Siglent SDS1000 7" oscilloscope anatomy (SDS1072CNL)

[rf-loop]

Tiny illustrated anatomy of Siglent SDS1000.

Edit: pictures are from SDS1072CNL  (70MHz without half speed "long memory")

... just because man is the most curious animal.

You can see big fan.  Fan is so silent that it is difficult to hear. This can do becouse they have selected enough size and also thick model so it can run very slowly.  First feel is "oh this is without fan".

[saturation]

Thanks a bunch rfloop! Studying it now.

[T4P]

Thanks a bunch rfloop! Studying it now.

*grabs notes* okay now i'm set for a new year

[free_electron]

looks like the emtpy spot for RL11 and RA54 were for a 50 ohm termination  ( one side of RA54 is connected to chassis)
a 595 shift register is apprently used to control some bits in the input system.

[A Hellene]

Thank you for the pictures, rf-loop.

It seems that the Siglent SDS1000 comes in four flavours:
SDS1000CFL: 4 channels, 70..300 MHz, 2 GSa/s, 24 Kpts memory depth
SDS1000CM: 2 channels, 60..150 MHz, 1 GSa/s, 2 Mpts memory depth
SDS1000D: 2 channels, 25..200 MHz, 500 MSa/s, 32 Kpts memory depth
SDS1000C: 2 channels, 25 MHz, 500 MSa/s, 4 Kpts/ch memory depth

Now, the Siglent SDS1000 design seems to be almost identical to the 2004 Rigol DS1000E design:
- The Sigilent mainboard has almost identical parts and arrangement to the DS1000E mainboard. Not to mention the strong similarities of the control/indicators interface and of their metallic and plastic cases.
- The power supply PCB seems to be the same exactly to Rigol's, and in the same exactly position and orientation with respect to the mainboard.
- The analog front-ends also seem to be exactly the same, with the exception of the differential amplifier (DS1000E: LMH6552) and maybe the input DC component amplifier also (DS1000E: AD8510).
- Both the Sigilent and the Rigol designs split the under test signal into its DC and AC components (see the Rigol DS1000E schematics).
- One DAC (DS1000E: AD5660ARJZ) with demultiplexers and sample&hold buffers based on the HC4051+TLC274 (DS1000E: the same).
- One dual high speed PECL comparator, probably the ADCMP562B that the DS1000E has.
- Several 74AHC595 to drive the amplifiers enable lines, the trigger selector chip (DS1000E: LMH6574), the trigger filters ('HC4053), and the 4.5V relays (DS1000E: UD2-4.5NU).
- Five dual high speed ADCs (DS1000E: Five AD9288).
- A Cyclone IV (or smaller) FPGA for the ADC timings and the ADC samples buffer (DS1000E: EP3C5F256C8N Cyclone III).
- An 18 Mbit pipelined SRAM for the SDS1000CM only, in the place of the unsoldered chip of the other three models (DS1000E: IS61LPS25636A 200MHz 9 Mbit SRAM).
- A MachXO An Altera MAX II CPLD memory manager (DS1000E: LCMXO256C MachXO).
- CPU: Blackfin DSP with Hynix system SDRAM (DS1000E: Blackfin BF531-SBSTZ400 and Hynix H57V1262GTR 128Mbit 133MHz SDRAM).
- A Spansion NOR boot FLASH (DS1000E: Spansion S29GL064N90 64Mbit 90ns).
- An ISP1362 USB On-The-Go controller (DS1000E: the same).
- An 100 MHz timebase, based on a 'AHC04 crystal oscillator.
- An analog interface TFT-LCD controller chip (DS1000E: UPS051 or HX8802).
- A 320x234 5.7' TFT for the 2-channel models, or a 480x234 7' TFT for the 4-channel model (DS1000E: TM056KDH02, 5.6' 320x234 analog interface TFT LCD module).

I am sorry to say that I am not surprised with the similarities I have found in these two 'different' products. Not surprised at all. Maybe Sigilent tried to improved that old design; maybe they just wanted a piece of the pie by using a wining recipe; I do not really know...


-George



EDIT:
For comparison purposes, here are pictures of Dave's Rigol DS1052E Teardown.

[free_electron]

no machxo in the siglent. its an altera max-II CPLD. the machxo is lattice. completely different beast...
sadly the siglent pictures are a bit blurry to see the partnumbers ( or i need better glasses )

[A Hellene]

You are right, it's an Altera on the Sigilent, not a Lattice. The image is blurry and seeing the 'MAX' part on it I mistakenly assumed it was also a MachXO. My mistake...

Thanks for the correction. I'll set right my post also.


-George

[T4P]

So... remember atten copied rigol? and that siglent is a cover name...

[tinhead]

So... remember atten copied rigol? and that siglent is a cover name...

no they don't. It seems that most ppl just spreading bullshit the whole time about "atten copying rigol".
Yes, there was was copyright discusion in China, and yes between Atten and Rigol - but it was an older model
where Rigol claimed to have (non existing) patents for TV signal trigger (actually the way how a specific transistor
was used to detect frames). Finaly the court decided that there is nothing wrong on atten devices.
Sure, when you look inside they look very similar, but honestly we talking about standard parts - and most
circuits are anyway based on app notes, so there is nothing wrong to copy them.

The software makes the difference, and there is complettly different software in DSP, CPLD and FPGA.

[rf-loop]

It seems that the Siglent SDS1000 comes in four flavours:
SDS1000CFL: 4 channels, 70..300 MHz, 2 GSa/s, 24 Kpts memory depth
SDS1000CM: 2 channels, 60..150 MHz, 1 GSa/s, 2 Mpts memory depth
SDS1000D: 2 channels, 25..200 MHz, 500 MSa/s, 32 Kpts memory depth
SDS1000C: 2 channels, 25 MHz, 500 MSa/s, 4 Kpts/ch memory depth

Older nearly obsolete models With 5.7" TFT have been long time. (SDS1000C,D,CM etc)
No one know who come first but in western markets Rigol come very popular fast. Maybe due to more marketing work in western countries and example with this "Purposely designed for hack". If look Rigol and Siglent they looks like same - nearly all bicycles looks like same. Maybe they also copy each others becouse all these have 2....  etc.  If human do not copy and learn from each others we still live in caves.

If look deeper to PCB, there are lot of similarities but also then lot lot of differencies in details. It mean that they are not directly copied with Xerox copy machine. 

But principle is same, and also this Rigol front end what is not at all Rigol design, it is copycat from Tektronix  but still also many differencies in details. (before years I have also tested and looked Rigols)

Rigol have not developed this front end principle, it is nearly as common public principle (all can free use it) and it can find in most this class of oscilloscopes and also it design principles can find in public textbooks - there is not ANY secrets and no one own this circuit principle. Some one select 499 ohm resistor and someone use 511 oh.. some use 33pF capacitor and some other use 27pF also components can add or subtract and positioned and routed different and it is not then copy.. then who copy who.  But sometimes it is fun becouse they copycat also errors, mistakes.

Also these kind of things best design help sources are component manufacturers application notes - who read these. And if really want study and learn there are old Hewlett-Packard journals. Very very pity this whole company have destroyed after Hewlett and Packard.  Agilent - everyone can do these cheap chinese boxes - even chinese but China is going too expensive country for cheap mass production.


There was some old case in China where Atten and Rigol go to court.  I do not know exactly what this case include and what was exactly result.  There are some words that Rigol win but what it include exactly is not clear for me.
(And some side note: Atten do not have oscilloscope product line in Atten factory!   Atten use OEM manufacturers afaik. Who all do  scopes for Atten and with what Atten "standards" this I do not know.   One OEM manufacturer is Siglent but who knows who others do sub manufacturing also for save money.  But all know that it is possible (with big lots) use buyer "standards" for exacmple save some money.)
Siglent is big ODM and ODM manufacturer. I do have any proofments who all use Siglent factory for they "own" products. You can buy all without any sign about Siglent name.. even PCB's may have your name.. it is only question about money. I have not any proofments that they do also for Rigol. It is so complex system of small sub manufacturers, OEM ODM manufacturers and cross manufacturing in China that maybe only a mountain troll know truth. 



Siglent product list afaik:
(There is also  old 5.7" TFT models but it is other case and it can simply regognize becouse they do not have last letter L)

All these "new" have 7" TFT
SDS1000DL    2 channel only,   25 - 200MHz  all in this serie 500MSa/s  32k sampling buffer

SDS1000CNL  2 channel only,  70 - 200MHz  this serie is 1GSa/s (exept 200MHz model have 2 x 1GSa/s (but not 2GSa/s max)) 40k sampling buffer  exept 200MHz model 5k sampling buffer.

SDS1000CML  2 channel only,  70 - 150MHz  only this serie (3 models) have 2M half speed long memory)  all have 1GSa/s  max 2M half speed sampling buffer

SDS1000CFL  2 and 4 channel,   70 - 300MHz  all in this serie  2GSa/s 24k sampling buffer.

CNL/CML factory claim is max 2000wfms/s. 
For slow speed there is in some models 6Msample continuous recording memory (and with extremely low speed there can record directly to USB stic).
(do not mix with frame recorder what is not continuous (lack of data between frames), this have max 2500 lenght )

Images quality are poor, sorry.  Purpose was only to tell the general picture of what kind of structure.
Also I do not know anything about original source of these images. But these pictures approached from the east and landed to my area without asking landing permission.

And this is not "new" unprecedented and progressive cross-breaking model, no one do not try claim it.

It is simple entry level and economy model without any big advantages. But still many also nice features in FW and building quality is not cheapest crap. 

But as all chinese, FW is not perfectly finished and there are many minor things what can do better. (but also big names have learn this from chinese and they have also started more ans more loose work.. look some agilent etc things. If these fails and errors happend 1980 or 1990 they need call back every equipment to factory - it is really BIG shame these old times good companies have started this copmpetition with prices and maybe also competite who can do most crap what still can sell only using old brand name.  How long they can keep this name..)


Signal quality is just what can expect from this construction. It can do better but then its price is littlebit different. Small amount of more lab work and some components more and some changes and some minor differencies in PCB routing etc..

It is just as one screw driver and you can see these hundreds of different cheap names. All these can turn screw when hobbyist do one table or fix his chair.  All peoples do not need Stahlwille tools.
All these same kind of oscilloscopes can show blinking led voltage changes "image" and other fun things.
Just if you like Rigol, buy rigol, if like UNI-T just buy it. There is not most bad and not most good. One have one thing good other have other good. There is not all good perfect equipment in this price class. All depend how You weight different things, pros and cons and also your feelings. 

I have seen real crap manufactured things. This is not at all this kind of product.
Also there can see they have not try save every jiao what is possible.


EDIT: there was "tinhead" explanation about some things what I did not know. tnx.

[poorchava]

AFAIR my warranty for Atten ADS1022 ends soon, so I might take it apart and do detailed photos when I have a free moment. It might be worthy to compare those two

[A Hellene]

Dear rf-loop, I am not defending nor am I pointing fingers at anyone. I only pointed out the strong design similarities of the two devices above.

I do not think I said that manufacturer A plagiarised a device that was designed by manufacturer B. Neither did I state that the original design was Rigol's; which is not: Free_electron informed us that it was Agilent the one who did most of the firmware development (and I supported that assertion with evidence) when they partnered with Rigol, for the latter one to manufacture for them an entry level DSO in order for Agilent to buy time to develop their own ASIC for their mid-range Infiniivision 2000X/3000X series DSO/MSO equipment.

As far as I can remember, Thomas had demonstrated in an older post the strong similarities the Rigol DS1000E design has with an ancient (meaning, of the nineties!) Tek DSO, whose model I cannot remember right now... And, as Thomas stated above, two devices having the same apparently components and PCB design does not necessarily mean that they also run the same firmware.

Who copied whom, or who 'improved' a third party design? Does it really matter for us, the consumers? It is obvious that it is for our benefit to have a wide range of choices.


-George

[rf-loop]

Dear rf-loop, I am not defending nor am I pointing fingers at anyone. I only pointed out the strong design similarities of the two devices above.

I do not think I said that manufacturer A plagiarised a device that was designed by manufacturer B. Neither did I state that the original design was Rigol's; which is not: Free_electron informed us that it was Agilent the one who did most of the firmware development (and I supported that assertion with evidence) when they partnered with Rigol, for the latter one to manufacture for them an entry level DSO in order for Agilent to buy time to develop their own ASIC for their mid-range Infiniivision 2000X/3000X series DSO/MSO equipment.

As far as I can remember, Thomas had demonstrated in an older post the strong similarities the Rigol DS1000E design has with an ancient (meaning, of the nineties!) Tek DSO, whose model I cannot remember right now... And, as Thomas stated above, two devices having the same apparently components and PCB design does not necessarily mean that they also run the same firmware.

Who copied whom, or who 'improved' a third party design? Does it really matter for us, the consumers? It is obvious that it is for our benefit to have a wide range of choices.


-George

Dear George

Yes. This was fully understood before. Only I want add also my some words to this "soup". 
Also these images show clearly that they are not copied but strong similarities in principle.
Sorry if you feel it was somehow negatively directed to you - there was not any kind of this meaning. 

Only quote was these models and this I want comment so that information is better updated for this time.

[A Hellene]

Sorry if you feel it was somehow negatively directed to you - there was not any kind of this meaning.

Oh, no my friend! I did not feel this way at all!

On the contrary, I was vaguely afraid that my comments could be taken offensively; my comments, that were not meant to be aggressive or even unpleasant, in the first place... Actually, I think that you have seen how much I enjoy analysing anything; from physical devices to intellectual concepts! But, though I do never intend to offend people, I often I manage to displease some individuals...


-George

[FenderBender]

Actually looks quite good! Thanks for the pictures.

[rf-loop]

Here is frequency response.

Edit: picture txt corrected (sweep start freq)

Cursors show -3dB level related to 1MHz level, this point is around 90MHz

[rf-loop]

Becouse sometimes is nice play..

There are different settings if look carefully (middle pictures have same signal, splitted with T and then cables littlebit different lenght (and without terminator so that nice ringing)

There is  some with sin(x)x and some linear, some picture have channel DSP filter and some have equ-time and also there is some persistence, then also becouse sometimes negative picture is also nice..  of course there are lot of many kind of settings, these just only imagine something just for fun.
(look also there is vertical mA in some picture)
In first risetime picture signal itself have <1ns risetime.

(note that jpg affect some amount!)

Last single image is same as number 7 in stack. (this is original .BMP)
(so problem is not Siglent image itself but normal jpg effect)

[rf-loop]

looks like the emtpy spot for RL11 and RA54 were for a 50 ohm termination  ( one side of RA54 is connected to chassis)

Some models have  also "50 ohm" input.
With this design it is not real 50ohm impedance over whole frequency area. It is only nearly same as feed thru termination installed outside to BNC input. (not exactly due to parasitic inductances and capacitances)
But in some cases better than nothing.  If look example some HP old oscilloscope input attenuator "hybrid" they are really RF designed and SWR is good over freq area. But this HP solution is more expensive. One HP input attenuator module price is perhaps nearly same as this whole oscilloscope.

[aghp]

Here is unofficial samplerate table for these models.



Normal "frame record" feature is not used in scan mode.

-aghp

[kingjay1]

Here is frequency response.

Edit: picture txt corrected (sweep start freq)

Cursors show -3dB level related to 1MHz level, this point is around 90MHz

I am a  beginner. Could you tell me why the amplitude suddenly up to 300mV at 150MHz.

[aghp]

Here is frequency response.

Edit: picture txt corrected (sweep start freq)

Cursors show -3dB level related to 1MHz level, this point is around 90MHz

I am a  beginner. Could you tell me why the amplitude suddenly up to 300mV at 150MHz.

There sweep generator return to 1MHz and start new sweep. Sweep time is 50s starting from 1MHz and ending to 150MHz and there start new sweep.

[electric_man]

Have somebody  tested  some functions of it ? How about?