New Rigol RSA800 Spectrum Analyser

[EEVblog]

https://www.rigol-uk.co.uk/products/spectrum-analysers-vnas/spectrum-analysers/rsa800-series-spectrum-analysers/

In the tiny DHO800 series scope form factor.
4GHz RSA804 is £1,650.00 (£1,980.00 inc. VAT)

[EEVblog]

Oops! 

[Cyclotron]

Oops! 

You know, the Measyrement, its right there between the Armpit and the Mode Meas.

Waiting for daves screwdriver.......

[tszaboo]

It's a real time analyzer, and it includes EMI option.
Comparable from them is the RSA5000 series:

RIGOL Spectrum Analyzer RSA5065N
Max. Ferquency Range 9kHz~6.5GHz
7914 EUR

RIGOL Spectrum Analyzer RSA808
Max. Ferquency range 5 kHz ~ 8.5 GHz
4163 EUR

So, higher bandwidth for half the price...

[Aiy]

I'm curious to see the hackability and hardware construction on these..

On normal RF test gear there normally are different hardware on different frequency tiers, but with Rigol you don't know..

Surely the analysis options are gonna be software, but the frequency options?

I guess we'll have to stay baited until Dave gets one to take apart

Best regards,

[EEVblog]

First shot at a reverse engineered block diagram.

Only thing I don't get is the PCB couplers that seems to be 180deg phase shifting the first 1st IF back into the LO input of the lower frequency path SIM221 mixer. And then also switching it back to the input 
Double checked it, but I could be wrong.
I can only presume this is for self calibration/testing.

Teardown photos:
https://www.flickr.com/photos/eevblog/albums/72177720335391455

[tv84]

Looks nice. Like the heatsink.

Let's wait for the performance.

[Aiy]

Thanks for the photos Dave,

Interesting with the unpopulated footprints.. from the look of it I'd say the VNA shares the same board, just with different components on it.

I was not expecting to see the TI LMX2820 in there, premium component - that's a surprise.

Can you pop a UART on the pin header and post some logs?

Boundary scan on the JTAG?

Looking forward to the Video when it comes out!

[EEVblog]

[Martin72]

I've never seen such a simple yet impressive example of the difference between SA and RTA as in this video with the remote control.

[pdenisowski]

Oops! 

Actually, I think I saw that spelling in Chaucer: Here bigynneth the Teste and Measyrement Enginerees Tale ....

[pdenisowski]

I've never seen such a simple yet impressive example of the difference between SA and RTA as in this video with the remote control.

Lots of RF field applications engineers (myself once included) have used that same trick for various demonstrations when they don't have a signal generator handy

Another fun thing you can do with most keyfobs in the United States is place them next to a GPS receiver to simulate over-the-air impairments or jamming.  Keyfobs here are usually at 315 MHz, the 5th harmonic of which is the GPS L1 frequency (1575 MHz).  If you place an American key fob right next to a GPS receiver antenna and keep pressing the button, most receivers will lose lock within about 30 seconds, at least in my experience.

(Keyfobs usually have relatively poor harmonic suppression)

[EEVblog]

[CC12068]

no sd cards this time... it's an emmc
bad news maybe....?

[Stanislaw]

First shot at a reverse engineered block diagram.

Only thing I don't get is the PCB couplers that seems to be 180deg phase shifting the first 1st IF back into the LO input of the lower frequency path SIM221 mixer. And then also switching it back to the input 
Double checked it, but I could be wrong.
I can only presume this is for self calibration/testing.

The main function of the zero‑IF suppression circuit is to cancel the local‑oscillator leakage signal coupled into the intermediate‑frequency path, so that the spectrum analyzer achieves improved noise floor performance in the very low frequency band.

[EEVblog]

The main function of the zero‑IF suppression circuit is to cancel the local‑oscillator leakage signal coupled into the intermediate‑frequency path, so that the spectrum analyzer achieves improved noise floor performance in the very low frequency band.

Thanks. Is this because the SIM221 isn't that great?
I presume a better/more expensive mixer wouldn't need this?

[Stanislaw]

No, it’s not simply because of the SIM221. The LO-to-IF isolation of all mixers is limited. At low frequencies, the LO frequency falls within the IF bandwidth, which affects the noise floor in the low-frequency range. In order to achieve extremely low measurement sensitivity, high-performance spectrum analyzers all have zero-frequency suppression circuits.Some low-cost spectrum analyzers simply do not target very high sensitivity in the ultra-low-frequency range, so they omit this additional zero-frequency suppression circuitry.

[eb4fbz]

The IF coupler orientation is wrong.

Anyway, I’m not sure if that can really work to cancel the LO to IF leakage if you can only adjust the phase and not the amplitude.

[Kurt_!]

Is there already a market for such a device?
Especially because of Rigol's "super" service (firmware updates....).
Perhaps it would be suitable for industries where only two or three processes are tested?
In the East...

[Aiy]

Hi everyone,

On the images of the motherboard, just between the big BGA and MCU daughterboard there is an LED with the label "FPGA_DONE".
Is the big BGA with RIGOL branding just a re-branded FPGA ?

In the firmware update package there is an FPGA folder with .bin files, and the header looks like Xilinx if Gemini is to be believed..

Code: [Select]

padding 0xFFs ...
FF FF FF FF FF FF FF FF
00 00 00 BB 11 22 00 44
FF FF FF FF FF FF FF FF AA 99 55 66

Best regards,

[bobof]

On the images of the motherboard, just between the big BGA and MCU daughterboard there is an LED with the label "FPGA_DONE".
Is the big BGA with RIGOL branding just a re-branded FPGA ?

In the firmware update package there is an FPGA folder with .bin files...

Seems highly plausible.  FPGA_DONE and the power net names also in the area on the PCB are all Xilinx house standards.  The packaging name on the top of the device, also.  Seems to be too much of a coincidence.  I've worked on an audio product where the "processor" device bought under license was just a rebadged Xilinx Spartan 6 - so it's clear they do this.

[EEVblog]

On the images of the motherboard, just between the big BGA and MCU daughterboard there is an LED with the label "FPGA_DONE".
Is the big BGA with RIGOL branding just a re-branded FPGA ?

Likely. I speculated that in my video.

[bobof]

In the firmware update package there is an FPGA folder with .bin files, and the header looks like Xilinx if Gemini is to be believed..

Code: [Select]

padding 0xFFs ...
FF FF FF FF FF FF FF FF
00 00 00 BB 11 22 00 44
FF FF FF FF FF FF FF FF AA 99 55 66


Yeah, AA 99 55 66 is the Xilinx syncword sequence and has been since at least Spartan 6, per my understanding this pattern is used to work out how the configuration data bits are connected at the start of configuration.  That and all the other clues mean it really is almost certainly a rebadged Xilinx FPGA.