Measuring Splitter/combiners

[prcguy]

Can we first agree that if you add 1watt to 1watt we get 2 watts? Theoretical.

You have to use a device to combine them and that will always have a minimum of 3dB loss, actually 3.01029996dB theoretical plus other losses. If you combine two 1 watt signals in phase using a combiner that has 3dB loss you will have very close to 2 watts at the common port. So yes I agree it can happen.

If you combine the same two 1 watt signals with the same combiner and its out of phase due to two different cable lengths you can have nearly zero watts so its all dependent on the phase. But the combiner still has 3dB insertion loss no matter if the signal is passing from the common port to an output or an output to the common port. Explain then how combing two signals in phase can give 3dB gain when the device combining them has 3dB loss? The answer is still you are starting with a 6dB increase by combining two in phase signals but loosing 3dB of it in the hardware combiner leaving you with 3dB net gain. Do you have a different explanation?

I should add that you can never see or realize the 6dB gain from combining two in phase signals because there will always be lossy hardware involved in the combining process. But the 6dB gain is there in theory to explain the resulting potential 3dB gain from combining two in phase signals. Hope that helps.

 

[MavamQ]

Do you understand that you have 3db loss because you have only 1/2 the power at each output of a splitter, and that half the power is a 3db loss? You only have the 3db loss because it is impossible to split the signal and still have the same energy at two outputs as you do at the input.

 

[prcguy]

Do you understand that you have 3db loss because you have only 1/2 the power at each output of a splitter, and that half the power is a 3db loss? You only have the 3db loss because it is impossible to split the signal and still have the same energy at two outputs as you do at the input.

But that same splitter used backwards as a combiner has the same 3dB loss when combining. How do you then get 3dB gain when combining two equal in phase signals? You would end up with the same power you applied at each input which would be two signals at 1 watt each ending up to be .5w each after the 3dB loss resulting in 1 watt total for two signals combined through a device that has 3dB loss. The theoretical 6dB gain from combining explains it.

 

[prcguy]

I might add that I asked the same questions the OP is asking of me in the early 80s while in a microwave components class where we made and tested things like stripline combiner/dividers and directional couplers. The instructor was a PHD and I asked how you get 3dB gain when the combining device has 3dB loss and he explained the 6dB combiner theory leaving you with 3dB gain. That's what I'm trying to explain in these posts.

 

[MavamQ]

You have to use a device to combine them and that will always have a minimum of 3dB loss, actually 3.01029996dB theoretical plus other losses. If you combine two 1 watt signals in phase using a combiner that has 3dB loss you will have very close to 2 watts at the common port. So yes I agree it can happen.

If you combine the same two 1 watt signals with the same combiner and its out of phase

we have settled that, the signals are in phase.

due to two different cable lengths you can have nearly zero watts so its all dependent on the phase. But the combiner still has 3dB insertion loss no matter if the signal is passing from the common port to an output or an output to the common port. Explain then how combing two signals in phase can give 3dB gain when the device combining them has 3dB loss?

you are assuming a combiner has 3db of loss, there are combiners with 0.2db loss and a resistive combiner has 6db loss.

I might add that I asked the same questions the OP is asking of me in the early 80s while in a microwave components class where we made and tested things like stripline combiner/dividers and directional couplers. The instructor was a PHD and I asked how you get 3dB gain when the combining device has 3dB loss and he explained the 6dB combiner theory leaving you with 3dB gain. That's what I'm trying to explain in these posts.

Maybe this will help, but you need to understand what the difference between dbm and db. The important line is, "
If the inputs are -10 dBm and -10 dBm, the output is -7 dBm. In this case, since the input levels are the same, the output power is doubled or increased by 3 dB. The result is therefore: -10 dBm + 3 dB = -7 dBm.

In both these cases, we have assumed no additional insertion loss. That’s not practical. Typically all combiners have an insertion loss specified in the data sheet that increases with frequency."
But you have to get over the idea that a combiner has 3db loss. A splitter has 3db loss because you are splitting the signal.

 

[KevinC]

I may not always agree with what he posts, but if you're arguing with @prcguy about combining/splitting and losses/gains you're not going to win.

And I'd bet all the money in my 401K he knows the difference between dBm and dB...but I could be wrong.

 

[prcguy]

we have settled that, the signals are in phase.

you are assuming a combiner has 3db of loss, there are combiners with 0.2db loss and a resistive combiner has 6db loss.


Maybe this will help, but you need to understand what the difference between dbm and db. The important line is, "
If the inputs are -10 dBm and -10 dBm, the output is -7 dBm. In this case, since the input levels are the same, the output power is doubled or increased by 3 dB. The result is therefore: -10 dBm + 3 dB = -7 dBm.

In both these cases, we have assumed no additional insertion loss. That’s not practical. Typically all combiners have an insertion loss specified in the data sheet that increases with frequency."
But you have to get over the idea that a combiner has 3db loss. A splitter has 3db loss because you are splitting the signal.

Its not my idea that a combiner has 3dB loss, its physics. Show me a 2-way RF combiner that has less than 3dB loss, make and model # please. I take that back, there are optical combiners like a reflector where two horn antennas point at it and when phase right you can have a fraction of a dB combining loss. They do exist. But the resulting gain of that is 6dB and there are losses going to an antenna and air which negate some of the gain. Its the same as multipath where a signal radiated from an antenna reflects off something and you have a direct and reflected path. If they signals are perfectly in phase you can have up to 6dB improvement at your receiving antenna and when they are 180 deg out of phase you can have 40dB or more attenuation. But for coaxial connector stripline hardware combiners its 3dB loss.

I have that $1000 cash waiting for the magical 2-way RF combiner that has less than 3dB loss. But if you pick a make and model and I prove it has at least 3dB loss, how much do I get? I'll be waiting for that make and model.

 

[MavamQ]

I may not always agree with what he posts, but if you're arguing with @prcguy about combining/splitting and losses/gains you're not going to win.

And I'd bet all the money in my 401K he knows the difference between dBm and dB...but I could be wrong.

Well then he can see that adding up signal levels and see -10dbm + -10dbm =-7dbm, and converted to db that is a 3db gain.

 

[prcguy]

I may not always agree with what he posts, but if you're arguing with @prcguy about combining/splitting and losses/gains you're not going to win.

And I'd bet all the money in my 401K he knows the difference between dBm and dB...but I could be wrong.

What's a dB? How much you got in that 401K? Are you willing to loose 60dB of it measured in dB$?

 

[prcguy]

Well then he can see that adding up signal levels and see -10dbm + -10dbm =-7dbm, and converted to db that is a 3db gain.

 

[ramal121]

dB$, boy I'm going to start using that.

 

[MavamQ]

Here is a decent paper on Combiner loses (CL). His overall point is "I personally came to believe that combining loss (CL) is the overall figure of merit that should be used (others may disagree this is why I put the idea here for others to pick at it)." rather than return loss isolation, etc. At the bottom right is the very low loss of his combiner looks about 0.1db, even at the band edges it is only up to 1.5db.

Combiner Loss Figure of Merit

Combiner Loss Figure of Merit

www.microwaves101.com

 

[prcguy]

No need to convert, its already in a dB format. Take -10dBm subtract 3dB for combiner loss that gives you -13dBm. Two signals of -13dBm combined at the output would then equal -10dBm for your way of thinking, which leaves a missing 3dB that needs to be explained. And until you post a make and model of an RF combiner that has less than 3dB of loss you have no argument that there is 3dB missing that must be explained, which I have and you don't believe.

 

[prcguy]

Here is a decent paper on Combiner loses (CL). His overall point is "I personally came to believe that combining loss (CL) is the overall figure of merit that should be used (others may disagree this is why I put the idea here for others to pick at it)." rather than return loss isolation, etc. At the bottom right is the very low loss of his combiner looks about 0.1db, even at the band edges it is only up to 1.5db.

Combiner Loss Figure of Merit

Combiner Loss Figure of Merit

www.microwaves101.com

That paper is discussing losses on top of theoretical 3dB and it happens I've built similar combiners in class and have used many them in the same field as the author is targeting. I was a principal engineer in the satelltie broadcast industry and designed uplink and downlink chains that used tons of splitters and combiners.

Anyway the key point to me in the paper is the first graph of four right below the green color representation of the Gysel power combiner. The chart is an S21 plot of insertion loss of the combiner which shows its got close to 3dB loss in center of band and about 3.5dB loss at band edges. Not bad. That's what the paper is about, improving performance of these combiners, which I have used many of.

You still need to find a two way combiner with specs that says overall insertion loss is less than 3dB and not the leftover resistive type loss on top of the theoretical 3dB. I'll give you some time but knowing its not possible we have to wrap this up sometime soon.

There are probably others watching this thread and in the end I want to make sure those watching see the correct answer to the questions being argued. Hope that comes soon.

 

[MavamQ]

dB$, boy I'm going to start using that.

Where did you see db$ ?

 

[KevinC]

How much you got in that 401K?

$14

 

[MavamQ]

prc guy, if you add two 1 watt signals do you get 2 watts?

 

[prcguy]

prc guy, if you add two 1 watt signals do you get 2 watts?

Yes, but explain how passing them both through a device that has 3dB loss gives you 3dB gain on top of that. The correct and only answer is that combining in phase actually gets you a 6dB increase and after hardware loss you are left with 3dB gain. Still waiting for your explanation or agreement or a make/model of a combiner that has essentially 0dB loss to prove your way of thinking.

 

[prcguy]

$14

You must mean at least 14 (10 to the 5th) dB$. If not I'll give you whatever Mr Q forfeits in this debacle.

 

[KevinC]

You must mean at least 14 (10 to the 5th) dB$. If not I'll give you whatever Mr Q forfeits in this debacle.

So if I have $14 into a two-way splitter, do I have $28 dollars or $7? I'm so confused!!!