RF power vs Conversion loss

[rakesh228]

in case of FET single resistive mixer, how the RF power and conversion losses related?

( why every one using LO power in positive dBm while RF power in negative dbm and vice versa....??? is there any reason to that?)

 

[jiripolivka]

In any mixer, diode or FET, the LO, local oscillator power controls the switching elements to on/off. Usually the LO power must be of the order of +10 dBm, typically, to do that.
The RF power (input or output) must be lower than LO, otherwise it would affect the switching function. In upconverters there is a rule that the RF output power , to keep the mixer linear (not saturated), is to be 7 dB lower than LO input. With +10 dBm LO input, the maximum undistorted RF output can be +3 dBm.
Conversion loss is the ratio of RF input to IF output (or reverse), mainly caused by switching resistance of mixer elements (FETs, diodes), and their matching to input and output ports. Good mixers have CL ~ 6 dB typical.

 

[rakesh228]

in case of down converter....whts the relation between RF power and LO power? if LO power is +10dBm then RF power must be in negatives i.e -5dbm??

 

[cnm]

In any mixer, diode or FET, the LO, local oscillator power controls the switching elements to on/off. Usually the LO power must be of the order of +10 dBm, typically, to do that.
The RF power (input or output) must be lower than LO, otherwise it would affect the switching function. In upconverters there is a rule that the RF output power , to keep the mixer linear (not saturated), is to be 7 dB lower than LO input. With +10 dBm LO input, the maximum undistorted RF output can be +3 dBm.
Conversion loss is the ratio of RF input to IF output (or reverse), mainly caused by switching resistance of mixer elements (FETs, diodes), and their matching to input and output ports. Good mixers have CL ~ 6 dB typical.

It probably depends on which kind of diodes/FET used (Si/InP/GaAs/GaN etc..)

 

[jiripolivka]

In any mixer, LO is like a heater with the vacuum tubes- it is the necessary condition to make a mixer operate.
In down converters (receivers) a mixer with a proper LO input can operate from a noise level input ( Pn = kTB) up to say LO power -10 dB. All the way down the mixer response is linear. If you exceed the RF input over LO-10 dB, any mixer "saturates", or its response is no more linear.
This is the reason there are several mixer "classes" according to LO input power. The lowest usually is +7 dBm, the highest +23 ...+27 dBm. This LO pumping level is determined by diode type (high or low knee on I/V curve).
There are also other mixer types: for instance, by introducing DC bias (the forward DC current brings diode junctions close to the I/V knee) such mixer only requires LO input of 0...+3 dBm . This saves money in mm-wave mixers, but such mixers are not linear with higher RF inputs.
You can have a single-diode mixer, a balanced mixer with a diode pair, a doubly or triply balanced mixer with two or three diode pairs. Each has some advantage in certain application. You can use also a FET or a transistor as a mixer (Gilbert cell has several transistor pairs), and others.

 

[tony_lth]

in case of down converter....whts the relation between RF power and LO power? if LO power is +10dBm then RF power must be in negatives i.e -5dbm??

No direct relation. Such as HMC520LC4, the LO power is 17dBm, the RF input P1dB is 12dBm.

 

[kellyza]

If let say, I have an up-converted single HBT mixer with DC bias at 2V, 10mA with the following settings and output:

LO port: 30 dBm, freq = 10 GHz
IF port: 10 dBm , freq = 2.4 GHz
I obtained about -3.8 dBm at RF output port, freq = 12.4 GHz and at the same output spectrum LO power appeared about -1.3 dBm.

Then I changed the power into these settings:
LO port: 25 dBm, freq = 10 GHz
IF port: 5 dBm , freq = 2.4 GHz
Surprisingly I obtained the RF power about -25 dBm while LO power appeared at about -4.5 dBm. I just reduce both powers with 5 dBm and why the output power dropped drastically.

Can anyone comment about the situation? What is the best input power I should use?

Thanks
-Kellyza-

 

[tony_lth]

For LO power 30dBm, the CL=13.8dB.
For LO power 25dBm, the CL=30dB.
That seems normal, because the LO power must be higher than some threshold, then the mixer can work. When LO power is higher than this threshold, the CL changes about a few dB. So you can test the threshold of your mixer. I am curious about what type mixer you are using, need 30dBm power? normally the LO levels are +13dBm, +17dBm, +23dBm, etc.

 

[kellyza]

Thanks tony_lth for your comment... actually I just try and error the value of the power... when I set the input power too low say negative dBm then the output power seems not too good that's why I try putting more higher values of the input power so that I could get the best possible output. It is ok for me to do that way or I should have more specific reason why I choose that value? pls advise me more.

Thanks
-Kellyza-

 

[tony_lth]

That's not a good idea. You should dowmload the datasheet of your mixer, do according to the manufactuer said. If you need more power to drive the mixer, there should be some problem, such as link cable, line,failure etc.

 

[kellyza]

That's not a good idea. You should dowmload the datasheet of your mixer, do according to the manufactuer said. If you need more power to drive the mixer, there should be some problem, such as link cable, line,failure etc.

For your info, I modeled and simulated the mixer using MWO platform. I'm not using available mixer but I designed and modeled it by my own. I used single BFP620 transistor for the mixer. I'm not sure my model is correct or not. May be you guys can give advise on how to design a mixer with optimum performance like its conversion loss, isolation or return loss may be. What are the conditions and circumstances that should be counted?

Thanks
-Kellyza-

---------- Post added at 07:49 ---------- Previous post was at 06:01 ----------

I obtained this graph when I plotted the conversion loss vs. LO power... I'm not sure it is wrong or right. Your comments and views are most welcome.



Thanks again
-Kellyza-