PicoScope 7 Software
Available on Windows, Mac and Linux
Gap time between buffers
Gap time between buffers
Hi,
I'm logging some contiguous data using a picoscope 4824 at 5MSps.
This fills 32 0.5 second buffers which I then export and join together in matlab to give a 16 second array.
I am having a problem with the gaps between the buffers, which reading around the forums seems to be caused by rearming time?
Ideally I would like to have no gaps, but I can cope as long as I know how long the gap is.
Are there any ways to
1) record the whole 16 seconds or longer, at 5Msps without gaps
2) know exactly how long the gap was (I can then pad it out in matlab)
3) any ideas how I could measure the gap myself. I thought about using a calibration signal of high frequency, but I wont really know how many cycles are missed in the gap without spending quite a bit of time carefully doing this.
thanks for any guidance
I'm logging some contiguous data using a picoscope 4824 at 5MSps.
This fills 32 0.5 second buffers which I then export and join together in matlab to give a 16 second array.
I am having a problem with the gaps between the buffers, which reading around the forums seems to be caused by rearming time?
Ideally I would like to have no gaps, but I can cope as long as I know how long the gap is.
Are there any ways to
1) record the whole 16 seconds or longer, at 5Msps without gaps
2) know exactly how long the gap was (I can then pad it out in matlab)
3) any ideas how I could measure the gap myself. I thought about using a calibration signal of high frequency, but I wont really know how many cycles are missed in the gap without spending quite a bit of time carefully doing this.
thanks for any guidance
Re: Gap time between buffers
Use the SDK and streaming mode
Then you can do at least 10MS/sec for as long as you like.
Benno
Then you can do at least 10MS/sec for as long as you like.
Benno
Re: Gap time between buffers
Thanks for the tip Benno. I've had a search as to whats involved and I think I'll give it a go. I'm more of a software tweaker than a writer, so if you or anyone else has any good tips to get me on my way and over the learning curve that would be much appreciated.
My application is always going to be very sensitive to gaps in data, but I can cope as long as I know where they are and exactly how long they lasted. I think some sort of time stamping is going to be required, and the SDK might be the best route for this,
thanks again,
My application is always going to be very sensitive to gaps in data, but I can cope as long as I know where they are and exactly how long they lasted. I think some sort of time stamping is going to be required, and the SDK might be the best route for this,
thanks again,
Re: Gap time between buffers
Example code is available on https://github.com/picotech in different languages. You will need ps4000a ones for the PicoScope 4824
One question before you go down this route, how many channels of data collection do you need and at what desired sampling rate ?
One question before you go down this route, how many channels of data collection do you need and at what desired sampling rate ?
Martyn
Technical Support Manager
Technical Support Manager
Re: Gap time between buffers
cheers for the link to the code.
Well, ideally 8 channels at 5Msps (I have many TB of space if this could be done), but I suspect I'll have to compromise over the usb etc?
I guess 4 channels at 100Ksps or maybe down to 10Ksps is still very useful for me.
I could do some analogue work to make 2 channels at 10 to 100 Ksps useful still, but anything lower than this it might not be worth the effort and I'll have to live with small sample sizes from the picoscope.
I am guessing from the advice to go down the SDK route that is it not possible to have an accurate "global" time stamping of each buffers start? If I know very accurately the real start time of each buffer (ie stamped from a global clock), I can cope with the gaps. For example if I can very accurately know the exact time difference of a peak in buffer 1 with respect to a peak in buffer 32,this is useful for me. If the buffers have variable unknown pauses between then, even micro second ones, I lose the time delta accuracy across the buffers I need.