EEVblog
HTML-код
- Опубликовано: 27 сен 2024
- A very detailed look inside the new Siglent SSA3021X 3.2GHz Spectrum Analyser.
How does a spectrum analyser front end work?
The entire RF section is broken down and analysed in depth with a block diagram overlay and signal flow.
Forum: www.eevblog.com...
Keysight AN150 App Note: literature.cdn....
Datasheets:
www.ti.com/prod...
ww1.microchip.c...
www.ti.com/prod...
www.analog.com/...
www.analog.com/...
www.analog.com/...
www.analog.com/...
www.physics.uto...
www.analog.com/...
www.analog.com/...
www.minicircui...
www.analog.com/...
www.analog.com/...
www.zcomm.com/...
www.analog.com/...
www.analog.com/...
www.analog.com/...
www.minicircui...
www.winnsky.com...
www.psemi.com/p...
www.analog.com/...
www.analog.com/...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
EEVblog Amazon Store (Dave gets a cut):
astore.amazon.c...
T-Shirts: teespring.com/s...
💗 Likecoin - Coins for Likes: likecoin.pro/@...
That block overlay of the PCB to help understand the design was *really* helpful! Thanks Dave!
Thanks, it took me a lot of hours to do that, but I thought it was worthwhile.
Awesome idea the block overlay 100% worthwhile for me. Very nice video and teardown, Thanks a lot Dave :D
Ya I couldnt agree more. One of the best teardowns ive seen just because you approached it that way this time.
Holy crap that block overlay was insane!!!! :) 22:21
Yeah, that was awesome! Love it :D
"3.2 GHz it's practically DC" - Dave Jones, 2016
The funny thing is the more time that passes the more true this will be
"The funny thing is the more time that passes the more true this will be"
Except, you canna change the laws of physics, captain!
Wie ich so Bullshitsprüche nicht leiden kann! Interessanterweise kommt sowas immer von Leuten die sich mit HF-Kenntnissen rühmen und wenn man dann genau nachfragt nur Dünnschiss kommt!
>Except, you canna change the laws of physics, captain!
I work in DSAs...this is very pertinent to me!
As everyone stated, best teardown ever, very good walk through the diagram blocks! Awesome!
Thanks. Worth the extra effort.
I second that
Yes. Great job on this one. Thanks!
Indeed. Best teardown ever.
This was one of my favorite teardowns you have ever done (ive watch almost every video you have posted), I learned so much on this episode. The way you matched the block diagram to the board then walked through it was phenomenal. Loved it! Thanks! More like this!
Thanks. I liked the result, will try to do more like this in future teardowns, but it takes a lot more time unfortunately.
Well the additional effort was noticed and appreciated. Thanks again.
Lots of work put into this teardown, absolutely awesome!
Yep, many more hours than usual, hence the lateness.
And how many chicken dinners ?
"I'm gonna have to eat every chicken dinner in this room."
Dave! If you are having trouble reading laser engraved part numbers, just wipe on a TINY dab of thermal paste with your finger and rub it off. The white paste stays in the etched areas leaving much higher contrast markings.
This is probably one of the best teardowns i've ever seen. Very informative! Thanks Dave!
Definitely! Big thumbs up, Dave!
I totally agree. I really appreciate all the work you put into this, Dave!
Thanks!
That was a very very positive review and teardown of that test gear which I found slightly surprising from you Dave :-) I hope the engineers and people involved in the project are watching this video looking at their work with pride. Great video thanks for posting.
I just got an email from Siglent thanking me, they liked it.
Those photos you took with explanations were incredible. One of your best to date.
You've always been the best, but you just upped your game!
Thanks.
Dave, this teardown and explanation were positively brilliant! This is your best yet. I liked how you superimposed your segmentation of each module of the main board and traced through the signal flow. I know how much work this was but I would rather see fewer, more detailed analyses rather than quick and simple. Seems you set the bar pretty high for yourself! ;-) But try not to be a victim of your own success!
One more thing: some of the viewers here may appreciate a tutorial like you used to do on Fundaments Friday. I recommend, if I may, you do one on mixers, heterodyning, and filtering. In this case, a quick mostly maths-free overview would be beneficial.
Thanks again, mate, and keep up the good work! I look forward to your next video!
Dave,
One of the best videos you have ever done. It was like listening to a college lecture on spectrum anaylzers! Loved it.
Best teardown you've ever done! Please make every teardown this detailed. I'd be happy to wait longer for them. Thank you so much!
I thought this was a briliant teardown Dave! I really loved the way you explained the PCB in blocks. Chapeau!
Dave, your electronics knowledge is genuinely astounding!
Wow! Big thanks for the extra effort with block diagram overlay and the detailed descriptions!
Glad everyone liked it. I was curious how it would be received.
I agree with you dave, a really nice design, i love modular systems too :-D.
Smart and clean sectional layout, well worth the pennies, you can see the effort and thought put into the design.
Imagine all equipment designed this way, im salivating LOL.
Nice vid Dave! Deep analyze of these RF guts is always welcome. Keep doing what U do!
Wow! Thanks Dave for the very detailed explanation of the functional blocks! Excellent functional overlays! Highly encourage you to make this more often! +1 :)
Thanks, I'll try.
oh wow, this video is awesome. A lot of effort went into the illustration. Excellent. Thank you Dave.
Thumbs up for the circuit overlay. If we can get this for other videos then brilliant!
Things like scopes are not so exciting, not as many building blocks to analyse.
That processor's JTAG is the header at 7:40 which you thought might be just for testing - it's the standard 20-pin ARM JTAG header. The other one just does the Spartan 6.
The DC-DC immediately below the 20-pin is most likely an ADP5052 - it's a quad DC-DC + 1xLDO specifically designed to power Spartan 6's + DDR2/3. It's capable of providing 3.3V, 2.5V, 1.8V and 1.2V simultaneously with sufficient current to power even the top of the line XC6SLX150T.
Thank you for the overlay and detailed analysis! It was definitely worth the wait! Great teardown!
Appreciate the extra effort in this one Dave.
Amazing filming and editing, really aided with understanding. Thanks Dave.
Thanks for the video Dave, i just may end up with one of those SSA3021X on my bench in the near future.
That block analysis was great, thanks a lot
For me is still woodoo.. Thank You Dave for open and show to Us. We again learned a lot! Lot of woodooo at least :)
Love the setup of this video! Great structured way of presenting it. Cheers!
Thumb up on the datasheet linking 👍👍
This video is so GREAT!!!! Really nice explanation for spectrum analyzer.
I like the new format of explaining the whole signal path. Could have been better if you could point at things. When you mention a part descriptor, it takes a little time to locate it by eye
Yeah, that's one downside of this, it's not easy to poke. I'd have the screen capture instead and use a big cursor. But then I wouldn't get the nice zoom feature which is done in edit.
Thanks and Recommended for all
Fantastic video! Loved your block analysis
Thanks Dave very useful mine came with a loose fitting faulty headphone jack plugging in an audio 3.5mm for a headset very loose fitting and almost useless contacted them via the agent they said it was a standard 3.5mm only option was to repack it it and send it back I only had it for 5 minutes so decided to do it myself your video was very useful love your vids buyer beware of this model if deemed important Dave UK
Great teardown! Interesting that there is an Altera FPGA on the back side.
Great teardown Dave! I loved the way you went thru the entire signal path. The part where you mentioned that the final IF frequency being less than 20 MHz is not correct. There is a phenomenon called "subsampling" where the sampled signal is at a higher frequency than the Nyqusit frequency. However the bandwidth of the final IF signal is less than 20 MHz, which is probably much less than 20 MHz since resolution BW of this SA is specified as 1 MHz.
Excellent work Dave!
Great teardown, Dave!
Fantastic video, great job Dave!
Dave, just wanted to let you know that you forgot to add the link at 0:08 when you said, "click here".
good job explaining the RF section!
I remember a spectrum analyser at polytech back in the 1980s. It was the size of a small bus.
I have always wondered why they start at several kHz, from a design POV. This means I need one specifically for when I want to do some audio stuff, unless I have one of the super expensive 20Hz - 20GHz R&S ones
I suppose that's done because the phase noise becomes too great for general use below a few kHz.
Aars Haar maybe too great for the price target... R&S ones go down til 2Hz ...
Dennis Lubert I wish I could say I knew more about spectrum analyzers.
I suppose they must be applying a trick to get down to 2Hz (which kind of blows my mind), maybe two signal paths for two halves of the total frequency range?
Aars Haar I never looked too much in detail into them, maybe I should next time I have that opportunity
They probably dropped the input protection because its integral capacitance interfered with the unterminated input BW linearity. At GHz it's damned hard to get a flat response with peaky capacitors (ie diodes) in circuit when there's no 50Ohm termination.
Agree
Wow! Great Teardown!
Awesome video. Thank you.
Great tear down video! Can you give us a tear down of the Keysight N9030A PXA so we can see a comparison to some of the higher end spec'an gear?
Aren't things slightly mixed up in the explanation starting from 27:30? Unless I'm mistaken, the reason for the multi-stage design is not primarily the need for a higher IF. You could use an IF of, say, 5.1 GHz in a single-stage mixer design just as well. It's just that doing narrow-band filters at that high frequency (for, say, an RBW of 10 Hz) is rather difficult, and of course an all-digital IF stage would then be rather impractical. Thus, you just have a pre-filter after the first mixer, but then mix down again, filter again, etc.
Great video.
Is it possible that they found a way to optimize the VHDL core file that goes on that big Spartan FPGA such that they were able to eliminate the missing components discussed at around 8:00 and pack them into the FPGA, but it was late in the design process and redesigning the board would have to wait until the next revision?
Good point, because it would just be shaping the IF signal that goes into the ADC and FPGA anyway.
I think that the system will run really when you 16384bins rather than going any higher which causes it to slow down, I just wonder why they could not have used a A10-7850k to process the data and they would have to have any more graphics hardware as that would be on the A10-7850k and they could have written the firmware in ASM80586 for the A10-7850k, why did they not use a quadrature phasing mixer which which would be able to convert the RF to AF (direct conversion) thing and give a nice clean output.
Dave, there are no zeds in "chassis". Bloody Aussies! (Said Alan from NZ.) :-)
5:59 - Crooked IC U104... everything else looks squared up but not that IC.
™Siglent Rust
It's both trademark for feature and colour- just try using one in your product and surely you're gonna get sued!
The mains filter cap being a Lelon isn't THAT bad, but there's one on the secondary too - WHY?!?! >_
I would quite like to see you design your own spectrum analyser, or any other type of test gear for that matter.
I'd suck at designing a spectrum analyser, RF design isn't my thing.
Anything other than RF then? I've been on the hunt for open-source lab gear and I feel like you'd be the perfect guy to design some.
this the click here belated April fools? Liked the block overlays
Oh, I'm not so sure it's "practically DC" if you have to use GaAs components...
Hi Dave, thanks for great video :). I want to discuss about 49R9 resistor (23:43), and what is purpose of this resistor. This resistor is protection resistor, and SA will switch input this this resistor to protect input path from large signal (Am I right?). All other component in input path has 50 Ohm impedance, and no need for extra resistor.
shaziiiii 😂 thank you btw
I missing only one label from the back cover... Yolo!,,,, :))))))))
1:33... No it is really easy to see... it must be a feature at that price :/
Do you shoot the actual teardown portion on the Tagarno?
Is that a micro SD slot at 5:01?
Many of the thick traces on rf board( presumably 50 ohm rf traces) have their solder or/and silkscreen removed on them and around them. Is it a common practice in RF design. Also all of the distributed component filters or matching circuits dont seem to have solder resist and silkscreen on them. Is it neccesary to do that when designing Rf board? Do silkscreen and solder resist have huge impact on the charateristic impedance pf the trace and distributed component?
it's easier to get consistent characteristic impedance with no solder mask. Slight changes in soldermask thickness would change the impedance of the trace. Soldermask may also change over time due to environmental conditions.
Thanks for the response. In this particular device the solder mask around the paths are not removed all the way down they're just removed from the sides about the twice width of the line. Is there a rule of thumb for the size of solder mask opening around the path?
Is that Micro Sd slot by USB port?
yep, close to the future resistive touch flat flex cable connector xDD
+Jotaemebé thanks
👍
Winner, Winner, Chicken Dinner
Hi Dave, Long time lurker, and I love your channel! I have a question if I may have a moment of your time. My son has begun to take a larger interest in these DIY kits available on Amazon (ie. ASIN#B0130LKYWE, B00XXZB5LW), and I was wondering what your opinion was on these for learning electronics. Myself, I recall the old "spring-based bread boards" from back in the day, but these kits seem to have great potential. Do you have any suggestions in particular for these kits, or possibly a recommendation for an "actually useful" DIY o-scope kit? Thanks in advance!
BULLSHIT ALERT: has anyone heard about "wireworld" cables? they sell all kinds of ultra expensive cables, i'd love to see Dave crush them with some hard science!!!
2:48 Bad soldering on that fat cap though.
Yes, did not look too good.
Its soldered on the bottom side, nothing wrong with that.
+gulftown
Usually soldered on both sides
I'm amazed how cheap the Chinese get milling done compared to casting when you look at the west. I worked at an RF company that was designing a satellite receiver and they talked about paying $800 for a one piece milled enclosure in low qty and $400 in small quantities yet getting investment castings done was viable at 50 pieces. They must be getting the milling done REAL CHEAP.
Also I suggest you use cotton gloves when doing tear downs as touching the aluminium causes almost immediate corrosion with skin oil contamination. This then in turn makes for a poor RF connection when it's reassembled.
aluminium oxidizes immediately on contact with air. No touching required.
Rhys D
Yes that is true but the thickness of the oxide layer is heavily dependant on the chemicals present. If you touch it with your bare hands you increase the thickness and chemical nature of the oxides depending to some extent on your biology and skin temperature. It makes a difference for RF seals where 1 Ohm is 2% reduction in conductivity. Companies don't like to waste money on unnecessary thing so they wouldn't buy gloves unless they made a difference.
This was an absolutely awesome teardown, so well done, so much detail. Love it!
If DC is so bad for Spectrum Analyzer, why are DC blocks not built into the unit?
Very good, hope to see more teardowns as detailed as this :)
What is the total value of all the gold in this thing ?
5:30 - 5:40
"nothing exciting unless SMPS float your boat, dead giveaway with the inductors"
Or....Dead giveaway with the SMPS silkscreen labels! haha :P
The final IF doesnt necessarily have to be in the 0-20MHz range; the ADC can undersample a different nyquist zone provided an input with appropriate passband filtering. I'm not suggesting they do this, but it is certainly possible. Great teardown video though from Dave, thanks!
I've tried to understand quadrature decoding, carriers, Fourier transformations, and intermediate frequencies, but I am still confused. Some day, I don't know.
Me too. Every time I think I have a superficial understanding, I try to apply it and find myself confused again. Should have taken more EE classes. Maybe studying more ham radio materials will help.
I always resisted against the idea that really everything comes from sine waves, so things could come from triangle waves for example. From the point of view that everything is really sine waves, a triangle wave for example has many frequency additives to make that resultant waveform. Quadrature sort of makes sense then because at a single instance a cos and a sin interpretation of a signal is made together, and then more data in time. So I was confused how it seemed arbitrary what was being sampled, but with data in time, it is relative and meaningful?
Also I was worried that carrier encoded data would in fact collide. You can say in a frequency graph there is a space between the channels where the data encoded is. That is my understanding. It seems a carrier vs the data is almost so arbitrary that it would go everywhere.
Maybe this will help. m.ruclips.net/video/LznjC4Lo7lE/видео.html
The obsolete Hitite parts may only be so because Analog Devices bought them out and the naming is in flux. They are probably changing it from Hxxx to ADxxxx :-)
Didn't let it fall down again haha
BEST BLOODY DISASSEMBLER I KNOW!!! Love your vids
I absolutely love this block diagram - real PCB - type of explanation :-)
I hate BGA chips!! This is a shame for electronic industry!
they should have put a z before the X then it would be the ZX Spectrum Analyser!
What a tremendous effort and excellent result on this video. Thank you!
Nice video ... liked the city tour on the circuit with road signs :-)
Great video ! Lots of useful info in there. Thanks Dave.
A RF capacitor can act like a inductor when it's connected by a quarter-wave line. Does anyone know more about this?
Smith chart will help you understand ;) Mark a capacitor's reactance and plot an arc corresponding to a quarter wave line. Now, because you have made half a revolution, you're on the other half, with opposite sign of reactance.
now you void your warntie on that Spectrum Analyser or have you get it for free
I goes back to it's owner.
Siglent Mona Lisa electronic art.
I like the photos you use in this video!
Wow Dave, this video is incredible! Much respect!
P.S. Videos like that remind me exactly why I got into electronics:)
Great review Dave of the new Siglent SSA3021X. Nice to see you finally got one to review, I'm sure Siglent will be happy with the results. And us radical fanboys will be rushing out to get one. So I look forward to the next installment of the EEVblog. CHEERS !
Don't kid yourself, likely Siglent sent this unit directly to him, I could imagine Dave has a lot of corporate insiders backing him for publicity. "And us radical fanboys will be rushing out to get one." Doubt it, that unit costs over $ 1,500 new who rushes out to spend that kind of money?
I would if the price is right, but due to me living in Canada. The taxes, exchange rate and import charges would make the price double. I bought an Signal Hound SA44B from the US on Ebay @ $560 USD my landed cost was $964 CAD. But it seems like a nice bit of kit, I like to wait for 6 months or more to see other reviews. As for my comment as towards radical fanboys, I watch too much American politics. Cheers ...
Very good, yes the politics oh my!
WOOOOO