Products > Test Equipment

Siglent SDS6000A DSO's 500MHz-2GHz

<< < (61/64) > >>

tomud:

--- Quote from: nctnico on August 28, 2022, 09:26:56 pm ---
--- Quote from: Martin72 on August 28, 2022, 09:17:22 pm ---
--- Quote from: randomOracle on August 28, 2022, 09:01:02 pm ---The SDS5000X has 5 GS/s ... the SDS6000A has 5 GS/s -- and of course, I was just referring to the BOM. Let it be $10. It is a negligible overhead to include for an otherwise nice scope.

--- End quote ---

It´s not only the material.
You have to change the pcb layout, you have to change the case (cutout) and what about all the scopes that were already build and sold...

--- End quote ---
Well, the input should have been in the design to being with. But it could be the synthesizer /PLL would have needed to be designed entirely different to facilitate a 10MHz input. It is not always trivial to generate all necessary frequencies from a single clock source while maintaining low jitter / phase noise.

--- End quote ---

Exactly.

As I mentioned, I am playing with ADC in terms of assessing the possibility of building an amateur 500MHz oscilloscope and the topic of clocks is not as simple as it seemed at the beginning (Ignoring the high costs and low profitability of this venture). But that's how it is. Until a man spends a large amount of dollars himself, he will not know how many problems await him :-DD

nctnico:
Yep. The best way to avoid failure is not to try.  8)

2N3055:

--- Quote from: randomOracle on August 28, 2022, 04:46:18 pm ---I think if users would have read my comment carefully, they would have understood: I do mostly measurements in sequence mode until memory is full. Then I go and fetch the whole content of the memory. And yes, then the difference is quite significant between Fast Ethernet and Gbit Ethernet or USB2 vs. USB3. The problem is: most people do not notice this difference because they do single-shot acquisitions, and then the overhead is quite high of "payload data" vs. "initiating transaction and protocol overhead to transfer this data to PC". Clearly, then you are not using your scope the right way for this purpose. On the SDS5000X this was not implemented properly: I could record in sequence mode but there was no command to fetch the whole memory as one chunk. I had to scroll through history (one command) and fetch trace by trace (another command), which was in my case even slower then repeated single-shot acquisitions. I notified Siglent about this issue and they fixed this within the limits of the architecture given by the SDS5000X.

Considering that they offer USB3 on the SDS6000A for a mouse/flashdrive etc. especially ridicules the fact that they do not offer USB3 for that interface where it would have mattered: scope to PC, where they only offer USB2.

Same for 10 MHz ref in/out: if you do not use and and do not know how to use it, then indeed you do not need it. It makes a huge difference in my long-running measurements. Besides, there is no OCXO in the scope. So whatever awesome timebase they put in there, it will suffer big time over temperature changes (e.g., day-night shifts if not in AC controlled room). Just because this is not important to you, it does not imply it could not be important to others. I'm still of the opinion that this would have been a nice feature.

These two are simply features found on "higher-spec'ed" scope and that would have been little effort to implement. It is not like I'm asking for twice the analog bandwidth.
Seems like people got offended by the term "professional". This may have been exaggerated, but just look at all scopes <$5k and then all scopes >$5k and you will notice that below that threshold, most scopes do not have the 10 MHz ref in/out ... and most hobbyists would not buy a scope >$5k; but the likelihood for a scope to have this is much higher if it costs more than $5k or $10k

--- End quote ---

I did ask you what are you doing... you might be an Oracle, we are not. I did ask what was important to you.
I also did ask you, did you do exactly this thing. And said that in this case your statement, from options mentioned Picoscope 6000E would be better for you.
All your statements how 25000 USD LeCroy "implements properly" something is a moot point (as I said) because it has different architecture, very powerful PC motherboard inside etc.

You didn't read that bottleneck in standard desktop scopes is not the interface itself but system throughput.
Not the raw interface speeds. Which obviously are faster if they are faster. Putting a 10 GBit Ethernet in a platform that can sustain 20 MB/s would be expensive marketing, nothing more.

You didn't read that Siglent has scope platform (in China only at this moment) that is rack mount, up to 512 channels combined system to achieve exactly this type of application you're doing. And other manufacturers too. There are also dedicated acquisition cards etc etc...

You are doing something that is NOT mainstream use of a embedded desktop scope and scopes of this type (not price class, but use type) are simply not optimised for this type of use.
The article you linked to calls for a specific synchronous sampling technique, and is calling specifically for special digitizer hardware that is not a general purpose scope. Like I said.

Most of the users here don't even like LeCroy scopes, because they are "not real scopes" to them. They feel like that for Picoscopes too.
Most people never downloaded a speck of data from scope to PC for analysis.
Most of them will save a screenshot from web server or to USB drive, or maybe data but to USB drive (fast USB for that is available). Many never connected scope to network at all, and some are not allowed to do so in their workplace. That is the market for this type of scope.

To you it all seems easy but it is not. Products are made on basis of platforms. You cannot make some Frankenstein contraption of random set of parts that are not working well together. You start with some reference design for basic computing platform that can be tailored for super fast internal data transfer and temporal synchronization between acquisition/trigger/primary data processing block and general computing/UI block. You do it on something you can have common codebase for. You also don't want to invent everything from the scratch (like OS) because it may give no gains but is expensive to make and maintain long term.

Also, this scope was collaboration with LeCroy. There is very "similar" scope from LeCroy, and it also has same specification in regards of all the things you mentioned.

It all started with your bit of rant about how it is makes an "unprofessional" equipment if a scope doesn't have specific features you need for your purpose.
Then I started to try to explain to you that is not an merit for calling something professional or not. It is a mere feature set and target market position (who do you sell to and for what kind of use). There are many minor subgroups between scopes (of same price range) targeted for different users. Long memory, high definition, many decoders and analysis packages, for digital, for power, automotive, for SI, etc etc.... There is also a scientific, nuclear, physics etc, and also automated testing applications.

By your logic, very expensive LabMaster 10 Zi-A is "not professional" because if used for power applications, it would perform badly for power converters measurements and you cannot use standard probing solutions... No, it would simply be wrong tool for the job.


That is all. No offense was meant.

Best,

nctnico:
Yeah, that will teach randomOracle for sure... to stay away from this forum.  :horse:

randomOracle:

--- Quote from: 2N3055 on August 29, 2022, 07:19:22 am ---You didn't read that bottleneck in standard desktop scopes is not the interface itself but system throughput.

--- End quote ---

let me explain:
- instead of heaving this loop(arming - trigger - fetch data)
you do this:
- arm - measure sequence - fetch sequence
so you save many-times over the "fetch data" command which is only issued once and then a big chunk of data is fetched that includes multiple measurements. this is independent of the scope bandwidth and already saves you a lot of communication which otherwise slows down the scope.

so you use a crappy scope and flood it with arm-trigger-fetch messages to fetch small chunks. no surprise that bandwidth goes down and I'm not arguing against that.
the better approach is to fetch large chunks of data as this increases the payload size relative to the protocol overhead (cf. link to LeCroy document). this improves system throughput a lot, even on crappy scopes because this type of load they can handle much better.

next thing: the SDS6000A _already has_ USB3 on the front for mouse/flash drive ... so clearly the system should be able to sustain USB3 speeds. so why not have USB3 to connect to the computer? your system bandwidth argument is illogical in the sense that for a scope that has USB3, it should be able to sustain the speed of it, too. If not, clearly, the designers made a mistake. Which is my whole point here: they simply "forgot" to include USB3 on the back and this is what I complain about here. You may disagree and that is totally fine since apparently it is not something you need. If others need that, you should be fair to accept that, too.


--- Quote from: 2N3055 on August 29, 2022, 07:19:22 am ---You didn't read that Siglent has scope platform (in China only at this moment) that is rack mount, up to 512 channels combined system to achieve exactly this type of application you're doing. And other manufacturers too. There are also dedicated acquisition cards etc etc...

--- End quote ---

Do I care about a device that is in China and cannot be bought by me? If this is internally based on the SDS6000A platform (which to a certain degree seems to be the case, based on the SDS6000L product number), then this further supports the fact that this is a pure marketing decision: they want to sell the SDS6000L with synchronization ... and not muddy the water by having the SDS6000A also offering this feature. I'm looking forward to tear-downs of both scopes and then we will know. Of course, this was their decision right from the start when they did the design. I'm fully aware how products like this are made and that this is a complex process. However, this is something they generally can do (SDS5000X, SDS6000L) and the extra dollars spent would not hurt much on a per unit basis (again: assuming they would have considered that in the design process).

Long story short: USB3 on the back and the synchronization was missed and/or a pure marketing decision. Not necessarily for technical reasons. This is truly sad. Siglent makes great hardware but with this one, they fell short of my expectations. But seems like people on this forum can't accept a good rant.


--- Quote from: 2N3055 on August 29, 2022, 07:19:22 am ---[...]

--- End quote ---

There was simply too much finger-pointing and bold font in the remainder of your post that I will ignore this part.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod