[SOLVED] Mixer's output power

[LandLack]

I am putting a 10.35 dBm signal into a mixer; the datasheet tells me that the Conversion Loss of the device is 6.7dB for a Single-Side-Band (SSB). I have two questions:

1) The total power output of the mixer will be 6.6dBm, right? I ask this because I think that the output of this mixer is Double-Side-Band (DSB) by default, so the output power will be in dBm: Pin-6.7+3 = Pout . (tell me if it's wrong, please)

2)How may I use a SSB instead of DSB? Do I need to use a filter on the mixer's output (If yes, can you tell me how can I filter the carrier power of the unused sideband, which has the biggest contribution)?

If you have an answer to any of these question, please reply, even if you aren't sure. Thank you,

Best Regards,
LandLack

 

[chuckey]

3, If you only want one sideband then you have to filter the output of the mixer. You say " how can I filter the carrier power of the unused sideband," there will be no carrier power due to the unwanted side band, only sideband power, no signal input - no power. The actual carrier the DSB is generated at is not present as a proper mixer, nulls it out ( balanced mixer) and a double balanced mixer nulls out the input frequency.
There are three ways of generating SSB, filter, phasing and the "third method", for a poor article see :- https://en.wikipedia.org/wiki/Single-sideband_modulation
Frank

 

[LandLack]

3, If you only want one sideband then you have to filter the output of the mixer. You say " how can I filter the carrier power of the unused sideband," there will be no carrier power due to the unwanted side band, only sideband power, no signal input - no power. The actual carrier the DSB is generated at is not present as a proper mixer, nulls it out ( balanced mixer) and a double balanced mixer nulls out the input frequency.
There are three ways of generating SSB, filter, phasing and the "third method", for a poor article see :- https://en.wikipedia.org/wiki/Single-sideband_modulation
Frank

Sorry for not underlining, but I'm talking about upconversion, from a baseband to a RF signal. I've been using minicircuits mixers, so the C.L. was referred to a SSB (and downconversion) situation; I've read that if both sidebands are present, the CL should be different, and this made me post this thread. I'll try to ask them directly...

 

[jiripolivka]

As you use the mixer for SIGNAL conversion, then the conversion loss applies without any relation to SSB or DSB mixer use.
There is an important difference in SSB or DSB mixer operation but only related to NOISE. In communication receivers, RF filter is used to only pass the desired signal and reject the unwanted sideband (possibly rejecting unwanted signals there).

Mixers generate many harmonic combinations of LO and RF (or IF) input signals. We use filters to select what we want and reject what we do not want. Mainly in upconverters, a RF band-pass filter must be used to define mixer conversion loss, otherwise it is undefined. Adjusting the phase shift between mixer RF port and filter input port often adjusts the harmonics and helps to reduce mixer conversion loss. Instead of 6-7 dB, a good upconverter can have 3-5 dB CL.

- - - Updated - - -

With +10 dBm mixer RF input, you also should choose a correct LO level (and also a "hard" mixer), which for a linear operation should be ~7 dB higher than the RF. Minicircuits offer mixers with +17 dBm LO input which is good for your case. Otherwise the mixer is overdriven and will be limiting due to saturation.

 

[LandLack]

As you use the mixer for SIGNAL conversion, then the conversion loss applies without any relation to SSB or DSB mixer use.
There is an important difference in SSB or DSB mixer operation but only related to NOISE. In communication receivers, RF filter is used to only pass the desired signal and reject the unwanted sideband (possibly rejecting unwanted signals there).

Mixers generate many harmonic combinations of LO and RF (or IF) input signals. We use filters to select what we want and reject what we do not want. Mainly in upconverters, a RF band-pass filter must be used to define mixer conversion loss, otherwise it is undefined. Adjusting the phase shift between mixer RF port and filter input port often adjusts the harmonics and helps to reduce mixer conversion loss. Instead of 6-7 dB, a good upconverter can have 3-5 dB CL.

- - - Updated - - -

With +10 dBm mixer RF input, you also should choose a correct LO level (and also a "hard" mixer), which for a linear operation should be ~7 dB higher than the RF. Minicircuits offer mixers with +17 dBm LO input which is good for your case. Otherwise the mixer is overdriven and will be limiting due to saturation.

Thank you for your reply; I need to use Surface Mount Technology, so most LO have an output power of 8to 14dBm, and most mixers are level 10 or something around. Still, the parameters in the datasheet seem ok: they said me that CL=6.7dB is referred to SSB; if I had DSB, it will be 3dB lower, which in practice means CL=3.7dB. Now I would like to ask you other two questions:
1)If the carrier is at F°=130MHz, then I'll have two sideband around this freq; but the other modulation results are around 2F°, 3F° etc., right? If yes, can you tell me an estimate of their amplitude compared to the original around F°?

2)I've used a simulator to see the frequency spectrum a IF baseband signal with 6.8dBm (DC)+ 3.54 dBm(the total power in the harmonics of the actual signal) to be upconverted (CL=3.7dB) on F°=130 MHz; can you tell me why the output power has a carrier of 9.8dBm? Or in general the total power balance of the upconverted signal (which seems to be bigger than the baseband).

I've attached the image...

 

[jiripolivka]

To your questions: 1. DSB is only related to noise, so with a signal use CL ´6.7 dB, not less. With a carrier at 130 MHz, you will have both sidebands wih all harmonics on both sides unless you use a filter to select one you need. Amplitudes of the sidebands depend on the diodes used, LO power, etc. This is why mixers mostly use the first side bands and reject others. Real amplitudes are measured, I do not trust simulations.

2. What do you mean by "IF baseband 6.8 dBm DC + 3.54 dBm"? If you apply DC to a mixer, you will rather burn the diodes, or at least cause a heavy imbalance. The AC power to IF port say 3 dBm with +10 dBm LO will generate a RF output of -3 dBm in my opinion. I do not trust imulations, I know good mixers from years of experience.

 

[LandLack]

To your questions: 1. DSB is only related to noise, so with a signal use CL ´6.7 dB, not less. With a carrier at 130 MHz, you will have both sidebands wih all harmonics on both sides unless you use a filter to select one you need. Amplitudes of the sidebands depend on the diodes used, LO power, etc. This is why mixers mostly use the first side bands and reject others. Real amplitudes are measured, I do not trust simulations.

2. What do you mean by "IF baseband 6.8 dBm DC + 3.54 dBm"? If you apply DC to a mixer, you will rather burn the diodes, or at least cause a heavy imbalance. The AC power to IF port say 3 dBm with +10 dBm LO will generate a RF output of -3 dBm in my opinion. I do not trust imulations, I know good mixers from years of experience.

Thank you again for your reply, you are helping me a lot;

1)the reason I said that CL was 3dB lower was because of this paper (see the section "CL") : **broken link removed**
Let me edit that part:

"Conversion loss is a measure of the efficiency of the mixer in providing frequency
translation from the RF input signal to the IF output signal. For given RF and LO
frequencies, two nominally equal-amplitude output signals are produced at the sum and
the difference of the RF and LO frequencies. Since only one of these products (or
sidebands) is utilized in most applications, the specifications given in Mini-Circuits data
sheets are for a single-sideband output. If both sidebands are utilized, the conversion
loss is 3 dB lower than in the single-sideband case.
Conversion loss of a mixer is equal to the ratio of the RF input power to the IF single-
sideband output power, expressed as a positive number in dB. All measurements are based(...)"

2)The signal is perfectly rectangular shaped, with 0V when "down" and 0.738V when "up". So the simulator "saw" that as a sum of a DC and a rect function; this is probably the reason why the baseband spectrum has a DC amplitude with 6.8dBm, while the remaining amount of power is the magnitude of a sinc function. I'll put it as another image:

 

[jiripolivka]

Thank you again for your reply, you are helping me a lot;

1)the reason I said that CL was 3dB lower was because of this paper (see the section "CL") : **broken link removed**
Let me edit that part:


2)The signal is perfectly rectangular shaped, with 0V when "down" and 0.738V when "up". So the simulator "saw" that as a sum of a DC and a rect function; this is probably the reason why the baseband spectrum has a DC amplitude with 6.8dBm, while the remaining amount of power is the magnitude of a sinc function. I'll put it as another image:

I cannot comment on the assertions of CL reduction by 3 dB as I never saw such a thing. In a down converting mixer, only one sideband is used and a filter rejects the other. The reason is to convert a desired signal which is present in one sideband while interferers can be present in the image band. In down converters, the CL is typically 6...8 dB and cannot be reduced by any way.

In upconverters, the IF signal generates two RF sidebands but again only one is desired. Usng a RF filter, one sideband is selected and the other rejected. By adjusting the phase shift between the mixer RF output port an RF filter input port, CL can be reduced by 2..3 dB. The phasing improves power transfer into the desired RF sideband. MIxer linearity becomes important in high-level mixers, and the linearity rule is to use LO input power minus 7 dB for the maximum IF input before the mixer becomes saturated.(and CL grows by 1, later by more dB).