Coupler or combiner?

[analog_chip]

Hi,
I want to combine RF and LO i.e. RF+LO and then input to my receiver which eventually squares and down-converts. Shall I used a combiner or a coupler? I have heard that combiner uses resistive or transformer technology and coupler uses capacitive coupling.

Also, to which category (discrete/HybridMIC) these components in this website belong? Do we just call them discrete RF/microwave components or something more specific?

**broken link removed**

 

[BigBoss]

Combiner has not be resistive, it may also be inductive ( transformer coupled) or capacitive.
If you use combiner, 2 signal will be combined under the same amount of attenuation at the combined output.If you use coupler, one signal will be attenuated by insertion loss but the other one will be coupled with a coupling coefficient say 10dB ,20dB etc.
It depends totally on your purpose..

 

[analog_chip]

Hi BigBoss,

Thank you very much for the reply! I want minimal power loss & minimal noise at the input since it is at the front of receiver.
I want to use a 180 deg hybrid. If I input RF and LO at the output ports (3,4) then I will get RF+LO at the sum port. Something as follows.
**broken link removed**
https://sigatek.com/Hybrids-90-180-...ds/SH12552-Hybrid-180-degree-2.0-4.0-Ghz.html

Whats your opinion on these Bigboss?

 

[BigBoss]

Hi BigBoss,

Thank you very much for the reply! I want minimal power loss & minimal noise at the input since it is at the front of receiver.
I want to use a 180 deg hybrid. If I input RF and LO at the output ports (3,4) then I will get RF+LO at the sum port. Something as follows.
**broken link removed**
https://sigatek.com/Hybrids-90-180-...ds/SH12552-Hybrid-180-degree-2.0-4.0-Ghz.html

Whats your opinion on these Bigboss?

Since your frequency sources are seperate, why would you pay more for nothing?? 3dB 180 degree Hybrid Couplers are for special applications to combine 2 signals or to divide these signal by 180 degree .That's why they are quite expensive and they will not be very proper choice for your combining purpose.
And insertion loss will be more than 3dB ( in practice ) and therefore they don't bring any advantage for you.If you design/buy a Wilkinson combiner/divider it will be more appropriate for you.You can design this combiner even by using lumped elements.
https://www.markimicrowave.com/2773/Power_Dividers.aspx?ShowTab=117

 

[analog_chip]

Hi Bigboss,
Thanks again! resistive dividers will increase the noise figure of the receiver. So I want inductive or capacitive combination. This is an experimental receiver where RF & LO is combined and then input to receiver. Are there combiners based on inductive/capacitive coupling?

 

[BigBoss]

Hi Bigboss,
Thanks again! resistive dividers will increase the noise figure of the receiver. So I want inductive or capacitive combination. This is an experimental receiver where RF & LO is combined and then input to receiver. Are there combiners based on inductive/capacitive coupling?

They are Wilkinson type (Microstrip Wideband) combiners.Look at their specifications..

 

[volker@muehlhaus]

And insertion loss will be more than 3dB ( in practice ) and therefore they don't bring any advantage for you.If you design/buy a Wilkinson combiner/divider it will be more appropriate for you.

Sure?

The Wilkinson is "lossless" only when used as a signal splitter, or as a combiner for in-phase signals. If you use it to combine out-of-phase signals from the "output" ports, the isolation resistors take effect and it is lossy. So for this application (combine different signals) it is not the best choice.

 

[RealAEL]

I want to combine RF and LO i.e. RF+LO

Just for clarification when you say RF+LO what exactly do you mean? Do you want to have a composite signal with two separate tones, one at Frequency RF and and the other at Frequency LO, symulatiously at the same output. Or do you mean you want a single IF tone at Frequency RF+LO (and/or RF-LO) at the output. Passive splitters, combiners or coupler will only do the former and present the two signals separately. You need an active mixer to get the two signals to truly 'combine' into a single IF output signal (and image).

 

[BigBoss]

Sure?

The Wilkinson is "lossless" only when used as a signal splitter, or as a combiner for in-phase signals. If you use it to combine out-of-phase signals from the "output" ports, the isolation resistors take effect and it is lossy. So for this application (combine different signals) it is not the best choice.

No 2 signal generator that work at different frequencies produce "in-phase" signals so if we talk about the real world, Wilkinson combiner/divider is lossy element.
I know what you mean, odd mode, even mode etc. They are math, but this is true..
Is there any more suitable combiner to combine two different frequency signals ?? Yes, there are but they are more complicated ( Butler,Hybrid, etc.)

 

[volker@muehlhaus]

No 2 signal generator that work at different frequencies produce "in-phase" signals so if we talk about the real world, Wilkinson combiner/divider is lossy element.

An example for in-phase signals would be power amplifiers where you split the signal into 2 paths (for 2 transistors) and then recombine.

But maybe you misunderstood my point? My point was that the wilkinson coupler used as a combiner is worse than the branchline coupler, because it has extra loss from the isolation resistor when used as a combiner for out-of-phase signals.

But I think RealAEL has raised a good pint: is there a fundamental misunderstanding in the initial question about "combining" signals? Does the OP search for a combiner or a mixer?

 

[shashy.br]

U want to Downconvert or upconvert?
RF+LO---> in order to add the frequencies u need a non linear device like Mixers!!!
How can u give input to Combiner and Expect a stable power level waveform at RF+LO?? the result will be
product radiative interference hence power level cannot be estimated properly!!!

Am i wrong??? :|

 

[BigBoss]

I guess that he wants just to combine 2 signals and the they will be driven to a diode mixer or something else to get Mixing products.
Therefore these signals are independent and not coherent.

 

[analog_chip]

Hi everyone,
Thank you all for the comments. I want to combine two tones at RF and LO and get a composite signal at wRF and wLO. I dont want any downconversion. The downconversion will be done by my receiver in CMOS which is somewhat like diode based as BigBoss said.

The power levels are extremely low e.g. -15dBm LO power.

I want to find the noise figure of this receiver and de-embed any noise due to summation at the input. The Wilkinson has T-line and resistors. I am worried about the resistors which will contribute to NF.

Also the input resistance to my designed receiver is 50 ohm. So I was looking for a solution like transformer or capacitor based combiner.

 

[ge]

Analog_chip,
For your application most combiners will have ideally 3dB loss. Reality will be higher (3.5 - 4.5dB). What frequencies is your RF & LO? You are right to avoid resistive. Wilkinson is the low cost / wideband method.

 

[volker@muehlhaus]

The power levels are extremely low e.g. -15dBm LO power.
(...)
I want to find the noise figure of this receiver and de-embed any noise due to summation at the input. The Wilkinson has T-line and resistors. I am worried about the resistors which will contribute to NF.

Usually, there is a low noise amplifier before the filter/mixer stages, with enough gain so that the noise contribution of the next stages is small.

But if your design is different, and you need to keep the RF path attenuation small, and you only need -15dBm of LO power for the combined signal, how about this: use something like a 20dB coupler that has very little insertion loss in the main path, and inject the LO with +5dBm at the coupled branch.

 

[biff44]

a microwave mixer is specifically designed to do this. But has three ports, the RF, LO and IF output port. You feed the LO power into the LO port, and the weak RF signal into the RF port and get the IF out and a weak LO +RF out the if port too.

Are you sending the LO and RF to the receiver remotely, so they both are very low level? I suppose you could shove the RF PLUS LO into a nonlinear device....something with an I = K V² and get a mixing product of RF-LO frequency, but it will be VERY poor conversion loss.

I did a receiver once where I received an RF signal, amplified, doubled, and retransmitted it. It had some real issues when there was any drop in input RF power...making it a lab curiosity but not really ready for prime time consumer product. Sounds like you might have the same threshold effect, where any drop in input RF power will send you off a cliff.

 

[analog_chip]

Hi ge,
My LO is at 2.4GHz & RF at 2.42GHz.

Hi volker,
Your idea is OK. But when there is as lossy stage in front of RX chain then noise figure increases due to two reasons. First, the loss amplifies the noise figure of next stage instead of suppressing. Second, its own thermal noise adds to noise figure if the loss is due to resistive/dissipative mechanism causing electron perturbation. I want to reduce the second. I know how to de-embed both but still I want to know options available in the market .

Hi biff44,
"Are you sending the LO and RF to the receiver remotely, so they both are very low level? I suppose you could shove the RF PLUS LO into a nonlinear device....something with an I = K V² and get a mixing product of RF-LO frequency, but it will be VERY poor conversion loss. "
Low level is used to ensure low power consumption. Square law mixers have good conversion gain. Why do you think it has poor gain?

 

[shashy.br]

If u dont want to use a coupler then other option is to use a Arbitrary Waveform Generator for generating Multi tone signal at the frequencies that u want!!!

Please look at the link below as an example...!!!

http://www.home.agilent.com/upload/...veform_EuMW2011.pdf?cmpid=MD1583&cc=IN&lc=eng

Also,

If are concerned only with NF while using Coupler method then why not look at the Datasheet and adjust the measured NF accordingly with the insertion Loss in Datasheet????

 

[volker@muehlhaus]

Your idea is OK. But when there is as lossy stage in front of RX chain then noise figure increases due to two reasons. First, the loss amplifies the noise figure of next stage instead of suppressing. Second, its own thermal noise adds to noise figure if the loss is due to resistive/dissipative mechanism causing electron perturbation.

Using a divider/combiner with 3dB loss in front of the RX will increase the noise figure by 3dB, no matter if there are resitive elements included. That's why I suggested to reduce the insertion loss in the RX, by using an asymmetric coupler. For the LO with its much higher power levels, we can afford higher insertion loss.

 

[Raghav Kapur]

You can search for Hybrid Couplers on everything RF. Couplers from most of the manufacturers are listed here. Use the parametric search tool to narrow down the results to match your requirement.

180 Degree Couplers - https://www.everythingrf.com/search/one-eighty-degree-hybrid-couplers
90 Degree Couplers - https://www.everythingrf.com/search/ninety-degree-hybrid-couplers