 should be faster compared to LAN since the oscilloscope is quite old and still runs Windows XP embedded, but please keep in mind, that the direct visualisation is of minor importance, but instead I need to copy the data quickly to the computer and evaluate the data. Additionally, without any visualisiation: Over how many waveforms can I approximately average when acquiring 5MS in a time of about 10ms or the next larger available time span?
My last question concerns the product line:
Is there a fundamental difference between the models 3406D and 6402D like a faster hardware in the 6402D? I know of the increased bandwidth of 250MHz, however a bandwidth of 200MHz should be sufficient as we use logarithmic amplifiers with a bandwidth of only 100 MHz. So my question is, whether the oscilloscope 6402D itself will react and process faster?
Thanks a alot for reading through my post. I hope I could clarify my questions and I am looking forward to any comment.