EEVBlog #805 - Siglent SDG2122X Arb Generator Teardown

[EEVblog]

The new Siglent SDG2122X X-Series Arb Signal Generator Teardown
Now in 4K resolution option.

Trio Test & Measure: http://www.triotest.com.au/shop/function-generators/3820-siglent-sds2024x-40mhz-2-channel-arbitrary-waveform-generator-1gsas.html

Datasheets:
http://www.ti.com/product/am3352
http://www.analog.com/media/en/technical-documentation/data-sheets/AD9122-EP.pdf
http://www.ti.com/lit/ds/symlink/ths3095.pdf
https://www.altera.com/content/dam/altera-www/global/en_US/pdfs/literature/ds/en2342qi_09520.pdf
http://www.ti.com/lit/ds/symlink/opa695.pdf
http://www.analog.com/media/en/technical-documentation/data-sheets/AD8510_8512_8513.pdf

[Armxnian]

Is this 1080 upsampled to 4k during your render?

[lukier]

Good teardown, thanks for taking the trouble to decode all the parts in the vital parts of the device. I've spotted DAC8580 (designator U27) I think, this will be probably auxiliary DAC for the offsets. Pretty decent in it's own right, 16 bit monotonic, 16 MSPS.

From the hardware point of view it looks like a very capable instrument. All thanks to the advances in telecom technology . Most of these high speed ADCs and DACs that we see nowadays in the budget gear has its origins in developments in RF for 3G/4G or core network (fibre optic, DWDM etc) equipment.

It's a pity it is above my budget, so I'll have to keep using my Hantek HDG2002B.

The only question left is: how is the firmware? My Hantek from hardware point of view is also fairly decent, especially for the price (~$300), but the firmware is absolutely awful 

[EEVblog]

Is this 1080 upsampled to 4k during your render?

Yes. Explanation in a video uploading now.

[AF6LJ]

Thanks Dave Great job.

[Radio Tech]

Pretty nice built kit for the price.
Like to see a performance review.

[Stephen Durr]

Forgive me if this question has been asked before, but can anyone elaborate on why the q.c. stickers on the back do not instill confidence? Dave mentions this at 1:54 in this video and I noticed the same comment at about 0:53 of the tear-down video for the Siglent 1000X, and I'm pretty sure I've heard similar comments about stickers on other product tear-down videos in the past. So what is it? Are the stickers considered cheesy? Do they make you suspicious of the manufacturer's QC process?     Sorry if this is basic but I'm a noob 

[tautech]

Forgive me if this question has been asked before, but can anyone elaborate on why the q.c. stickers on the back do not instill confidence? Dave mentions this at 1:54 in this video and I noticed the same comment at about 0:53 of the tear-down video for the Siglent 1000X, and I'm pretty sure I've heard similar comments about stickers on other product tear-down videos in the past. So what is it? Are the stickers considered cheesy? Do they make you suspicious of the manufacturer's QC process?     Sorry if this is basic but I'm a noob 

Yeah I feel these snipes are just cheap shots too.
They show the instrument has been checked against published specs.

But what really matters is the Cal certificate which is the official record of Quality Control and proves the instrument meets its design spec. Without a Cal cert as the official record, sure anybody can wack whatever sticker they like on gear and no doubt some do.
Let's hope Dave does spends a moment in future vids revealling what equipment DOES NOT come with a Cal cert, that will be more meaningful than mention of a particular "looking" sticker.

@ Stephen Durr
You might have noticed Dave did mention the additional Cal label on the SDS1000X series and this label seems like a new addition to the Siglent range.

[FaultlineAudio]

Great tear down Dave.  I'm in the market for a Arb Gen and I had though of buying the Rigol DG4162 until I saw the Siglent in the Oscilloscope review you did earlier.

I hope you do a review/comparison video between the two models.  The Siglent seems to be a better, higher quality choice than the Rigol.

Thanks
Chris

[c4757p]

Forgive me if this question has been asked before, but can anyone elaborate on why the q.c. stickers on the back do not instill confidence? Dave mentions this at 1:54 in this video and I noticed the same comment at about 0:53 of the tear-down video for the Siglent 1000X, and I'm pretty sure I've heard similar comments about stickers on other product tear-down videos in the past. So what is it? Are the stickers considered cheesy? Do they make you suspicious of the manufacturer's QC process?     Sorry if this is basic but I'm a noob 

They are a style of QC sticker that is well known for being used on cheap crap. It's entirely a subjective comment, but I feel the same way - those stickers immediately make me think of crap products where the only quality control test is "does it fit in the box?". They remind me of a soldering station I bought once that had a sticker on it that said, I shit you not, "QENUINE". The transformer went up in smoke after a couple months, naturally.

Siglent are improving in actual, internal quality, but they still have a way to go before the instruments look like it. Seriously, I want to peel off the sticker on the front just to hide that MS Office 97 WordArt-looking "True Arb" label! Sure, it's subjective, but looks are more important to make a sale than we engineering-types often want to admit.

[Armxnian]

I agree that they make the product look cheap. They are also meaningless. Both my Siglent units have the stickers, both of them have rust...

[SteveLy]

Those QC Passed stickers are seen on the cheapest dodgiest $2-shop items, so indeed they don't instil any confidence

[Smokey]

The whole QC procedure is sticking stickers on.  If they stick...

QC PASSED!

[daqq]

Hi Dave, thanks for the teardown.

The EN2342 is expensive, but all you need for it is a few caps and resistors on the output - it has an integrated coil inside. The weird shape on the bottom is the coil winding. Basically it's the equivalent of the LMZ series from Ti.

[Aaguilar4]

Thanks for the teardown Dave. I,ve been a huge fan for at least three years, but this is my first time on the forum.
I never undrstood who would want to buy these power modules like Enpirion before. They are very expensive compared with building your own 4A converter, and all they offer really is integrated magnetics (also the reason why the package is so thick).
It seemed to me like this was a good idea for prototyping and gettig something out the door quickly. But for serious products built down to a price, doesnt make much sense at all.

My guess as to the ground plane is two-fold:
1st) thermal dissipation of power loss in the FETs and the inductor
2nd) low parasitic inductance of ground connection to the low-side switch in buck converter. In this type of package, the die is flipped, connected directly to the leadframe, which gets rid of bondwires typically found in QFN. These bondwires add parasitic resistance and inductance. in the case of this Enpirion part, large strips of PCB copper connect pretty much directly to the switches inside your die.

All the best and thanks for all your videos!
Alvaro

[GNU_Ninja]

Those QC Passed stickers are seen on the cheapest dodgiest $2-shop items, so indeed they don't instil any confidence

Just had a look at the back of a UNI-T UT58A multimeter that I brought into work to test. Sure enough, there's a gold QC passed sticker on the back. The company that makes those stickers must be coining it 

Mind you, it also says CAT II 1000V rated on the front of the UNI-T. Yeah, right 

[daqq]

I never undrstood who would want to buy these power modules like Enpirion before. They are very expensive compared with building your own 4A converter, and all they offer really is integrated magnetics (also the reason why the package is so thick).
It seemed to me like this was a good idea for prototyping and gettig something out the door quickly. But for serious products built down to a price, doesnt make much sense at all.

Actually, they are quite useful for a lot of reasons:

1. Just-add-water philosophy. All you need is a few extra components which you can source from pretty much anyone with very little fuss. Compare it to rolling your own: Source the inductor, source the caps, the main IC, diodes, FETs... Here all of the critical components are given to you.
2. Optimized spatially - you'll have to try very hard to outperform things like these modules in terms of power to volume ratios. For a similar construction you'd need more board space and you might still not get the same specs. See the http://www.ti.com/ww/en/simple_switcher/qfn-package-modules.html#nano nano series for an extreme of this.
3. Optimized performace wise - to get a good DC DC converter, all you need to do these days is slap together a few components that have close-enough specs. To get a really good DC DC converter you have to pick your components for quite some time, test them, poke and prod them, test the whole setup... In these modules someone did it for you. As they did the EMC testing. Also the mechanical design. Essentially, this: http://www.ti.com/lit/ds/symlink/lmz31530.pdf gives you 30 Amps from a package the size of a 30Amp inductor, without any hassle.

Yes, at the end of the day you are just saving time. They offer nothing you could not do on your own, eventually, with the right tech (which might get costly). But it can (depending on the situation)save you time, board space, efficiency.

edit: This one's nice as well: http://www.intersil.com/content/dam/Intersil/documents/isl8/isl8273m.pdf

[EEVblog]

Thanks for the teardown Dave. I,ve been a huge fan for at least three years, but this is my first time on the forum.
I never undrstood who would want to buy these power modules like Enpirion before. They are very expensive compared with building your own 4A converter, and all they offer really is integrated magnetics (also the reason why the package is so thick).

Welcome!
What this part offers is a guarantee that it's going to work with the Altera FPGA they are using (think obscure power-up current spike issues etc). It's proven, it's no fuss, one less thing the designer has to worry about about. Quite common practice. They obviously aren't shaving every last cent off the BOM cost.

[Fungus]

Forgive me if this question has been asked before, but can anyone elaborate on why the q.c. stickers on the back do not instill confidence?

Maybe a car analogy would help:

How would you feel if you bought a new car and it had one of those stickers on the back?

[Godzil]

Thanks for the teardown Dave. I,ve been a huge fan for at least three years, but this is my first time on the forum.
I never undrstood who would want to buy these power modules like Enpirion before. They are very expensive compared with building your own 4A converter, and all they offer really is integrated magnetics (also the reason why the package is so thick).

Welcome!
What this part offers is a guarantee that it's going to work with the Altera FPGA they are using (think obscure power-up current spike issues etc). It's proven, it's no fuss, one less thing the designer has to worry about about. Quite common practice. They obviously aren't shaving every last cent off the BOM cost.

Yeah but in professional product, price of component doesn't matter right? I mean they try to made the best tool that work well in all conditions, no matter the price!

What, why someone tells me that some company doesn't care about quality and just do that for money and try to make the cheapest possible device?

[Earendil]

Could someone explain why this instrument only have 20Mhz arbitrary wave output, when its DAC is running at 1230Mhz, please?
I feel confused. 

[rf-loop]

Could someone explain why this instrument only have 20Mhz arbitrary wave output, when its DAC is running at 1230Mhz, please?
I feel confused. 

Very much simplified for imagine what is going there.
What you thing if Arb memory have 5 cycle on sine wave. Then you set Arb frequency 20MHz. What is output frequency. It is 100MHz.   Oh but Arb waveform max is 20MHz. Yes, but  just One Arb waveform have 5 cycle in this case..


20MHz is how fast it can repeat whole Arb memory (it also depends how many Arb memory points waveform have)

For thinking. You have 100 memory points long Arb waveform what you have made. Then you set Arb frequency to 10MHz.  How fast it need run from point to point in arb memory. What is stepping frequency. 10MHz x 100=1000 MHz

[ivan747]

Forgive me if this question has been asked before, but can anyone elaborate on why the q.c. stickers on the back do not instill confidence? Dave mentions this at 1:54 in this video and I noticed the same comment at about 0:53 of the tear-down video for the Siglent 1000X, and I'm pretty sure I've heard similar comments about stickers on other product tear-down videos in the past. So what is it? Are the stickers considered cheesy? Do they make you suspicious of the manufacturer's QC process?     Sorry if this is basic but I'm a noob 

This exact model of sticker is often used in some of the cheapest, lowest quality products China has to offer.

[ivan747]

Forgive me if this question has been asked before, but can anyone elaborate on why the q.c. stickers on the back do not instill confidence? Dave mentions this at 1:54 in this video and I noticed the same comment at about 0:53 of the tear-down video for the Siglent 1000X, and I'm pretty sure I've heard similar comments about stickers on other product tear-down videos in the past. So what is it? Are the stickers considered cheesy? Do they make you suspicious of the manufacturer's QC process?     Sorry if this is basic but I'm a noob 

 They remind me of a soldering station I bought once that had a sticker on it that said, I shit you not, "QENUINE". The transformer went up in smoke after a couple months, naturally.

New signature right there.

[nixxon]

Airflow
4:36 into the video, it is commented that the ribbon cables are blocking the vents.

Maybe they deliberately block the upper vent holes with the ribbon cables to increase the airflow onto the components which are located at the bottom.

Between 3:55 and 3:58 into the video, the instrument was starting to fall down. Let's hope it didn't fall too hard.