Microwave SSB Modulator

[jazzalex]

Hi all,

I am looking for a single-sideband (SSB) - modulator/demodulator in the microwave domain (X-Band: 7 to 9 GHz or Ku-Band: 9 GHz to 13 GHz). Does anyone know where to search for it ? Any hint appreciated !

Thanks

Alex

[gregariz]

probably your best bet would be a transverter

http://www.downeastmicrowave.com/

[vk6zgo]

probably your best bet would be a transverter

http://www.downeastmicrowave.com/

I would agree, making filters sharp enough to  produce SSB at microwave frequencies is virtually impossible.
Up-converting from a more useable frequency is normal practice.

Another method was used for a time,where a lower frequency SSB signal was envelope demodulated,the associated RF signal limited to remove any amplitude variations,then multiplied to the required frequency &  again modulated by the  envelope to give an SSB signal at the higher frequency.

This is a poor description,but it apparently worked OK.
As soon as stable upconverters were developed,this method was mostly dropped.

VK6ZGO

[jazzalex]

ok - thanks for the info ! Actually, I need this for a lab experiment, in which I am stuck at carrier frequencies between 9 and 10 GHz. However, my signal is very narrow in terms of bandwidth (a sine of 5 MHz or more). In that context I understand that what I have to do is basically getting rid of the carrier and the upper sideband and add the carrier afterwards. However, then -- regarding the demodulation -- I ask myself: Is a detector diode able to work with only one sideband and the carrier or does it require both sidebands ?

Thanks

Alex

[vk6zgo]

ok - thanks for the info ! Actually, I need this for a lab experiment, in which I am stuck at carrier frequencies between 9 and 10 GHz. However, my signal is very narrow in terms of bandwidth (a sine of 5 MHz or more). In that context I understand that what I have to do is basically getting rid of the carrier and the upper sideband and add the carrier afterwards. However, then -- regarding the demodulation -- I ask myself: Is a detector diode able to work with only one sideband and the carrier or does it require both sidebands ?

Thanks

Alex

I'm a tad confused,do you mean you have an original baseband signal at 5MHz,with a fairly low bandwidth,which you need to modulate
onto a microwave carrier,or is your baseband signal 5MHz wide?

 A few possibilities  for modulation types:
(a)If you  get rid of the other sideband & suppress your carrier by 60dB or better,you have the sort of SSB signal,commonly used by Hams,etc.
With this sort of signal,you can use a detector diode,but you have to supply a "replacement"carrier,which in the normal case is not phase locked to the original carrier.(You may as well use a balanced demod,as it is exactly the same as the modulator youwill have to use at the  Transmitter end).

(b)If you use a similar signal to (a),but only suppress the carrier by around 26dB ,you can use the very low level carrier to phase lock a carrier generator at the Rx end,so you will have the same stability of demodulation you would have with full carrier AM.

(c)Another way is get rid of the other sideband,but send full carrier---this is called  Compatible SSB,& can be received with a normal AM
(single diode) detector.

(d)Yet another method is Vestigial Sideband (VSB),where you don't suppress all of the other sideband,& run full carrier--this is similar to (c),but easier to achieve.The downside is problems with Quadrature Distortion.

VK6ZGO

[jazzalex]

I'm a tad confused,do you mean you have an original baseband signal at 5MHz,with a fairly low bandwidth,which you need to modulate
onto a microwave carrier,or is your baseband signal 5MHz wide?

 A few possibilities  for modulation types:
(a)If you  get rid of the other sideband & suppress your carrier by 60dB or better,you have the sort of SSB signal,commonly used by Hams,etc.
With this sort of signal,you can use a detector diode,but you have to supply a "replacement"carrier,which in the normal case is not phase locked to the original carrier.(You may as well use a balanced demod,as it is exactly the same as the modulator youwill have to use at the  Transmitter end).

(b)If you use a similar signal to (a),but only suppress the carrier by around 26dB ,you can use the very low level carrier to phase lock a carrier generator at the Rx end,so you will have the same stability of demodulation you would have with full carrier AM.

(c)Another way is get rid of the other sideband,but send full carrier---this is called  Compatible SSB,& can be received with a normal AM
(single diode) detector.

(d)Yet another method is Vestigial Sideband (VSB),where you don't suppress all of the other sideband,& run full carrier--this is similar to (c),but easier to achieve.The downside is problems with Quadrature Distortion.

VK6ZGO

It's the further: Original baseband signal is a sine of 5 MHz (or multiples of 5 MHz --> 10, 15, 20 MHz up to 100 MHz) and the carrier has a frequency of approx. 9.5 GHz.

It looks like a) or c) makes most sense to me. In that context one more general question:

If one single sideband+carrier is sufficient for the demodulation with a single diode what impact does then the second sideband have in conventional DSB-Demodulation ? What happens if e.g. due to transmission errors the upper sideband differs from the lower sideband ?

Thanks for your help,
best

Alex

[vk6zgo]

Hi Alex,

If ,say,the USB has errors that aren't the same as that of the LSB,the two sidebands will not be exactly the same,&  the resulting demodulated signal will be adversely affected.

An example of this is "fading",in HF radio,where not only are the propagation time & losses slightly different between sidebands,but the difference varies over time,resulting in cancellation at one moment,& reinforcement the next,as the path distance, & hence,phase varies,the degree of cancellation or enhancement depending on the relative path losses.

The useful demodulated output from a detector for an AM (DSB,full carrier) signal is:

F(demod)=(f2-fc)+(fc-f1) where fc is the carrier,f1 is the LSB,& f2 is the USB.

In a perfect system (f2-fc) & (fc-f1) are identical & in phase,so add directly.

In a real system,they may not meet this criteria,& some of the effects quoted above will be noted.

In Compatible Single Sideband (CSSB):
 
F(demod)=(f2-fc),or F(demod)=(fc-f1),depending on which sideband you suppress.

Re: Generating SSB--if your lowest frequency is 5MHz,you shouldn't have a lot of trouble with filters if you use a lower frequency to generate the signals,then up-convert.

Are you really locked into  AM/SSB for this?
Commercial practice is pretty much to use FM for microwave systems so there may be a lot more information out there for that form of modulation.

VK6ZGO

[jazzalex]

Thanks for the comprehensive explanation - now the picture is complete !

Are you really locked into  AM/SSB for this?
Commercial practice is pretty much to use FM for microwave systems so there may be a lot more information out there for that form of modulation.

Yes, I know and I'd rather prefer using different techniques. However, my lab experiment exhibits an unconventional transmission behavior, which makes me believe that the currently applied DSB is not optimal and that SSB will improve the situation. Why and how is something I cannot explain briefly. I will sketch a graph and provide measurement results here soon and I am very curious if you share my point of view or even have a better suggestion.

Talk soon, best

Alex

[jazzalex]

Yes, I know and I'd rather prefer using different techniques. However, my lab experiment exhibits an unconventional transmission behavior, which makes me believe that the currently applied DSB is not optimal and that SSB will improve the situation. Why and how is something I cannot explain briefly. I will sketch a graph and provide measurement results here soon and I am very curious if you share my point of view or even have a better suggestion.

As proposed I sketched a drawing, which explains the reason for considering SSB instead of DSB: My experimental setup exhibits small "gaps" of lowest group delay (which appear periodically in 5 MHz intervals) and this is precisely where I want the signal to be transmitted. This works ok but not as accurate as it should be (see first two graphs with DSB). Hance, my idea is using SSB instead or shifting the sidebands to the next gap -- as the remaining graphs illustrate.

Any comment appreciated.

Thanks

Alex