Some peoples asked from me that if I can somehow explain History function and Sequence function.
People who sit with me of course with oscilloscope + pen and paper it is easy to explain and also show.
After then I make freehand paper draw + text bit more clean. Perhaps it may help someone to imagine somehow better what these are.
Note that in SDS2kX there is lot of more wfm memory (140M) + also of course then much more waveform history buffer memory.
Amount of waveform history buffer memory is not defined in specification so this have perhaps also confused peoples.
When user turn Sequence mode on it also may feel strange because depending settings, what you see on display looks like very slow.
It need remember that even if it capture 80000 waveforms (segments) in one Sequence, it do not update display at all. There is only raw ADC sample points pushed to memory. After Sequence is ready, it start process every individual waveform from buffer for display, producing perhaps line or sinc curve between dots in every waveform (segment) and this all for every 80000 segment and also producing intensity gradient. Every 80000 waveform is stacked to one display (old error when it show only reduced amount of segments after Sequence ready is repaired time ago. Also in this error in older FW when search memory, they all are there).
After then it stop or continue next Sequence, depending user selections.
Sequence mode segment buffer and normal use waveform history buffer is same. Also view tool is same "History".
Attached this some kind of "explanation" image.
Edit/Add:It need explain bit more. There was question about history record data.
First it need understand that it is very different thing if we talk about display memory or acquisition memory. It can think that display memory is just pixel map for display. There is not anymore (exept in some special cases) pure raw ADC data.
When scope capture "waveform" say example 14kpts this go from ADC to sampling memory (this is lenght what often see in specifications "memory length" or how they name it).
Displayed waveform length is always 700 points horizontally. Siglent captured waveform length is always compressed to display waveformd width (700pixel). What is captured length is what you see, there is not overlapp. No part of waveform unvisible out from right or left side. Whole capture length, in this case (example) 14kpts is compressed for display 700pts. In this case 20 data points to one point (horizontally, and vertically as they are). It is bit different than some other scopes.
Now, when it capture one 14k it is compressed and produced "image" compatible for display memory and mapped it to display memory together with possible previous waveform there.. Then capture next and again compress and map it to same display memory together with previous capture it continue until display memory is updated for view on TFT. More fast wfm/s more these captured waveforms are over each other in this display memory. More they hit same pixels also intensity grow. (it is bit more complex how to produce intensity in details but this is not important now)
Some other oscilloscopes may record these compressed for display waveforms or sequentially whole display memory using some interval.
In these cases there may be also that display map have only small part of captured waveform and rest of capture length is out from display. Often in these scopes have some kind of image what imitate display area and whole capture length. This is also case in Siglent SDS1000CML and similar.Back to Siglent SDS1000X models and waveform history buffer. Same time it capture waveform to memory (in this case 14kpts) and produce "image" from this it also copy this true capture 14kpts to waveform history buffer. Every time new waveform is captured, raw 14k data is pushed to history buffer (and stamped with trigger time). Blind time in sequential waveform buffere is same as running scope blind time. All captured real raw data is there. But still there is of course not data what is not captured when it have dead time waiting nect possible trigger event. But it is still very very different in most cases than "History" or "waveform record" what many oscilloscopes in this price class and even bit over have.
After stop scope for looking history, you have access to full captured raw data instead of just display memory snap shots. In this example case up to 3912 stored 14kpts waveforms. (total bit over 54Mpts). (in this case scope settings is 1us/div 1Ch in use and 1GSa/s used. In this case wfm length is 14kpts.
After you stop and look these stored waveforms they look same as in window. 700pts wide. But, there is 15kpts data for every single wfm. Just turn window zoom on and look. Or in stop mode, of course also whole screen can zoom in and out just using t/div.
You can measure, use cursors and so on. You can turn it dots mode, lines, you can turn Sin(x)/x on and off (and see how it draw via true sampled points).
But still, even when this is feature is really nice and what Rigol or most of competitors do not have at all you need underrstand how digital oscilloscope work and it is not continuous history. It have blind time between every capture. Just as normally running scope have.
After every capture, it is waiting until it can capture next waveform. If it is R&S car price RTO or expensive Keysight top model. All these have dead time between waveforms. Only in very slow (more or less slow) some oscilloscopes can record continuous stream.
In this example I use not so good 1us/div, 14kpoints, 1GSa/s.
If there is signal what give trigger events enough fast, scope can capture up to 4220 waveform/s.
One waveform length is 14us. One capture period is 1000000us/4220=234us So, there is (average) 220us blind time and 14us captured data.
But, different time bases, different signal, different memory setting and so on all may change this blind time ratio.
Remember it also depends your signal. In worst case trigger system is just agen recovered and ready for next trig but in your signal it happend just before scope can trig again. Now it is waiting when signal have next change what meet trig settings. Or if in auto modde there come auto trig before your signal trig. This is also important to remember. If this must not happen do not use "auto trig".
If need less blind time then there is Sequence mode. With same settings (1Ch, 1us/div) it can capture max 65000 wfm/s so in this example case 1000000/65000 = 15.4us. 14us wfm lenth and 1.4us blind time cap between waveforms (segments). But seqment mode have other limit. Between whole sequences it need long pause for process all captured data at once for diplay memory for view)
For max amount of waveforms in history buffer you can look table about wfm/s speeds where is also sequence speed and all needed infomation. Using 10ns/div wfm history buffer can keep 80000 waveforms. (54M wfm buffer lenth is not available with all t/div settings. It is compromize with many variables)
Summarized in brief:
In history waveform buffer, there is real raw full used memory length waveforms. It is NOT display memory map history.
It is real captured waveforms "raw data" history buffer, Image processing is totally out from this process when it push every single captured full memory length waveform data to buffer.
Home
Help
Search
Login
Register
