Continuous sampling at higher frequencies

[TaiShengY]

Hi, we just got an OpenScope MZ and were wondering if there's a general rule of thumb that describes how long of a gap there will be between buffers if you're forcing triggering over WiFi. I've seen that streaming data through the logger functionality tops out at 50 kSa/sec, so I would assume this is the maximum frequency continuous data can be collected at hardware wise. We'd like to collect data as continuously as possible in the 200 kHz to 500 kHz range, is there a general relationship between this sampling duty cycle, buffer size and frequency or should we expect WiFi transmission/overhead to dominate the sampling?

Thanks!

[Fa-b]

Hi @TaiShengY,

I think the transmission is the bottleneck for continuous acquisition. When using USB the baudrate is at 1.25MBaud and then we're talking bits so including 1 start and stopbits there is 125kB/sec. 12 bit resolution probably sent in 16 bit raw integers makes 62.5kSa/sec in theory. Then you have to account for protocol overhead.

The OpenScope hardware though is capable of 6.25MSa/sec.

 

I don't know about the WiFi transmission rates achieved by the OpenScope, but I expect the protocol stack to consume a lot of the potential speed. I can even imagine that data are sent from the scope at the same rates as if it where connected through USB. Interested what Digilent will say.

 

Regards

Fabian

[JColvin]

Hi @TaiShengY,

Are you looking to measure a periodic waveform or a less standard/non-period waveform?

If you are looking to measure a periodic waveforms, the OpenScope MZ can trigger the oscilloscope channels and capture a buffers worth of data at 6.25 MS/s. However, the delay between one external trigger (whether event driven or manually done or otherwise) is very much dependent on other factors outside of the OpenScope MZ's control, such as the network speed and traffic, and how cleanly WaveFormsLive processes and displays the data. The drawback with these things is that any delays and processes are variable in their length of time, so not only will there be a gap in the data, you won't know how long the gap will be. But anytime you are collecting data, you will collect the data at full speed and fill up the buffer with it.

If you are instead looking at a less standard set of data, I would actually consider the OpenLogger which has been specifically designed for logging. Rather than external triggers and a lot of handshaking between the host board and the receiving end (mobile phone, pc, etc), each set of data is collected by a clock and placed into a buffer that is then simultaneously sent to WaveFormsLive or SD card.

Let me know if you have any questions about this and let me know a bit more of what you are looking to measure.

Thank you,
JColvin