Zooming *IN* [actually: signal reconstruction]

[2N3055]

Max ZOOM ratio (acquired any way) is ratio of available memory. Scopes with same memory and different sample rate will drop at same sample rate after you just change several timebase positions.
Actually we can argue that scope with higher sample rate will see it's higher sample rate only on few fastest timebase positions.

So it's best to start out with as high a sample rate as possible...?

Another point in favor of the MSO5000.

Are you making a point not think at all before writing an response?
I know you can do better than that..

Max sample rate has no bearing to detail acquired when sampling large amount of data if sample rate had to drop because memory is limit... On high  memory length to detail ratios, memory will be the limit.

Lets presume you have 2 scopes, both with 1Mpoints of memory, one with 10GS/s and other with only 1GS/s.
At 10 GS/s you spend 10X more memory than at 1 GS/s  per second. Okkay?
So if you have 1 Mpoints of memory, at 10GS/s you use up all memory when sampling 100us worth of data. If you want to sample 1ms worth of data , both of those scopes will have to sample at 1GS/s.. and will be the same. You will have same detail on both. That is why we say that long sampling memory is important, to keep native sample rate as long as possible. And that is why Keysight Megazoom is good "analog emulation" scope but loses even to DS1000Z when you need to sample longer period and zoom in to detail...

[balnazzar]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

[bdunham7]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

How much do you want?  Current model entry level DSOs have an astounding amount of memory compared to older models, even those from A brands. 

[Bud]

That is why we say that long sampling memory is important, to keep native sample rate as long as possible. And that is why Keysight Megazoom is good "analog emulation" scope but loses even to DS1000Z when you need to sample longer period and zoom in to detail...

Not true for event-based sampling. Keysight with its segmented memory can sample and store only when triggered. With that, you can have information captured for much longer if events are spaced apart in time than a scope with larger conventional memory, which will get filled in a few milliseconds.

[Fungus]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

I assume it's because they'd have to up the processing power to match.

(...so all the EEVBLOG forum haunters don't spend every waking hour pointing out how slow they are)

I think some teardowns have seen much more RAM inside them than actually gets used.

Also: Note that they don't have 256Mb or 512Mb of RAM or some "computer" number. It's always a nice round number that the processor can handle.

[rf-loop]

That is why we say that long sampling memory is important, to keep native sample rate as long as possible. And that is why Keysight Megazoom is good "analog emulation" scope but loses even to DS1000Z when you need to sample longer period and zoom in to detail...

Not true for event-based sampling. Keysight with its segmented memory can sample and store only when triggered. With that, you can have information captured for much longer if events are spaced apart in time than a scope with larger conventional memory, which will get filled in a few milliseconds.

Bit different thing this segmented memory acq.
 
And it is also limited to available amount of memory and also limited to max 1000 segments least in some Megazoom IV infinii vision X series 1000X to 6000X.

In many scopes there is more memory - also for segmented acquisition of course. Example even in Siglent SDS1004X-E up to 109Mpts  and lot of more than max 1000 segment.

Many oscilloscopes today have some kind of segmented memory acquisition - even in Rigol. But, also more memory for it. So what is point...  most of modern normal scopes have it.

[2N3055]

That is why we say that long sampling memory is important, to keep native sample rate as long as possible. And that is why Keysight Megazoom is good "analog emulation" scope but loses even to DS1000Z when you need to sample longer period and zoom in to detail...

Not true for event-based sampling. Keysight with its segmented memory can sample and store only when triggered. With that, you can have information captured for much longer if events are spaced apart in time than a scope with larger conventional memory, which will get filled in a few milliseconds.

That is absolutelly not what we are talking about in this topic. True, but not a topic.

[balnazzar]

I assume it's because they'd have to up the processing power to match.

As I thought. So, the bottleneck, in the end, is always processing power...

[tautech]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

How much do you want?  Current model entry level DSOs have an astounding amount of memory compared to older current models, even those from A brands.

FTFY

[balnazzar]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

How much do you want?  Current model entry level DSOs have an astounding amount of memory compared to older current models, even those from A brands.

FTFY

That's true. If only lower-status brands would care to do a lot of other things better, they'd become the new A-brands.

[2N3055]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

How much do you want?  Current model entry level DSOs have an astounding amount of memory compared to older current models, even those from A brands.

FTFY

That's true. If only lower-status brands would care to do a lot of other things better, they'd become the new A-brands.

That is weird and false statement.

Siglent and Rigol are for years not doing MUCH better than ANY A brand in area of less than 2000 USD scopes.
They have same quality and reliability and loads of features more than big brands.
It is only that there are many people out there who expect unrealistic things, disrespecting how much effort goes into making things.

And when, for instance, Siglent makes new SDS2000X HD that is better in its class than anything from A brands, they complain it is expensive.. While when Keysight asks you same money for a scope that just looks funny next to it than it is money "well invested" into good brand...
 

[balnazzar]

That is weird and false statement.

Siglent and Rigol are for years not doing MUCH better than ANY A brand in area of less than 2000 USD scopes.
They have same quality and reliability and loads of features more than big brands.
It is only that there are many people out there who expect unrealistic things, disrespecting how much effort goes into making things.

And when, for instance, Siglent makes new SDS2000X HD that is better in its class than anything from A brands, they complain it is expensive.. While when Keysight asks you same money for a scope that just looks funny next to it than it is money "well invested" into good brand...

First of all, you shouldn't mark as false statements other people's opinions.

A few days ago a SDS5000 got stuck in my hands and never recovered. Now it shows the spash screen every time you tried to boot it. I was just probing a small circuit stimulated at 200mV.

Second, I sometimes I use scopes from A-brands (Agilent) at the lab, and they are quite different (not sure about the price range because they are no longer in production). The difference is remarkable.

Third, this thing that one tells how he thinks and you scream disrespecting here and there has to stop. Some people think that Siglent and Rigol are built down to a price, while KS, Tek, etc are built to last.
Other people think that they are equivalent, and the big brands rip you off by making you pay for their name. And even if it was exactly so, nothing wrong with that, because paying a premium for a brand that is old and has built a solid reputation across decades is perfectly normal in any kind of manufacturing domain. Is it OK to pay 1 million for a Ferrari when a Dacia can bring you from A to B all the same (well, kind of..)? It's up to each individual to decide. It's not that people who buy the Ferrari are idiots. Or if you want a more realistic analogy, put a BMW in place of the Ferrari.

Stop patronizing people for their opinions. You should just tell them that you think differently, and articulate your thoughts.

[Fungus]

Some people think that Siglent and Rigol are built down to a price, while KS, Tek, etc are built to last.

If only there was some sort of explanation:

[noisyee]

At this point i feel obliged to ask: why manufacturers don't release scopes with more mem depth. They seem to be on the verge of generous with other specs (see indeed rigol with sub-1k$ 350 MHz 8 GSa/s), but memory depth is almost always the weak point of accessible scopes..

Nowadays most scope equips with more than enough memory. Few Mpts to tens of Mpts is a good trade-off between response speed and available sampling rate at long time base.
I had used some top class scopes with hundreds of Mpts memory, but only enable it at specific case. Like debugging high-speed data bus, in which case I want to capture as many packets as possible while keeping enough sampling rate. The response time is horrible: just take a cup of tea while waiting the scope processing it's data. So you MUST relay on proper trigger set-up, not a good luck, to capture what you want which again is a time consuming work.
So why bother significant slow down while in most case you don't need GSPS at ms/div. Needless to say entry level scopes don't have the power to analysis full data at large memory depth. You don't want to capture a large frame and then zoom in to find what you want, trust me.

[balnazzar]

Needless to say entry level scopes don't have the power to analysis full data at large memory depth. You don't want to capture a large frame and then zoom in to find what you want, trust me.

Indeed! (About both your statements).