Author Topic: Siglent - 11/20 - New SDS1104X-U, 4 channel 100MHz, 1Gsa/s economy oscilloscope  (Read 18717 times)

0 Members and 1 Guest are viewing this topic.

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 12726
  • Country: 00
Rigol FFT is totally wrong. I have seen notes that Rigol 1kZ have 64k FFT instead of 2k. So 64k and 128k not very big difference.

Yep.

All data points are cherry picked, there's inaccuracies and missing competitors (where's the cheap Keysights/Insteks?)

Also this kind of tables are not very useful because it do not expain any thing any deeper than just name, number and v or x.

Yep. They tell you nothing about the usability, the amount of useful information shown on screen, how that information is shown, how many button clicks and knob twists is takes to turn on a simple measurement, if they can zoom out after a data capture, etc., etc.

But still... par for the course in the marketing world.

« Last Edit: December 04, 2020, 08:22:58 am by Fungus »
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
Rigol FFT is totally wrong. I have seen notes that Rigol 1kZ have 64k FFT instead of 2k. So 64k and 128k not very big difference.

Yep.

All data points are cherry picked, there's inaccuracies and missing competitors (where's the cheap Keysights/Insteks?)

Also this kind of tables are not very useful because it do not expain any thing any deeper than just name, number and v or x.

Yep. They tell you nothing about the usability, the amount of useful information shown on screen, how that information is shown, how many button clicks and knob twists is takes to turn on a simple measurement, if they can zoom out after a data capture, etc., etc.

But still... par for the course in the marketing world.

Cherry picked or not. This is marketing and normal and every single manufacturer use it. Keysight perhaps one of leader in this area of things. If you have read enough HP-Agilent-Keysight sales brochures but not only these, even and more clever can find in many application notes. They have done it tens of years, they are very experienced in this. 

But, If do Chery picking, cherry need be true not lie cherry.. It is still not extremely fair but without direct lies.

Btw, Rigol manual still tell FFT length is 16k or somehow from diplay area 1200pts  :-//

And what is Rigol Frame Record... they tell it is fast, hardware based and max 60000 frames. If look afterwards one saved frame... is it possible look it with FFT or decode it or make some new measurements based to ADC data there. Or change interpolation.
Time when I have Rigol 1kZ I did not use it at all because it was for one project where it totally fails, I do not even remember if this feature was there in this time (just after  1kZ born).
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 10317
  • Country: my
  • reassessing directives...
Apparently one is on the way to me. Wioll be interesting to see how stripped down the design is compared to the existing model.
will be interested to see your review. the table in your next post shows comparison with Keysight DSO, but majority of us own DS1054Z so we hope to see comparison with that too. we would love to see some bashing and cursing and how crappy our DS1054Z vs this brand spanking budget scope. so maybe we can consider (based on your acedemical review) if its worth another buy.

Comparison (from Siglent)
edit: just checking it.... there are small lies... there are big lies.. (attached) if they based on that and just add 2Mpts to competite, they were highly deluded.. enough for me... there is no possible way to get 10MPts (bins) FFT out of this thing...
http://scopechart.com/rigol/rigol-ds1054z-ds1054z-1
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 22254
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Comparison (from Siglent)
edit: just checking it.... there are small lies... there are big lies.. (attached) if they based on that and just add 2Mpts to competite, they were highly deluded.. enough for me... there is no possible way to get 10MPts (bins) FFT out of this thing...
http://scopechart.com/rigol/rigol-ds1054z-ds1054z-1
Nit picking.
SW engineers have made some wonderful enhancements to DSO memory in the past from the initial release of a product yet the datasheets are nearly never updated to reflect the changes.
So which datasheet do we compare from ?

A case in point was the original first SDS2000 with 35 Mpts memory that later with FW was enlarged to 70 Mpts yet no datasheet ever reflected this.
What is right to do ?
Update datasheets and cause confusion in the marketplace or keep datasheets to match initial product release ?

Or later release a new tweaked model of much the same instrument like SDS2000X, now with 140 Mpts memory.

Best is a apples vs apples comparison.
Avid Rabid Hobbyist
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 10317
  • Country: my
  • reassessing directives...
A case in point was the original first SDS2000 with 35 Mpts memory that later with FW was enlarged to 70 Mpts yet no datasheet ever reflected this.
Or later release a new tweaked model of much the same instrument like SDS2000X, now with 140 Mpts memory.
i'm not sure if i understand. you mean siglent scope can be hacked to 35/70/140Mpts? can you point me to the hack thread? or are you saying, the scopes come out of factory at 35/70/140Mpts capacity? its just that not updated in datasheet? and what price? $500 budget? if spec/datasheet is not what it actually is, whats the point discussing here? btw, in earlier post i was referring to the stated (outdated) 12Mpts DS1054Z memory (actually now 24Mpts out of factory) when Siglent made the comparison chart and claimed their scope has 14Mpts (if actually bigger?) thats their problem, but when they stated (under rated) the competitor's spec, thats trouble! can we allow laziness in updating datasheet/chart?
« Last Edit: December 04, 2020, 07:19:25 pm by Mechatrommer »
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 3453
  • Country: us
So which datasheet do we compare from ?

The one that is current when you make the comparison.  I get your point, but everybody with any current knowledge of budget scopes knows that the 1054Z comes with 24M for a while now. My honesty expectations for marketing types are pretty low, on the order of politicians, but this does seem to fall below even that standard.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 12726
  • Country: 00
....and what price?

SDS2000-level, so $1500 entry point. Not the cheapies.

 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 21576
  • Country: nl
    • NCT Developments
So which datasheet do we compare from ?

The one that is current when you make the comparison.  I get your point, but everybody with any current knowledge of budget scopes knows that the 1054Z comes with 24M for a while now. My honesty expectations for marketing types are pretty low, on the order of politicians, but this does seem to fall below even that standard.
Not just that but technically the hackability should be taken into account as well...

In the end these comparison charts are utterly useless. And some stuff in the datasheet too. Calling a 7" screen large  :-DD
« Last Edit: December 04, 2020, 07:58:32 pm by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 12726
  • Country: 00
In the end these comparison charts are utterly useless.

I remember certain Siglent fanboys endlessly posting a Siglent vs. Rigol chart here.  :popcorn:
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
A case in point was the original first SDS2000 with 35 Mpts memory that later with FW was enlarged to 70 Mpts yet no datasheet ever reflected this.
Or later release a new tweaked model of much the same instrument like SDS2000X, now with 140 Mpts memory.
i'm not sure if i understand.

Originally old SDS2000 (yes without X) come with single or dual* 35M memory length. And there was no any hack but some peoples have made BW modification in hardware.

Later one Siglent FW update rise memory length to single or dual* 70M. (Perhaps due to more wise memory mapping.) It was already there in hardware.

Next model, SDS2000X have single or dual* 140M max acq. memory length just from this model start.

And now SDS2000XPlus  single or dual* 200M.

*) In 4 channel models where construction is  2x(1ADC+1MEM)
In 2 channel models construction is 1x(1ADC+1MEM)

SDS1104X-E (4 channel) construction is 2x(1ADC+1MEM)
SDS1104X-U (4 channel) construction is 1x(1ADC+1MEM)

--------------------------------

Other question, not releted to this quoted msg at all.

What is real truth about Rigol DS1000Z    FFT
Manual tell it is 1200pts and then some kind of memory mode 16kpts.

Now I have heard it is 64kpts.
I have not seen any example or test image where can see this exactly. I have not seen after some search where can see amount of bins or df.  Why it so difficult.
Also this compare tables do not tell anything about true V/div things... yes I know it is display scale but imho more important is, what is full resolution V/div and what are then just magnified. As example in some Keysight model 1mV/div is digitally magnified from 4mV/d.
If one have true 1mV/dif and other have this magnified  1mV/div  is there apple vs apple. My opinion is - not. Even if we do not talk sensitivity but we change it to siplay scale.  Next time some make 100uV/div display scale and advertise in data sheet and sales brochures they have  100uV/div...

Of when we put rigol frame redorder to same row where other are segmented memory acquisition or sequence. Apple vs apple... oh well  I have seen many kind of apples.
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 
The following users thanked this post: Performa01

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
Siglent have officially corrected this FFT 2kpts mistake what is in DS1000Z, afaik.
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
In the end these comparison charts are utterly useless.

I remember certain Siglent fanboys endlessly posting a Siglent vs. Rigol chart here.  :popcorn:

As far as I can see it was Dave who post these  tables - due to unknown reason. Perhaps media business need.

And then next I try tell error there.
Also there is many other errors or other way not directly comparable things compared.  One example V/div things.
« Last Edit: December 05, 2020, 05:55:39 am by rf-loop »
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 22254
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
As far as I can see it was Dave who post these  tables - due to unknown reason. Perhaps media business need.
And why KS 1k series, it's a 2 GSa/s DSO FFS and in another market sector and price.  ::)
Avid Rabid Hobbyist
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 10317
  • Country: my
  • reassessing directives...
then i guess he got a chart made by some clown who claimed to know the SDS1104X-U spec and doesnt know well the DS1054Z spec.
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline Martin72

  • Super Contributor
  • ***
  • Posts: 1838
  • Country: de
  • Testfield Technician
Here a direct comparison from welectron, official siglent distributor.


Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
And here other one


Probes 100MHz = Probes included with 100MHz model
Probes 200MHz = Probes included with 200MHz model


But other thing.
I have asked this also before without answer or least I have not hit this answer.
Rigol manual and data sheet tell FFT  1200pts  and FFT in "memory mode" (what ever this is)  16kpts.

But somewhere I have seen claim about  64kpts.

Where is documented or directly displayed it is 64kpts. Who can show it or where I can find this excatly,  it is changed and when, after what Firmware version.
I like facts not rumors.
----------------------------------------------------------------------------------

ERROR corrected: Table is corrected. Available memory was previously wrong 14M 1Ch/7M 2 channels/ 3.5M 3 or 4 channels.

After I have tested it this is clear.

This is correct:
1Ch ON, max acquisition length 14Mpts (1x14M) max 1GSa/s
2Ch's simultaneously ON, max acquisition length for both channel 7Mpts (2x7M)  max 500MSa/s
3-4Ch's  simultaneously ON, max acquisition length for every channel 7Mpts (4x7M) max 250MSa/s


Also tested it with saving .CSV to USB.  2ms/div,  all channels on, 7M max memory. 250MSa/s.
It produce huge 490 000 488  bytes file. All 4 channels 7Msamples + header where is setttings.
« Last Edit: December 19, 2020, 10:02:19 am by rf-loop »
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 3940
  • Country: hr
I also need to make a not here about discussions on FFT:

Many are confusing number of sample points and number of FFT bins.

You need 2048 points to have 1024 bins..
If you sample at 1 MS/s, divided by 2048 will give you 488 Hz frequency between bins, in 1024 bins from 0Hz to 500 kHz (1/2 sampling rate).

So if a scope claims 64 KPoints of FFT, it will as a result have 32 Kbins.
It will also have 15,26 Hz resolution per bin at same 1 MS/s sampling rate.
Or 1,526 kHz per bin at 100MS/s with 50 MHz bandwidth
or 15.26kHz  at 1GS/s and 500 MHz bandwidth.

That is how you determine how much bins/points Rigol has, by figuring out best RBW at specific sampling rate/bandwidth.
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 10317
  • Country: my
  • reassessing directives...
Rigol manual and data sheet tell FFT  1200pts  and FFT in "memory mode" (what ever this is)  16kpts.
But somewhere I have seen claim about  64kpts.
Where is documented or directly displayed it is 64kpts. Who can show it or where I can find this excatly,  it is changed and when, after what Firmware version.
I like facts not rumors.
ok here is confesion from 1kZ owner for not the first time... the built-in 1kZ's FFT is crap! whatever it is, 64K? 16K? 2K? let call it 200 points, bottom line its crap... happy now? ps: now dont ever confuse yourself with its memory size 12M or 24M? dont bother! dont ever ask Siglent to put it in their comparison chart ok? ;D
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
I also need to make a not here about discussions on FFT:

Many are confusing number of sample points and number of FFT bins.

You need 2048 points to have 1024 bins..
If you sample at 1 MS/s, divided by 2048 will give you 488 Hz frequency between bins, in 1024 bins from 0Hz to 500 kHz (1/2 sampling rate).

So if a scope claims 64 KPoints of FFT, it will as a result have 32 Kbins.
It will also have 15,26 Hz resolution per bin at same 1 MS/s sampling rate.
Or 1,526 kHz per bin at 100MS/s with 50 MHz bandwidth
or 15.26kHz  at 1GS/s and 500 MHz bandwidth.

That is how you determine how much bins/points Rigol has, by figuring out best RBW at specific sampling rate/bandwidth.

Yes, naturally.

And in this thing, example Siglent, and many others do not leave anything what need guess.

This is SDS1104X-U
128kpts. 131072points aka samples and 65536 bins.  In this image sampling frequency is 50MHz and this give FFT range 0 - 25MHz. (as can see it is zoomed in lot from full scale)

50MHz/2/65536 is  381.47Hz   and this information it tell always on the screen. If user have selected less length then it show it.

Then these different windows have different response. Here example "flat-top" used in example. (not very good highest freq resolution but better for level.) 

Picked up just from wikipedia.

« Last Edit: December 06, 2020, 12:51:10 pm by rf-loop »
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3612
  • Country: fi
  • Born in Finland with DLL21 in hand
Rigol manual and data sheet tell FFT  1200pts  and FFT in "memory mode" (what ever this is)  16kpts.
But somewhere I have seen claim about  64kpts.
Where is documented or directly displayed it is 64kpts. Who can show it or where I can find this excatly,  it is changed and when, after what Firmware version.
I like facts not rumors.
ok here is confesion from 1kZ owner for not the first time... the built-in 1kZ's FFT is crap! whatever it is, 64K? 16K? 2K? let call it 200 points, bottom line its crap... happy now? ps: now dont ever confuse yourself with its memory size 12M or 24M? dont bother! dont ever ask Siglent to put it in their comparison chart ok? ;D

All we know this,  but question was what it is in scope itself. I wonder what makes it so difficult.
If I have DS1kZ on my table I know it after 2 minute if not sooner and I have answer for this kind of simple question.
Also I have not get answer if Rigol "record" frames can afterwardd use for automatic measurements, for FFT, for decode etc. They advertise it like it is seqmented acquisition...
About Siglent we have lot of data and details what it is and what not... but about Rigol... nearly like "guess what is this - play".
Do peoples buy these just for decorating some hobby corner or just playing some fun.

So, without further data/knowledge, there is user manual and data sheet what tell it is 16kpts in  "memory mode" -  what ever it is. I keep this as fact, until some data what overrides this information.

If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Huawei HarmonyOS 2.0  | 6G | Arcfox α-S
 
The following users thanked this post: Performa01

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 10317
  • Country: my
  • reassessing directives...
what ever it is. I keep this as fact, until some data what overrides this information.
for what? why bother with rigol if you already have siglent? its like you buy a DSO just because of its FFT? man you already have your SA? you make "rf" to loop now you want to see FFT in DSO :palm:

About Siglent we have lot of data and details what it is and what not... but about Rigol... nearly like "guess what is this - play".
Do peoples buy these just for decorating some hobby corner or just playing some fun.
for building circuit, for download data to PC so we can do serious FFT, not toyish 1Mpts FFT also possible upload to MATLAB for anyone interested. we are returning customer from 1kE to 1kZ. downloading data to PC from Siglent is "guess what is this - play" since forever, but someone hit a kill switch by saying they dont need data to PC, fine! Rigol already have name in people's heart for a decade for budget $400-500 scope, it gathered people into this forum when Dave make the first review 1kE. there was no $400 Siglent so it was no brainer. now Siglent just want to start the race, i wish them good luck! Rigol already has in-house 10GSa/s 2GHz for $10K, Siglent has half that and half that for not half that price. we will wait until Dave make review and compare both, he has both scopes. cheers.
« Last Edit: December 06, 2020, 02:07:04 pm by Mechatrommer »
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline switchabl

  • Regular Contributor
  • *
  • Posts: 71
  • Country: de
About Siglent we have lot of data and details what it is and what not... but about Rigol... nearly like "guess what is this - play".
Do peoples buy these just for decorating some hobby corner or just playing some fun.

I can only speak for myself, but: I already know the built-in FFT is horrible in the Rigol, so if I know I need it, I use something else.  :-// People "playing some fun" probably don't use it either, because it is no fun at all.

But I was curious because of all the different numbers mentioned, so I turned it on FFT for the first time in years:
I think the width of the side-lobes for a rectangular window should be (sampling frequency / number of points), so that suggests roughly 4k points, probably zero-padded to 16k?
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 12726
  • Country: 00
probably zero-padded to 16k?

Why would they zero-pad it?

When it was initially launched it used "on-screen" data like all the other measurements, ie. 1200 points.

"Memory mode" does some extra work to fetch more sample data than what's on screen. I'm guessing they build a composite buffer by fetching many batches of 1200 points, then do the FFT on that.



 

Online jemangedeslolos

  • Frequent Contributor
  • **
  • Posts: 370
  • Country: fr
I don't understand the positioning of this scope ?
it is perhaps intended for large institutions which will order large quantities and therefore save a lot of money compared to the X-E.

But the price difference is too small compared to the loss of specification for a one shot purshase for hobby or small business.

300€ or $ net and they will have a hard time manufacture enough I think.
In my opinion they missed the opportunity to destroy the ds1054z and all the unfinished chinese Aliexpress scopes.
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 22254
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
This is SDS1104X-U
And in this thing, example Siglent, and many others do not leave anything what need guess.


128kpts. 131072points aka samples and 65536 bins.  In this image sampling frequency is 50MHz and this give FFT range 0 - 25MHz. (as can see it is zoomed in lot from full scale)
Ahh, so you have one.
Being stuck in China due to CV does have its advantages.  ;D
Avid Rabid Hobbyist
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf