The Siglent SDG1000 and SDG800 thread

[ivan747]

I have noticed there is a lot of info and research on the SDG1000 series generators, but it is spread all over the forum, so it's hard to find. I'm starting this threads in hopes that people can find this information easily. If you make a finding in your SDG1000 series generator, give us a heads up. Sometimes information is buried on threads not directly related to the instrument, on those cases we will link the particular post that mentions the instrument. In cases where a forum thread or other resource has, for example, a review and a finding, there will be a reference to the thread or resource on each category.

Anyone can contribute, contributions are asked and welcome. You can also post small findings here, if you happen to find something big and important (like a hack), definitely tell us about it and give us a link so that can post that link for everyone to see your findings.

Let's get started:

SDG1000 series

Members of the series:
Siglent: SDG1005, SDG1010, SDG1020, SDG1025, SDG1050. The last 2 digits indicate the max. sine wave bandwidth in MHz.
The SDG1005 and  1010 appear to have slightly different hardware, this is known by the lack of a fan in them, except for the early versions.
See here: https://www.eevblog.com/forum/testgear/function-generator-comparison-help-please/msg272666/#msg272666
LeCroy: WaveStation 2012, WaveStation 2022, WaveStation 2052. These are 10, 25 and 50MHz respectively.

Specs
Official datasheet http://www.siglent.com/2014EnglishWebsite/Documents/Datasheet/SDG1000_DataSheet_DS02010-E07A.pdf

Specs known but not officially listed
External modulation input sampling rate: 500ksps
External modulation input filtering: none, there is no Nyquist frequency filter, aliasing can occur. Feeding in a 250kHz sine wave, from Ch.2 (which is in phase) results in 2 discrete modulation values. In this condition, AM modulation looks like ASK modulation on a scope, and FM into FSK modulation on a spectrum analyzer.
Frequency response bode plots for the SDG1025 (also read the posts that follow): https://www.eevblog.com/forum/testgear/the-sdg1000-and-sdg800-thread/msg612806/#msg612806

Firmware
Reportedly, Siglent's firmware versions always end in an odd number. LeCroy's firmware versions end in an even number. It is known that the LeCroy firmware can be uploaded to a Siglent unit (will void your warranty), and it gives you a more pleasing visual theme, swapped channels (Ch1 willl be the left BNC instead of the right and vice versa) and removes the built in counter. It will also enable you to work with LeCroy's LabVIEW drivers without editing header files, and possibly other PC connectivity things detailed in LeCroy's programming manual. More details on that on reply #1

 But please do not try until you read the related thread. Obviously, this voids your warranty.
Also, please note that as of April 2015, Siglent has released LabVIEW drivers. Details down below.

*Official Siglent firmware download center: http://siglenteu.com/gjjrj.aspx?id=15
*LeCroy's download center (WaveStation 2000 series, NOT 3000): http://teledynelecroy.com/support/softwaredownload/documents.aspx?sc=24
*Put LeCroy's firmware in a Siglent thread: https://www.eevblog.com/forum/testgear/siglent-sdg1000-(aka-lecroy-wavestation)-firmware-updates/

Manuals
*User manual: http://www.siglentamerica.com/USA_website_2014/Documents/UserManual/SDG1000_UserManual_UM02010-E07B.pdf
*Service manual: http://www.siglentamerica.com/USA_website_2014/Documents/service_manual/SDG1000_ServiceManual_SM02010-E01B.pdf
*Programming manual: https://www.eevblog.com/forum/testgear/the-sdg1000-and-sdg800-thread/?action=dlattach;attach=127867
*LeCroy's user manual (describes in MUCH more details the PC connectivity): http://cdn.teledynelecroy.com/files/manuals/wavestation_om-e.pdf
*LeCroy's programming manual: http://cdn.teledynelecroy.com/files/manuals/wsta_scpi_manual_reva.pdf

PC Software
Official EasyWave software:  http://siglenteu.com/gjjrj-xq.aspx?id=867&tid=14
*Lecroy's software is exactly the same.
*Siglent's LabVIEW drivers: https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/msg656874/#msg656874

Authorized distribuitors (or anyone who offers a warranty)
Siglent offers a 3 year factory warranty, sellers listed provide this warranty. If they offer a different warranty, that will be pointed out.

*Forum member tautech, New Zealand: http://www.tautech.co.nz/
*Forum member rf-loop,Finland: https://www.eevblog.com/forum/profile/?u=413 He offers this warranty as a complement.

Reviews
*Video review by Connor Wolf:
*Some Amazon reviews: http://www.amazon.co.uk/product-reviews/B00EJ9BKO8
*Some TEquipment reviews: http://www.tequipment.net/Siglent/SDG1025/Function-Generator/?readReview=75c58#tab-reviews
*My humble opinion: http://www.tequipment.net/Siglent/SDG1010/Fully-usable-AWG-at-a-very-low-price/c629c3f1890b4/

Teardowns
*SDG1010 photo teardown https://www.eevblog.com/forum/reviews/siglent-sdg1010-function-generator-teardown/
*Another teardown, with text description: https://www.eevblog.com/forum/testgear/siglent-sdg1025-arbitraryfunction-generator-under-some-tests/msg129024/#msg129024
*Video teardown:
*Teardown of an early SDG1020 (2012) with fan: https://www.eevblog.com/forum/testgear/siglent-sdg1020-waveform-generator-teardown-pics/

Issues (unsolved) Potentially outdated!!!
*(2-jan-2015) USB driver is partially broken in Windows 8 by the lack of driver signature. A workaround is the following: The SDG1000 series have 2 USB modes. One is a device-specific protocol called USBRAW by the user interface. This mode requires proprietary drivers to work. Unfortunately, these drivers are not signed despite the manual sugesting they are, so Windows 8 will refuse to install or work with them under normal circumstances. There are two workaroundsL temporaly disable driver signature enforcement in Windows 8, which is re-enabled after a restart or use the second USB mode. This second mode, USBTMC is similar to how HID devices, such as USB mice, joysticks and keyboards work. It requires no device-specific drivers. Just like HID, Windows comes with a general TMC driver, that it can use to communicate with the SDG1000. The included software works with USBTMC.
*The PC software is lacking waveform clipping, amplitude normalization, compression etc: https://www.eevblog.com/forum/testgear/looking-for-reasonable-dds-function-gererator-need-opinions-please/msg577154/#msg577154
*There is no readouts in dBm, you can input values in dBm, but these get converted to volts and displayed as volts, not as dBm. A workaround is to print a table with volts and dBm conversions.

Issues (solved)
*SQUARE WAVE GLITCH Solved only for newer units. (in brief, the work-aroundis to use the pulse function when working at low frequencies, either that or a hardware mod): https://www.eevblog.com/forum/testgear/siglent-sdg1020-arbitrary-fungen-first-impressions-and-a-possible-signal-glitch/
As far as I am aware, the SDG800 does not have this problem, thanks to the EasyPulse technology.
*Sometimes your firmware will come in Chinese, this is how you change it to english: press utility button. Press 5th blue button once. Press 2nd blue button three times.
*TFT contrast issue, solved in subsequent firmware upggrade (download the latest firmware): https://www.eevblog.com/forum/reviews/siglent-sdg1010-function-generator-teardown/msg140663/#msg140663
*TFT contrast issue, detailed: https://www.eevblog.com/forum/testgear/siglent-sdg1025-arbitraryfunction-generator-under-some-tests/msg126941/#msg126941

Findings
*Sine wave purity and modulation tests using spectrum analyzer: https://www.eevblog.com/forum/testgear/siglent-sdg1025-arbitraryfunction-generator-under-some-tests/msg126941/#msg126941
*Jitter tests (pulse mode): https://www.eevblog.com/forum/testgear/siglent-sdg1025-arbitraryfunction-generator-under-some-tests/msg127312/#msg127312
*Same as previous 2, by the same poster, but with more pictures: http://siglent.freeforums.org/tests-sdg1000-function-arbitrary-waveform-generator-t5.html
*You can probably rack mount the instrument if you wish, buy the LeCroy accesory: http://cdn.teledynelecroy.com/files/manuals/wavestation_rackmount_assembly_instructions.pdf
*You can install LeCroy's firmware on your Siglent: https://www.eevblog.com/forum/testgear/siglent-sdg1000-(aka-lecroy-wavestation)-firmware-updates/
*Overshoot measurements and some USBTMC discussion: https://www.eevblog.com/forum/reviews/new-siglent-sdg1020-vs-used-fluke-281-wavetek-39a/msg137823/#msg137823
*External modulation input sampling frequency (correction: the maximum internal modulation frequency is 20kHz) (starts halfway into the review): http://www.tequipment.net/Siglent/SDG1010/Fully-usable-AWG-at-a-very-low-price/c629c3f1890b4/
*There is a method for loading CSV files directly into the SDG1000 series, that aparently does compression: https://www.eevblog.com/forum/testgear/looking-for-reasonable-dds-function-gererator-need-opinions-please/msg577256/#msg577256
*There is a way to add a temperature controlled crystal oscillator (TCXO) for higher accuracy:
*Of course, you could also get a 10MHz reference, as accurate as you'd like, put it in a nice project box and connect it to the external 10MHZ input. Just remember there will be jitter on the output anyways, because it's a DDS.

Other information of interest
*Siglent has released a Power Amplifier with a bandwidth of up to 1MHz and a power output of 10W (when ouputting between 500Hz and 200kHz). It's tailored towards their signal generators.
    Product page: http://www.siglent.com/ENs/prodcut-fjxx.aspx?fjid=1071&id=145&tid=16&T=2
    Manual: http://www.siglent.com/2014EnglishWebsite/Documents/UserManual/SPA1000_UserManual_UM60010-E01A.pdf
*SDG1000 and SDG5000 series comparison: https://www.eevblog.com/forum/chat/siglent-equipment-old-or-new-design-!!/
*Opinions on Rigol vs Siglent generators: https://www.eevblog.com/forum/buysellwanted/best-cost-effective-option-for-arbitrary-function-generator/
*3 mods you can do: https://www.eevblog.com/forum/testgear/looking-for-reasonable-dds-function-gererator-need-opinions-please/msg579056/#msg579056
*Wiki-style page, including IC names, and teardown photos. The Wiki belongs to an open source project, which I recommend you check out: http://sigrok.org/wiki/Siglent_SDG1010





SDG800 series

Members of the series
(pending)

Specs
Official datasheet http://siglentamerica.com/USA_website_2014/Documents/DataSheet/SDG800_DataSheet_DS02008-E02A.pdf

Specs known but not officially listed

Firmware
*Official firmware download center: http://siglenteu.com/gjjrj.aspx?id=15
Verify that the model number is correct before downloading

Manuals
*User manual: http://www.siglentamerica.com/USA_website_2014/Documents/UserManual/SDG800_UserManual_UM02008-E02B.pdf
*Service manual: there is no service manual officially published online. If there's one on the CD, please upload it for us. Thank you.
*Programming manual: http://www.siglentamerica.com/USA_website_2014/Documents/Other/SDG800_RemoteControlManual_RC02008-E02A.pdf

PC Software
Official EasyWave software: http://siglenteu.com/gjjrj-xq.aspx?id=868&tid=14

Authorized distribuitors (or anyone who offers a warranty)
Siglent offers a 3 year factory warranty, sellers listed provide this warranty. If they offer a different warranty, that will be pointed out.

*Forum member tautech, New Zealand: http://www.tautech.co.nz/
*Forum member rf-loop,Finland: https://www.eevblog.com/forum/profile/?u=413 He offers this warranty as a complement.


Reviews
(pending)

Teardowns
*Teardown of the SDG805 with photos of the internals: https://www.eevblog.com/forum/testgear/teardown-of-siglent%27s-sdg805-arbitrary-waveform-generator/

Issues (unsolved)
(pending)

Issues (solved)
(pending)

Findings
*You can probably rack mount the instrument if you wish, buy the LeCroy accesory: http://cdn.teledynelecroy.com/files/manuals/wavestation_rackmount_assembly_instructions.pdf

[Macbeth]

Nice thread 

I am not aware if the LabVIEW thing is possible with Siglent's firmware

I can confirm it is possible to get the LabVIEW interface running with Siglent's firmware by modifying the LeCroy LabVIEW "code". IIRC you need to look for the *IDN? SCPI command in the "code" (or WTF you call that OO GUI crap?) and replace the strings from "Wavestation nnnn" to "SDGnnnn". Of course LabVIEW is a total dogs dinner, I only got to play with the trial version and it had me tearing my hair out.

Stick with something more natural to a programmer, like Visual Studio, Excel even, or MATLAB which I have not had the pleasure of using yet, but Shahriar (thesignalpath) seems to use all the time. Also make sure you download and install the FULL NI-VISA and not just the runtime version.

Oh, one other thing that got my piss boiling for at least a day  was not knowing that the SDG1000s defaulted to a USBRAW format, and if you want to use NI-VISA, you have to change to USBTMC in the Utility panel.

[ivan747]

I edited the post, thank you 

[ivan747]

Nice thread 

I am not aware if the LabVIEW thing is possible with Siglent's firmware

I can confirm it is possible to get the LabVIEW interface running with Siglent's firmware by modifying the LeCroy LabVIEW "code". IIRC you need to look for the *IDN? SCPI command in the "code" (or WTF you call that OO GUI crap?) and replace the strings from "Wavestation nnnn" to "SDGnnnn".

Specifically, what you have to do if you want a LabVIEW driver for the SDG1010, SDG1020, SDG1025 or SDG1050 (yes, this is for search engine ranking ) is: install LeCroy's "driver" as usual (that is, unzipping the folder and copying it to C:\Program Files\National Instruments\LabVIEW 2014\instr.lib\), open the labview project file called "Teledyne LeCroy WaveStation.lvproj", then in the project explorer, look for a file called "LCWS Initialize.vi", double click, then open the block editor by double clicking one of the UI controls just in the right spot, look for a pink block called "WST," and change the text to "SDG,", voila. Done 

[commongrounder]

Hi ivan747.  Thanks for starting this thread. I know it will very useful.  I just wanted to mention, in the interest of being thorough, that there is a model SDG1005, too. This is the model I have.  Not surprisingly, it has no fan, which is great for the audio work that I do.
Is there any other user that wishes there was a way to change the output level readout, say from volts to dbm, without having to reenter a value?

[ivan747]

What do you use dBm for, may I ask?
You can print a table with volts and dBm values, if you want, but you are right. It is something that can be added to the settings menu.

[aviphysics]

*There is a way to add a temperature controlled crystal oscillator (TCXO) for higher accuracy:

Does this solve the jitter problem with short pulses generated at low frequency or does it just improve precision?

[ivan747]

It doesn't solve the jitter problem at low frequencies. That problem CAN be solved with another hardware modification. It consists of replacing a couple of SMD resistors on the board with lower value ones. It is linked in the main thread.

[rf-loop]

*There is a way to add a temperature controlled crystal oscillator (TCXO) for higher accuracy:

Does this solve the jitter problem with short pulses generated at low frequency or does it just improve precision?

First need particularize what jitter we are talking and produced by Square or Pulse function. (they have totally different operating principle and jitter mechanism is different)

Also need first know what is "low frequency" in this case. Is it 1mHz (milli Herz) or is it 1Hz or 1kHz or what. Also 1MHz is low frequency for others and "high frequency" for some others.

If we talk Pulse mode, there is system operating principle generated 8ns jitter independent of frequency (+ other jitters) This is because it have 125MSa/s and in this working principle it need some "jumping" for adjust pulse period/frequency.

If we talk about Square wave, and it is very different case due to its operating principle.
There can do some modifications what can improve cycle to cycle jitter some amount or more depending what manufacture version (and lot) it is.
In this model Square wave is derived from sinewave. (it works so that internally it produce sinewave in this Square wave mode - this is common classic way to do Square / rectangle.)

This is done: Sine --> in Comparator (with adjustable treshold and hysteresis) out --> Square/rectangle.
In this operation, specially with very low frequency  side product is timing jitter. This is because there is not ideal pure zero noise sinewave, there is not ideal Comparator without noise and not treshold voltage without noise and so on.)

There is two separate way to improve sguare/rectangle  wave timing jitter related to this comparator. (Comparator there is fast PECL comparator). Today (and starting long time ago) Siglent have changed it. It is possible that still can do some minor improvement. If people have older unit, it is possible it can improve lot of.   

In old case there was too narrow hysteresis window. It leads to corner glitches and high amount of time jitter with low frequendcies and even pulse bursts in worst case near rise and fall.
This can fix adjusting hysteresis window more wide. This can adjust with hysteresis set resistor value. But also this improvement need compromise. (if want also 20% and 80% duty timing jitter minimize then 50% time jitter can not minimize. If want get best possible 50% duty jitter, then 20/80% duty goes more bad. So it need compromize. For hyst set around 20k resistance values are ok but if put more weight for 50% duty time jitter then there can use example 15k or even more low but then perhaps too much  reduced quality for 20 and 80% duty.

Other source for time jitter is internal sinewave in PECL inputs. Jitter can some amount improve if filter higher noise frequencies from signal. In simply way it there is 50ohm (49.99)  terminator resistors for internal sinewave lines to PECL inputs.  Parallel with these resistors can solder small capacitors (these can solder over resistor). (I do not remember values exatly but we talk here about picofarads.  Perhaps something around 30pF (depending model).  It can select using tests specially if want find best compromise).

Comparator is ADCMP562.


In this image (from old unit) can see that Hysteresis setting resistors are 40.2k (R82, R87)  and this individual unit have high amount Square corner glitches and low freguencies cycle to cycle jitter (specially under 10kHz frequencies)

Internal sinewave transmission line terminators (50 ohm)  in signal path to comparator are R193, R194)

[jimon]

Are SDG10xx hackable for max frequency nowadays ? I'll prefer to have Siglent to Hantek HDG2002B.

[tautech]

Are SDG10xx hackable for max frequency nowadays ? I'll prefer to have Siglent to Hantek HDG2002B.

Maybe, but to date nobody that I'm aware of has made the effort.
You'd want to be aware the low frequency(>10 MHz) don't have cooling fans while over 10 MHz they do.

[rf-loop]

Are SDG10xx hackable for max frequency nowadays ? I'll prefer to have Siglent to Hantek HDG2002B.

No. It looks like Siglent do not use this marketing trick. (least yet)

(For sinewave, of course with reduced specs example SDG1010 can produce 50MHz using Arb and equation: 10*(sin(10*x)).
Now with 5MHz set for Arb it give out 50MHz.)

[commongrounder]

What do you use dBm for, may I ask?
You can print a table with volts and dBm values, if you want, but you are right. It is something that can be added to the settings menu.

I'm just used to having precision attenuators on my FGs. I'll sometimes start with a particular output voltage and then want to drop it by a certain db amount.  It would be convenient to be able to switch level reading types on the fly. When in the middle of a repair, or test, it is nice to not have to be distracted by a calculation or chart. Audio is my main business, and most gain stages are measured in referenced db values.  I have an external precision attenuator, but it takes precious time to set up.  If the SDG unit would just let one open the level set menu without entering a value, the measurement type of interest could just be selected, and the unit would convert the existing level to it.

[ivan747]

Are SDG10xx hackable for max frequency nowadays ? I'll prefer to have Siglent to Hantek HDG2002B.

No. It looks like Siglent do not use this marketing trick. (least yet)

(For sinewave, of course with reduced specs example SDG1010 can produce 50MHz using Arb and equation: 10*(sin(10*x)).
Now with 5MHz set for Arb it give out 50MHz.)

Personally, I think they do use the trick, and the only difference is the fan. We would have to start some formal research on this.

[rf-loop]

Are SDG10xx hackable for max frequency nowadays ? I'll prefer to have Siglent to Hantek HDG2002B.

No. It looks like Siglent do not use this marketing trick. (least yet)

(For sinewave, of course with reduced specs example SDG1010 can produce 50MHz using Arb and equation: 10*(sin(10*x)).
Now with 5MHz set for Arb it give out 50MHz.)

Personally, I think they do use the trick, and the only difference is the fan. We would have to start some formal research on this.

Really?

I can tell least this. Every single unit is different. Every SDG1010 is different. Every SDG1025 is different. Every SDG1005 is different. Every SDG1020 is different. Every SDG1050 is different.

Main stream models are SDG1010, 1025 and 1050.  There 1025 and 1050 have fan.
Fan is needed.  If we run unit 24/7 using these specifications limit values where stress is most high it need survive it without risk that failure rate grow too high. It is statistics. It need also include knowledge about components tolerances and all combinations what may occur, also worst cases. So if one individual survive it do not tell yet nearly anything. It need statistics using lot of equipments.

After manufacture there can see some units are better and some not if looks some test values what may include elliptic filters quality and end amplifier situations with different loads(!) with different waveforms including all kind of impedance mismatches and so on...  and just example some units have better square, less jitter and more jitter, and including that difference may occur only if go to other than 50% duty values.

Some units whole circuit (mid way amplifiers, filters, attenuators and end stage freq response is different, not only designed different but also due to all componets tolerances (and it need also know component values drift after ageing).   

If product can not meet some (company internal) specifications in product test for example 1025 model what to do.  I think no one believe they go to recycling.  So what to do?  First idea what every people can himself think is try if it still meet some other model specs.  If it meet SDSG1020 factory internals test specifications so that it can be this model, then it is just ok.
I do NOT KNOW if this is true in Siglent case but if I do this kind of products I do it.

But then, if you take SDG1005 and SDG1050  do you really think only designed(1) HW difference is fan. I do not "believe".

(1) there is least two kind of differencies in HW.
- designed differencies
- not designed differencies from every single component tolerances and all combinations with these, including also differencies in plain main boards what also are individuals still after designed to be equal.

 Example around end amplifiers there is really small things what affect lot of.  (and example in attenuator/elliptic filer/filter section)
Just for tiny example some pF change and all is different and many other things.

I know it, I have changed SDG1000's end amplifiers just because different manufacturer "compatible" is not at all compatible (accidentally - or trying money saving - or just due to availability in needed time - Siglent used (in some manuf. lot) different manufacturer compatible end amplifier.).  In this time also I have investigated end stages due to some ringing in cases where output is loaded with some level range using just some other than pure resistive load.  After these and knowing how small differencies there need it is easy to understand that 50MHz model is not same as 25 or 10 or 5 MHz model. Even if they looks like same they are not same (designed or not designed differencies - I do not want to take a stand.). 

Or take example Sine to Square circuit. Small differencies in components real values inside their tolerances and other unit have much more cycle to cycle time jitter specially with low frequencies.
I have investigated this thing inside SDG1000  quite deeply and also made some experimental's there in circuit. (results are allready published)

Solution is extremely simply: If need 50MHz model buy 50MHz model. If need 5MHz model, buy it but do not want after then want it still is like 50MHz model if you add fan to it.   

[MartyD]

Solution is extremely simply: If need 50MHz model buy 50MHz model. If need 5MHz model, buy it but do not want after then want it still is like 50MHz model if you add fan to it.   

I could not agree more! Besides the possible technical differences between them, you are also trying to steal intellectual property from the manufacturer by "unlocking" higher specs than what you paid for. And you do not pay alot for what you get in a Siglent.

Just my 2 cents

[tautech]

Authorized distribuitors (or anyone who offers a warranty)
(pending)

Myself & rf-loop plus other members.
I have a dozen SDG810's in the field with 3 year factory warranties.
Happy for you to link my website. 

[rf-loop]

Authorized distribuitors (or anyone who offers a warranty)
(pending)

Myself & rf-loop plus other members.
I have a dozen SDG810's in the field with 3 year factory warranties.
Happy for you to link my website. 

Yes.

I have n amount of SDG's and other Siglent equipments in the field with 3 year factory warranties.

But today also more than just 3 year "factory warranty".
(beginning 15.Jan. 2015)

Here, locally in Finland, of course with 3 year factory warranty as  Siglent 
(Siglent Technologies Europe GmbH
Liebigstrasse 2-20, Gebaeude 14
22113 Hamburg
Germany)   

have stated about warranty in European Union area.
( http://siglenteu.com/shfw.aspx?id=16 

For some models what I have accepted (due to longer time experience) I give also in Finland 1st year my own warranty extension to these units what have passed in my own Q.C. before sold. 

For customer it means that customer send or bring broken device to me after he first get RMA from me. Just simply, if equipment is  warranty failed (what warranty cover), customer contact me and afters sooner or later he have working unit. 

Depending failure, in co-operation with the Siglent be decided which is the wisest way to proceed in this individual case.

And of courseif customer have him self made failure doing some "oops..." type mistake, we can also repair these using Siglent spare parts. But this kind of service is not free.

[jimon]

@rf-loop
SDG1025 board photo
SDG1010 board photo
(Haven't found a pic for SDG1050)
Looks like exactly the same design to me. Maybe they indeed use a different components, but don't think so, why bother in a first place ? If Siglent do a design for 0-50 MHz DDS (which they do because SDG1050 is there), then it gonna work for 0-5 MHz without any changes at all, so there are no point for them to pay even 1 cent more because summary of all components + assembly costs for 50 MHz model likely to be smaller then selling price for 5 MHz model (remember how iPhone hardware was estimated around 200$, but selling price is 600+$).

The only question why they don't sell 50 MHz model for a price of 5 MHz, and an answer is quite simple - product development requires a budget (+office +revenue +tax +whatever), so each unit price contains one part for hardware, and another part is to cover development. Also they don't plan to sell a lot of them, so development costs per unit are quite significant. So in result unit is expensive, so they don't sell a lot because users aren't willing to pay so many in a first place.

So would you mind to show us hardware differences between models and prove that my assumptions are wrong

[Macbeth]

Every SDG1010 is different. Every SDG1025 is different. Every SDG1005 is different. Every SDG1020 is different. Every SDG1050 is different.

What? So if you open up a few SDG1025's they are all going to be different? Like they just randomly stick any old junk to hand at the factory? They don't have a standard schematic and service manual? This doesn't inspire me with any confidence. In fact I think I should just throw the thing in the bin and buy a proper arb gen if the Siglents are just bodged together like a dogs dinner with any old shit Won Hung Lo finds on the scrap heap that day. 

[tautech]

Every SDG1010 is different. Every SDG1025 is different. Every SDG1005 is different. Every SDG1020 is different. Every SDG1050 is different.

What? So if you open up a few SDG1025's they are all going to be different? Like they just randomly stick any old junk to hand at the factory? They don't have a standard schematic and service manual? This doesn't inspire me with any confidence. In fact I think I should just throw the thing in the bin and buy a proper arb gen if the Siglents are just bodged together like a dogs dinner with any old shit Won Hung Lo finds on the scrap heap that day. 

Shit Macbeth, did you not take the time to read rf-loop's post carefully. 

These units perform well against compeditors because care is taken they WILL meet specifications.

Some units whole circuit (mid way amplifiers, filters, attenuators and end stage freq response is different, not only designed different but also due to all componets tolerances (and it need also know component values drift after ageing).   

Check this thread to see how my 10 MHz unit performs against similar units:
https://www.eevblog.com/forum/testgear/show-us-your-square-wave/

[rf-loop]

Every SDG1010 is different. Every SDG1025 is different. Every SDG1005 is different. Every SDG1020 is different. Every SDG1050 is different.

What? So if you open up a few SDG1025's they are all going to be different? Like they just randomly stick any old junk to hand at the factory? They don't have a standard schematic and service manual? This doesn't inspire me with any confidence. In fact I think I should just throw the thing in the bin and buy a proper arb gen if the Siglents are just bodged together like a dogs dinner with any old shit Won Hung Lo finds on the scrap heap that day. 

No. 

But still every single unit is different.

Every single resistor, every single capasitor, every inductor, every amplifier and so on, even every PCB, all they are different, more or less.

Every component have its tolerances. In some cases all components are quaranteed inside tolerance limits and in some cases it is guaranteed 2SD  is indside limits or what ever is spesified or promised.

Product have lot of these components. As all componets are random (inside limits) there is lot of combinations.

Take example one filter. There is perhaps many resistors, capasitors and inductors... do 100 pcs with real designed components and after then measure every filter shapes - do you believe they are same and you find 100 individual filter.  Analog electronics world is...

If we do digital things...  you take component what make 8 and 8 bit AND. You make 100 these circuits...  and if not total failures they all give same result.

In SDG there is lot of analog circuits. Amplifiers, comparators, attenuators, filters...

It is possible to make 100 SDG1050 and find that some SDG1050 is example "borderline" with specs but if sell as 1010 it is ok. So... (I do not claim they do just this but also it is possible... and if I do products I want use also this... )

Possible differences in values of some  components in different models, I do not want and I can not take a stand.
There may be some differencies also in different HW versions and different manufacturing lots.

[nctnico]

If product can not meet some (company internal) specifications in product test for example 1025 model what to do.  I think no one believe they go to recycling.  So what to do?  First idea what every people can himself think is try if it still meet some other model specs.  If it meet SDSG1020 factory internals test specifications so that it can be this model, then it is just ok.
I do NOT KNOW if this is true in Siglent case but if I do this kind of products I do it.

That makes no sense. You are not going to produce a product and hope it meets specs and sell it as a lower specced if it doesn't. That way you can't be sure you have enough units to sell. There must be other differences like component values between the low frequency and high frequency unit. I'm pretty sure there is a model number encoded somewhere either as a jumper option or somewhere in a flash. The output circuitry can also be different to have a lower frequency roll off.

[rf-loop]

If product can not meet some (company internal) specifications in product test for example 1025 model what to do.  I think no one believe they go to recycling.  So what to do?  First idea what every people can himself think is try if it still meet some other model specs.  If it meet SDSG1020 factory internals test specifications so that it can be this model, then it is just ok.
I do NOT KNOW if this is true in Siglent case but if I do this kind of products I do it.

That makes no sense. You are not going to produce a product and hope it meets specs and sell it as a lower specced if it doesn't. That way you can't be sure you have enough units to sell. There must be other differences like component values between the low frequency and high frequency unit. I'm pretty sure there is a model number encoded somewhere either as a jumper option or somewhere in a flash. The output circuitry can also be different to have a lower frequency roll off.

I have no any idea for dispute what is case exatly in Siglent with these.

This is not at all relevant with this case but still intersting (and really very extremely normal):

Take example this fun about one "famous" ADC.
ADC is designed and produced for 100MSa/s
AD sell them 100, 80 and 40MSa/s versions. Circuit is same, designed components in the chip are same...

If I do SDG1000 Idesign  do all same 50MHz models.
I then do some very trivial separation using some automatised test.

What ever day If need sell more SDG1010 what are not in this fraction allready. Just take from 50MHz fraction and set model for SDG1010.

But I have no any idea how Siglent do and I have also see some differencies...,  if someone ask "I do not remember"

[nctnico]

Chip yield is something different because the process is hard to control. You can't compare that to a complete electronic circuit.