LO in dBm in mixer designs

[STOIKOV]

in mixers' designs I've read in books or papers that the LO is specified in dBm. This means dB relative to miliwatts at 50ohm load or dB relative to milivolts ? supposing the LO is a square wave.

 

[bcl]

Re: LO dBm

dBm is refered to dB relative to milliwatt.

 

[DDavid]

Re: LO dBm

Hi,

How much dBm it's power
square it's volt

 

[shwoo]

Re: LO dBm

dBm is power and we don't care about the impedance.
So, telephone engineers think 0dBm is equivalent to 0.775Vrms for
600 ohm systems, while RF engineers think 0dBm is equivalent to
0.225Vrms for 50 ohm systems.
You can easily convert dBm to power:
0dBm = 1mW
10dBm = 10mw
20dBm = 100mW
30dBm = 1W

2mW = 3dBm
4mW = 6dBm
20mW = 13dBm
I hope you understand my point.
S. H. Woo

 

[bcl]

Re: LO dBm

For sinusoidal signals in 50 ohm system, P(dBm)=10*log(Vrms*Vrms*1000/50)
where Vrms=Vpeak/√2

 

[STOIKOV]

Re: LO dBm

now it's clear for me that dBm is power referenced to 1mW with a load of 50Ω in RF systems. But now my doubt is in a square wave VCO I have that supplies a mixer. The LO port of this mixer is not 50Ω, is a higher impedance because is capacitive (gates of MOS). So in this case how I can specify the power of my VCO ?

 

[shwoo]

Re: LO dBm

I am wondering whether you are specifying LO power level to purchase
the VCO or you are designing/buidling one.

If LO is fed to a very high impedance port, you can add 50-ohm shunt
resistor to make sure input impedance of the mixer is controlled close
to a fixed value. Of course, the power of square wave is different from
sine wave. How clean is the square wave? It is difficult to make high frequency
square wave. How would you want to measure the power of the square wave?
Power meter? Spectrum Analyzer? RF voltage probe?
How do you want to determine optimum LO level?
If it is low frequency, say 10KHz, you may want to specify the peak to peak
voltage measured with a high impedance scope probe.
By the way, did you mention your frequency?
S. H. Woo

 

[biff44]

Re: LO dBm

Mixers do not work well with square waves as the LO signal. That is because a mixer is pretty much an analog sampler, and if you do not give it a sine wave at the LO port like it wants to see, the sampling function gets screwed up.

So your orignal question is similar to asking "what miles per gallon does my car get if I use corn syrup instead of gas in the tank"! The answer is simply, who knows if you can even get the car started, let alone a specific detail like miles/gallon.

 

[safwatonline]

Re: LO dBm

Mixers do not work well with square waves as the LO signal. That is because a mixer is pretty much an analog sampler, and if you do not give it a sine wave at the LO port like it wants to see, the sampling function gets screwed up.

So your orignal question is similar to asking "what miles per gallon does my car get if I use corn syrup instead of gas in the tank"! The answer is simply, who knows if you can even get the car started, let alone a specific detail like miles/gallon.

well , i am not sure about that i think square wave is ideal for switching as it gives maximum slope and hence more ideal switching , where near ideal switching gives at least better noise performance and better linearity

 

[bcl]

Re: LO dBm

The mixer ports usually like to see good impedance matching, 50 ohm or otherwise. As far as average power goes, you can refer back to the definition of average power;
Pavg=[1/(T*R)]*∫|f(t)|^2dt, integrated over period T, and divided by R.
The LO port, when driven with a sinusoid, would rectify the LO signal and, in the process, generate harmonics of the LO. If you start out with a squarewave as LO drive, since square wave is rich in harmonics, you may experience performance degradation in IMD and harmonic rejection of the mixer.

 

[biff44]

Re: LO dBm

If you understand samplers, you would realize that you are sampling a sinewave with a long pulsewidth (your squarewave), thereby canceling out most of the sine wave voltage. You need to sample with a spike, which is sort of what the top of the LO sine wave is.

 

[echo47]

LO dBm

A mixer is not a sampler.

A square wave make an excellent LO - it essentially toggles the polarity of the RF signal. This provides higher conversion gain, lower noise, and lower LO feedthrough than if you used a sinewave LO.

That's why high-level diode-ring mixers perform so well. The LO sinewave gets clipped into a square wave.

 

[safwatonline]

LO dBm

actually the sinsuidal LO doesn't sample, and the result of switching with a sinosoidal wave should be IFout=sgn(LO) * RF which means that u are multiplying with a square wave already, but giving the slow slope of the sinosoidal doesn't perfectly switch and hence it gives more noise and lower linearity

Added after 2 minutes:

and about the more harmonics , this is also not true as for perfect switching by a sinewave u get sgn(LO) which is already a square wave and hence have the same harmonics if u used a square wave from the first place

 

[bcl]

Re: LO dBm

Rectified sine wave has even harmonics, square wave that has rise and fall time has both even and odd harmonics.

 

[PaulHolland]

Re: LO dBm

biff44: Go back to school !.

Putting a square wave into a mixer is fine. All mixers are working in the saturated region of diodes or FET's so a square wave does only help you and not give difficulties. Normally people do not use square waves since its difficult to make a good square wave at high frequency.

Will work but reduce your power 2 dB and your fine !.

Paul.

 

[echo47]

LO dBm

Here's a really nice introduction to mixer theory and design:

 

[rosaeidi]

LO dBm

Hi,

h**p://
this link doesn't work.

 

[safwatonline]

LO dBm

the link works