So, I was thinking long and hard ( and me being not very smart it is kinda slow process ).
If you set manual sample memory to 100 Mpoints.
You set timebase to 1 ms/div.
At 1GS/sec you will have 100 ms worth of data in memory, 10 ms on the screen.
And then you zoom in separate window to 20 ns/div.
That will give you effect of TRIPLE timebase. Full length sampled at one length (not visible on screen but in memory), a intermediate length in main window (that you freely move around inside sampled data with normal timebase controls) and a third time length in second window (zoom window, that you move around with zoom controls).
To accomplish that effect on scope without manual memory, you would need TWO independent zoom windows:
(Attachment Link)
I can see how that might be useful sometimes, on a scope without multiple zoom windows..
Exactly what I tried explain in my last post (answering to Tautech).
Many scopes can do this partially. In runtime they can show, without zoom function open, just full window "zoomed in" part from whole length and with selected time scale.
When zoom window function open they can show same part of signal using more fast time scale or some other part from whole memory length with same or more fast time scale than main window. Just like I previously told.
1. Whole length with its time scale as main window time scale so there is not 3 time scales, only two (but many scopes in normal mode do not show this whole signal length in display except small bar graph top part of scree.)
2. Zoomed in main window with user selected time scale (main t/div setting) what is just this normal screen. Only that we do not normally say it is "zoomed in" display. Many scopes work normally in this full window zoomed mode with more or less part of capture is outside of screen what can not see in runtime, just visually blind.
3. Zoom function in use, screen vertically splitted. Same or more zoomed in part of signal in "zoom window"
If all are born in digital scopes era without anything about analog scopes, do we still talk about timebase when we talk oscilloscopes UI settings. Perhaps not. We perhaps talk about time scale as vertical, example voltage or current etc scale.
In this kind of "normal" scope there can be two visible window what shoe part of whole capture length (if scope is in this working mode) 1st is based to memory length in use and samples interval but I think it is wrong to say it have different third time base aka horizontal time scale. It have this time scale what we set when we are looking it. Even if length is 10ms and our main window have 10 divisions and our setup is 1ns/div and our window to 10ms long captured signal is only 10 ns slice and 9999990 not visible in runtime. Oh well, runtime can zoom out, and set 1ms/div and see whole length (if scope do not force overlap)
Cheap solution is add fixed memory length function as usually is in many many scopes - and is still also in some Siglent more old oscilloscopes so it can not say Siglent do not know what it is. Even more, some older Siglent scopes have Dual Independent time scale horizontal alternate mode with independent dual trigger system what is very rare. And fixed memory length with normal main window zoom and split window zoom.
How to keep whole user selected memory length visible, if user want, on screen and least two separate different time scale and position sub windows on screen... this is challenge. If there need be full freedom for set time scales in both sub window and fully free time position for both, this need different hardware.
Why do like others, walk your own road, Siglent! Plese let us see dual independent time scale aand position "zoom" windows with full length diaplay with user selectable fixed length or autoselected memory length where user can set limits and or force sample interval.
Also please add true dual independent time scale axis mode (noy alternate) with independent triggers for both (principle like in Siglent old SDS1000CML what have dual timescale independent trigger alternate mode), so that it can work least with two channels simultaneously and what can imitate (poorly but somehow) real dual
independent beam scope. It was not mistake when in history Tektronix made 7844 dual beam model.
Scopes what have dual fully separate ADC and dual memory like 4 channel Siglents have good change to do it.