Can Two RF Signal Sources Connected in Parallel Increase Signal Amplitude?

[hilbert2]

TL;DR

About connecting two identical RF signal sources to the same antenna or measuring device.

Suppose I connect two identical signal generators to a dividing coaxial cable with two input ends and one output end. Is the signal amplitude from the output end the same as from one source, is it twice that or something in between? If this should be seen as analogous to two DC voltage sources connected in parallel, then you'd expect the signal intensity to be the same as from one source, but I'm not completely sure about this.

 

[DaveE]

This is all about exactly how that signal is combined into one cable. If you do it exactly wrong, you get zero signal out the end of the cable (and you might break a signal generator, if they are cheap). If you do it correctly the signals add to twice the power. Or, something in between.

It's not a DC thing. It's all about waves. Perhaps you can think of water waves traveling in two canals that are then merged into one canal. What if one wave is at it's highest peak at the junction, but the other wave is at it's lowest, how will they add up? What if they are in phase, i.e. both at their highest peak? That's not a perfect analogy, but it will give you the right idea.

I think this subject alone encompasses at least 25% of radio engineering. You can search for these terms: waveguide, combiner, splitter, impedance matching, VSWR...

 

[hilbert2]

Thanks for the advice. Good to see I knew what's a relevant question about this subject.

 

[sophiecentaur]

Summary:: About connecting two identical RF signal sources to the same antenna or measuring device.

you'd expect the signal intensity to be the same as from one source

Yes - each (parallel) source would be supplying the half the current (and half the power)

It's not a DC thing. It's all about waves.

It's also to do with the Impedances in the circuit, in practice. Combining signals can always lose some power unless they are of equal phase and amplitude. There are many different systems for combining RF signals, involving transformers, directional couplers etc. If minimising power loss is important then there's no point in just using parallel connection.

 

[DaveE]

It's also to do with the Impedances in the circuit

Yea, thanks for clarifying. You can never be reminded to many times that in RF electronics impedance is important. I'll do my very, very best to remember that for next time!

 

[davenn]

Summary:: About connecting two identical RF signal sources to the same antenna or measuring device.

Suppose I connect two identical signal generators to a dividing coaxial cable with two input ends and one output end. Is the signal amplitude from the output end the same as from one source, is it twice that or something in between?

It's double the power, 3dB, minus any circuit losses

It's a very common thing to do in amplifiers by combining the output from multiple modules/devices
A simple PCB one ...

It's also a very common thing to do with different spot frequencies within the same band, used in repeater systems
A combiner is used for this purpose which matches the 2 incoming 50 Ohm ports to a single 50 Ohm output port
A combiner can be used in reverse to become a divider, say 1 input to 2 output ports. eg when one coax feeds 2 antennas

 

[sophiecentaur]

It's double the power, 3dB, minus any circuit losses

Only if the impedances are right and if a circuit such as the one you illustrate. Two 50Ohm sources, connected in parallel will be the equivalent of a single 25Ohm source and the same source Volts. The ratio of powers, delivered from an emf of V to the 50Ohm load is 1/4 : 4/9, which is 1.78 times the power delivered from a single source. That's not as good as the X2 that a properly matched network will give you. There's a shortfall of about 0.5dB. There are other problems with that simple connection as there can be a mismatch, seen by each of the generators and the load (on the end of a feed line). Your system is much better in many ways.

Edit: all this applies when the two sources are the same frequency and in phase. If you are trying to combine two different transmitters on different channels into one antenna feed then you need to have 'notch' filters `( or equivalent) so that each transmitter feed appears as an open circuit at the operating frequency of the other. When you're dealing with tens or hundreds of kW, Power loss can be a major issue but the technology has been used for many decades. Hats off to the RF Engineers of the past who had to make do with very basic measuring equipment and no computers to help them.

 

[f95toli]

Only if the impedances are right and if a circuit such as the one you illustrate. Two 50Ohm sources, connected in parallel will be the equivalent of a single 25Ohm source and the same source Volts.

I believe Davenn assumed that the sources were connected using a combiner/splitter (it the same typically the component) which will provide isolation between them and ensure that everything stays at 50 ohm

Connecting two sources directly (via say a BNC-T) is usually a bad idea; and I would never try it using an actual RF source. There is real risk that you could damage the output of the sources,
At the very least I would include some protection by adding some resistors in series with each source; this won't help the RF performance but will reduce the risk of damage.

 

[sophiecentaur]

adding some resistors in series with each source;

There are a number of resistive combiners but they always waste power. Favourite is a 3dB coupler which will give all the combined powers of two identical sources out of one port and any imbalanced power out of the fourth port - which you would normally terminate with 50Ohms. There are many designs of 3dB couplers. I remember Sage Wireline which could be cut to length (λ/4). I don't know if it's still available but a cheap solution if you don't have good PCB making facilities.

 

[f95toli]

There are a number of resistive combiners but they always waste power. Favourite is a 3dB coupler which will give all the combined powers of two identical sources out of one port and any imbalanced power out of the fourth port

I'm going to nitpick here and point out that what you are describing is neither a coupler nor splitter, but a 3dB hybrid (if I understood the description correctly)

Splitter/combiners (3 port device) , hybrids (4 ports) and couplers (usually 4 ports, but one port is typically terminated at the factory) and all be used to "divide" signals; but they work somewhat differently and if you buy one you need to make sure it does what you want.

 

[sophiecentaur]

I'm going to nitpick here and point out that what you are describing is neither a coupler nor splitter, but a 3dB hybrid (if I understood the description correctly)

Yes. A rose is a rose. Hybrid used as coupler or splitter. I couldn't remember the name that was used to describe the technique of combining a number of solid state UHF amplifiers to obtain a useful output for broadcast relay transmitters but an array of "hybrid" couplers for splitting the drive and then combining the outputs was common when the individual device power was limited to just a few watts. It can be done with stripline on a printed board or with wireline . It's fault tolerant because a failed component will not mean the transmitter loses all output power.
You can do the same thing with wound components at lower frequencies.

 

[f95toli]

My comment above was not meant as criticism. It is just that it usually better to use purpose made splitters instead of hybrids. Splitters are usually cheaper and have wider bandwidth.
If you are working at low frequencies (up to a few hundred MHz) you can buy cheap splitters from places like Minicircuits.

 

[sophiecentaur]

Splitters are usually cheaper

3dB couplers work over about an octave. What sort of 'splitter' are you referring to`````?

 

[f95toli]

Splitters/combiners can cover more than an octave. The ones I use in the lab typically cover 2-18 GHz (I mostly work between 3 and 10 GHz, but sometimes it is good to be able to go to higher frequencies).
Typical insertion loss (in addition to the 3 dB) is something like 1dB which is a bit higher than ones that only cover a decade (which have insertion losses of about 0.4 dB or so) but the isolation is still close to 20 dB and that is usually the most important parameter

See e.g.
https://www.atlanticmicrowave.com/c...plitters/2-way-2-18ghz-power-divider/splitter

For lower frequencies (5-2500 MHz)
https://www.minicircuits.com/WebStore/dashboard.html?odel=ZAPD-2-252-N+

Another difference between couplers (4 port devices) and splitters (3-port) is that couplers are -usually- directional whereas splitters can also be used as combiners which makes them a bit more versatile.

I use all different types in my lab (splitters, hybrids, combiners); which one works best for "dividing" a signal depends on the application.

One advantage of couplers is that they can handle much higher powers than splitters (which are typically limited to about 1W), so if you need to split the power from a power amplifier you can't really use the latter. Hence, for UHF transmitters you would indeed use a coupler but for regular lab use a splitter is usually more convenient.

 

[Jody]

I think any chapter or section from a textbook talking about Wilkinson might help answer the question (Pozar's Microwave Engineering book is a classic and it's in Chapter 7). OP didn't give specs or specifics such as bandwidth and so the side conversation about couplers versus hybrid might be confusing to non-RF readers.

 

[Klystron]

Summary:: About connecting two identical RF signal sources to the same antenna or measuring device.

Suppose I connect two identical signal generators to a dividing coaxial cable with two input ends and one output end. Is the signal amplitude from the output end the same as from one source, is it twice that or something in between?

If this should be seen as analogous to two DC voltage sources connected in parallel, then you'd expect the signal intensity to be the same as from one source, but I'm not completely sure about this.

If the OP is searching for an analogy from basic electronics to teach and/or better understand RF, consider visual waveform representations rather than 'parallel direct current' models. Even pulsed DC analogues leave out vital details.

Before personal electronic computers became common, I often found Lissajous figures helpful teaching beginning RF including comparator circuits, RF mixing, signal conditioning and parametric oscillators/amplifiers. The latter intrinsic parametric equations also serve as introduction to understand masers and antennae designs.

For measuring small changes between your two input waveforms consider using Moire patterns as an analogue for wave interference; again, a highly visual representation of interpenetrating fields.

In keeping with your antenna model, simplified visual representations of phased array radar systems may offer guidance. Perhaps the expanding radiation wave front in this excerpted illustration might serve as a replacement model for DC voltage amplitude? Or am I over complicating the question?



I concur with prior comments that 'wave' and beam models help students visualize and then manipulate RF electromagnetic fields. Also agree that understanding impedance, particularly distributed impedance, remains vital to understanding RF circuits and systems; justifying the additional comments concerning impedance matching and coupling components.

(Pardon for using online encyclopedia as references but most of the equations appear satisfactory and the visuals help understanding.)

 

[Paul Colby]

Summary:: About connecting two identical RF signal sources to the same antenna or measuring device.

Suppose I connect two identical signal generators to a dividing coaxial cable with two input ends

On thing to keep in mind is that a passive network device like a splitter etc. can't generate power. They can only dissipate power. If your cable is providing more than the power nominally supplied by the sources then it is doing so because you've coaxed more from the sources. This can happen if the sources are feeding a better impedance match. If all things are ideal, no power loss and no impedance mismatch, you should see just the sum of the power supplied by each source.