# What is the functional difference between DSB and 2SB mixers, and why exactly would 2SB be better for radio astronomy?

The Astronomy SE question 2SB mixer in radio astronomy? asked in 2014 in its entirety:

In literature about the instrumentation of radio astronomy, I frequently come across a type of mixer described as "2SB". I'm familiar with single-sideband (SSB) and double-sideband (DSB) mixers, but I have never studied a 2SB mixer.

What is a 2SB mixer and how does it compare to a DSB mixer? Ideally, I would also like to know of formal references that directly address the topic for further reading.

2SB is Dual Sideband, as opposed to DSB - Double Sideband. Here are a couple of papers you might find relevant and interesting:

leaves me still not really understanding the difference between 2SB and DSB mixers nor exactly when one would choose one over the other.

My impression is that DSB mixers suppress the carrier while allowing both sidebands to pass through a single output, while 2SB mixers further separate each sideband to a separate output.

Question(s):

1. Is the separation of the two SBs to two outputs the functional difference between 2SB and DSB? If not, what would it be?
2. The abstract of the second linked paper says that 2SB mixers' "main advantages being to avoid spectral confusion and to diminish effective system temperature by a factor two with respect to double sideband (DSB) receivers." Is it possible explain in fairly simple terms what that is so?

1Hesper et al. A Sideband-Separating Mixer Upgrade for ALMA Band 9 from the 20th International Symposium on Space Terahertz Technology, Charlottesville, 20-22 April 2009 2A calibrated digital sideband separating spectrometer for radio astronomy applications

2Finger et al. A Calibrated Digital Sideband Separating Spectrometer for Radio Astronomy Applications also available at Pub. Ast. Soc. Pac.

• I'm not sure the demodulation tag is a good fit, but there are no tags for mixing, mixers, or sideband, ssb, dsb...
– uhoh
Aug 16 '20 at 23:28

Is the separation of the two SBs to two outputs the functional difference between 2SB and DSB? If not, what would it be?

Almost certainly yes, although it seems that the only hits that "2SB mixer" gets is for astronomy sites. See the figure, from the first paper you cite. It clearly shows two sidebands being received independently.

Basically, you're getting two SSB channels out of this, which, presumably, can be selected downstream.

The abstract of the second linked paper says that 2SB mixers' "main advantages being to avoid spectral confusion and to diminish effective system temperature by a factor two with respect to double sideband (DSB) receivers." Is it possible explain in fairly simple terms what that is so?

Spectral confusion: DSB mixers take the sideband above the carrier and the sideband below the carrier and mix them equally to baseband. The output of a DSB mixer is the sum of the downconverted USB and the downconverted LSB. For uncorrelated signals, the power in the USB and the power in the LSB are added.

Noise temperature: Noise temperature is a function of the ratio of noise power to desired signal power. Because the DSB mixer output power is the sum of the LSB and USB power, it's output has the noise contribution from both of these sidebands. Because (based on the context of the article) the desired signal is in only one sideband, you're getting the noise power from two sidebands, and signal from one.

With a SSB mixer (remember you get two SSB channels out of the 2SB mixer) you get the noise power from one sideband, and the signal from one sideband. So the power/noise ratio is doubled -- this results in the noise temperature being halved.

• Thanks for the speedy answer, this is great! I'll review it further later today.
– uhoh
Aug 17 '20 at 2:17
• If it answers your question, mark it as such. Aug 17 '20 at 21:03
• yep yep, fear not! I've accepted several thousand SE answers over the years. I usually let a question sit for several days to allow for discussion, voting and for the possibility of additional answers to be posted. Since this answer is insightful and I'm not an expert, I'd like the chance to read it carefully and think about it as well.
– uhoh
Aug 18 '20 at 0:57
• Oh I think I understand "Noise temperature is a function of the ratio of noise power to desired signal power." to mean that (based on the article) one would have cool the front end further for DSB than for SSB if the useful signal were only in one sideband. $$T_N = \frac{P_N}{k_B \Delta f}$$ and two sidebands will have twice the $\Delta f$ of a single sideband. That leaves me wondering why not simply use SSB instead of one output from the fancier 2SB mixer, but that may be beyond the scope of this question.
– uhoh
Aug 20 '20 at 1:44