Bird|otherbox has quit [Remote host closed the connection]
Bird|otherbox has joined #scopehal
Bird|otherbox has quit [Remote host closed the connection]
Bird|otherbox has joined #scopehal
Degi has quit [Ping timeout: 252 seconds]
Degi has joined #scopehal
bvernoux has joined #scopehal
<d1b2>
<rapzak_> Hi.. Got the SCPI on CDC up an running, if i like to make a continues streaming to scopehal, i think SCPI has a lot of overlay, so think about make a command there can put it into stream mode where it pushes out the data, until it recieve a stop command... will there be any format there is more easy to copy? - also allowing for several channels
<d1b2>
<azonenberg> The common way to do this on a single stream is to have a command that returns a binary data blob with a length prefix
<d1b2>
<azonenberg> But it's absolutely possible to do multi stream stuff too. the question is how tricky it will be on the software side
<d1b2>
<azonenberg> on Ethernet, we're moving a lot of stuff over to a dual-socket protocol
<d1b2>
<azonenberg> with one for control plane that is SCPI ASCII request/response format
<d1b2>
<azonenberg> and another for data plane that is just a firehose of raw binary sample data with headers on each trigger
<d1b2>
<azonenberg> i've never seen something do that over USB but it's absolutely possible to e.g. expose two CDC endpoints and have one for control and one for data or something
<d1b2>
<azonenberg> i'd suggest diong a single stream protocol to start for simplicity
<d1b2>
<azonenberg> then optimize later
<d1b2>
<rapzak_> my mind was, doing the scpi... one command changes it to raw/binary dataset streamed out - until another scpi is sent to stop it...
<d1b2>
<rapzak_> during the stream, no settings can be changed
<d1b2>
<rapzak_> the device rx line will stay idle/mostly idle, so it can come out of the raw mode
<d1b2>
<rapzak_> i like the option to do "no data missing" stream... because the samplerate is quite low
<d1b2>
<rapzak_> i hope there could be trigger options in scopehal, software wise
<d1b2>
<rapzak_> then it is more like a DAQ unit
bvernoux has quit [Quit: Leaving]
<d1b2>
<azonenberg> You might want to look at how the ThunderScope implements this
<d1b2>
<azonenberg> since it sounds like you have a similar use case
<d1b2>
<azonenberg> How low a sample rate are we talking?