SA-LF-CON Spectrum Analyzer Low Frequency Converter

[TopQuark]

Came across this odd bit of test gear on Taobao, though it is also available on ebay (https://www.ebay.com/itm/283625193935).

It claims to extend the lower frequency limit of a spectrum analyser by mixing the signal under test with the 10M internal clock of a spectrum analyser, shifting the lower frequency signal under test by a 10M offset.

Thought this would be wonderful for looking at noise spectral density of circuits if it works as advertised, and from initial tests, it does work pretty well.

I generated a test signal at 5, 50, 500, 5k, 50k and 500kHz, at -50dbm from my SDG2042X and feed it into the SA-LF-CON. The output of SA-LF-CON is connected to my SSA3021X Plus which only has a lower frequency limit of 9kHz, a -10M frequency offset is applied for easy reading of the original signal. The 10M reference output of the SSA is connected to the 10M input port of the SA-LF-CON.

The results look really good, the amplitude of the signal is preserved, the 5Hz signal is detected quite accurately. I think I'll run more test and see what else I can do with this thing.

[mawyatt]

That is very interesting indeed!!

Do you know what type mixer they are using and any circuit details? Very curious about how this was implemented.

Looking forward to your test results!

Thanks for posting

Best,

[TopQuark]

Quite interesting indeed! I took apart the device to have a look under the hood.

The main workhorse is the SBL-1-1+ mixer from minicircuit, plus a bunch of external supporting devices.

Here's the signal flow from what I gather:

LF Input (BNC) -> 3000uF AC coupling -> 3dB attenuation -> SBL-1-1+ IF input

10M ref clock input (SMA) -> AC coupling, clamping -> SN74LVC2G04 Dual Inverter Gate -> BP filtering -> 3dB attenuation -> SBL-1-1+ LO input

SBL-1-1+ RF output -> 3dB attenuation -> SBB5089Z gain block -> 3dB attenuation -> RF output (SMA)

[TopQuark]

A couple points of interest here:

The SBB5089Z is rated for 50MHz-6000MHz, a bit too high for 10MHz. One non technical explanation for this odd selection is the SBB5089Z is available for really cheap and in large quantities in the Chinese 2nd hand / grey stock market, and quite a few RF products I bought from Taobao uses this gain block chip. A lot of gain block chips suffers from low frequency instability, that's probably why the chip was rated to 50MHz only. There's a 330 ohm resistor plus AC coupling cap feeding back the output to the input of the gain block, that's probably how they are able to tame the gain block chip and extend the operational frequency down to 10MHz.

The 10M ref input clock processing is quite interesting. I verified with a scope that the output of the filtered SN74LVC2G04 output is not clean, but they are using it as the LO anyways. I'll probe around and analyse the signals on board to see how the whole thing works.