10,000 Waveform Capture Limit - 6402D

Post any questions you may have about our current range of oscilloscopes
Post Reply
Posts: 0
Joined: Thu Apr 27, 2017 8:41 am

10,000 Waveform Capture Limit - 6402D

Post by hutleyj » Thu Apr 27, 2017 9:17 am

Hello all,

I am bit of a newbie when it comes to PicoScopes. I am currently evaluating a PicoScope 6402D for a missing pulse counter application. The criteria for a 'missing' pulse is one with an area less than 70% of a normal pulse.

My application at a maximum of 400 pulses per second (PPS) and the pulse width is fixed at 3.2us (very low duty). Based on this, the PicoScope is set to 500ns/div, that is triggered every 2.5ms. The test is run over 2.5 minutes, but can be as long as 5 minutes in certain cases. It is imperative that I measure and integrate every pulse.

I have used both rapid triggering and equivalent time sampling (ETS) that has allowed me to capture 10,000 frames of pulse data. The limitation of 10,000 frames is imposed by the PicoScope software and is mentioned in the user documentation. For the function I am trying to achieve, is need to capture at least 60,000 frames of pulse data (and as much as 120,000). With the low sample settings I have used, I should not be filling the memory buffer of the device (512MS for the 6402D, but I a considering purchasing a 6404D with 2GS memory buffer).

My question is why is the limit set to 10,000 frames, and can anything be done to increase this limit? Or is it possible to sequentially capture additional sets of 10,000 frames to achieve the desired amount? If I can capture the required amount of data, the integration and post-processing should be fairly straight forward.

I hope this all makes sense. I have emailed this to tech support, but thought I'd post a query here to see if other users have opinions or experiences. I look forward to the replies.

Kind regards,
Joe Hutley - BEng (Hons) MIET

Posts: 441
Joined: Mon Aug 11, 2014 11:14 am

Re: 10,000 Waveform Capture Limit - 6402D

Post by Gerry » Fri Apr 28, 2017 12:48 pm

Hi Joe,

I did answer your Help desk ticket regarding this query, but I'll include it here for completeness:

I can see the problem that you have, as you just don't have enough buffer memory in any of our PicoScopes to be able to capture the periods in between the pulses as well as the pulses themselves, and you're sampling too fast for streaming. So your only option is to try and trigger on the pulse event, but there is no way around the 10,000 waveform limit in our PicoScope 6 software.

However, you can use our Software Devlopment Kit (https://www.picotech.com/library/oscill ... nt-kit-sdk) to create your own application that will integrate the pulse waveforms and calculate the energy to be able to count the missing pulses. When using the SDK, there is no limit to the number of waveforms that you can capture (well you are limited by the available memory for your application), and you also have the option of streaming the data to the PC, because you can stream much faster in the SDK (although triggering and capturing 120,000 waveforms would be the better way to go).

What I should have added is that, if you were to go to a deeper buffer scope, it might be possible to do this, but this will depend to some extent upon what pulses you expect to capture, and how willing you are to compromise on their signal integrity.

Let's say, for instance that you're willing to consider doing the analysis with pulses that have limited harmonics (i.e. you are OK with rounded pulses being your staring point). If you were to go for a 6404D with a 2GS buffer, you would actually have 1.5GS available in PicoScope 6 (because of the available sample rates for the fixed Timebases used). Now, instead of capturing events you could just capture nearly all of the data between pulses as well (to get around the restriction of 10,000 waveforms in PicoScope 6. If, for instance, you use a Total Capture time of 50mS then you will only need to capture 6,000 waveforms, to capture all the data you need. If you then request the number of samples to be 25,500, which, for the 6404D, will restrict the actual number of samples to 252,016, with a sample rate 5.04MS/s, you're maximum bandwidth will be 1/5th of the sample rate, i.e. 1MS/s. The 3.2uS pulses represent a 6.4uS period square wave with a fundamental frequency of 156.25kHz. So, using the mentioned sample rate you should be able to capture up to 1M/156.25k, i.e. the 6th Harmonic (which for a pulse waveform with only odd harmonics will be the 3rd harmonic) which will give you something resembling a pulse with rounded edges.
However, this is where the acceptable pulse shape comes in, as you would be able to use or adjust your 70% criteria for missing pulses if all pulses are the same shape and perhaps just change in terms of height, but if they change in terms of shape, e.g. if they become more rounded, or worse still change in different ways, then it will be hard to apply a criteria to a sampling process that already filters the data that you're capturing (affecting some pulses more than others).


Technical Specialist

Post Reply