Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Keith Broughton]

I have recently seen a diagram of a setup where multiple com transmitter outputs are sent to a passive combiner/splitter and each output of the base stations (and IEM TX) are connected through an isolation device.
Could someone expand on why this is required and how the ISO actually works?
I am presuming it is to prevent reflected TX signals folding back on other TX outputs.

[Jason Glass]

I have recently seen a diagram of a setup where multiple com transmitter outputs are sent to a passive combiner/splitter and each output of the base stations (and IEM TX) are connected through an isolation device.
Could someone expand on why this is required and how the ISO actually works?
I am presuming it is to prevent reflected TX signals folding back on other TX outputs.

Hi Keith,

Your presumption is correct, in that an isolator is effectively a one-way valve for RF signal flow.  They are indeed used to prevent signals that reflect from various points along the path from leaking back into active gain stages.  That sort of leakage is where a lot of intermodulation products are created and retransmitted.

The way they work is that they're a magnetic (usually ferrite) semicircle, with an input and two outputs.  That first part is known as a circulator.  To make a circulator into an isolator, the second output is terminated with a resistive load.  Because of the electromagnetic relationship between the circle's flux and the signal's h-field, It can only travel in one direction along the ferrite.  Signal enters the input, travels around the circle, and exits the first output.  Any signals reflecting back into the first output can only travel in the direction of the second output, where they are dissipated as heat in the load.

It's a bit like a crazy kind of traffic roundabout, with exit 2 being a turn onto a one-way dead end street.

[Keith Broughton]

Thanks Jason. Thanks for the details.
 At first I thought the ISO was an RF cable isolator which is similar to an iso transformer in audio.
Now I know what it does, it makes sense in combining multiple TX sources.
Always something new to learn in RF world

[Henry Cohen]

Your presumption is correct, in that an isolator is effectively a one-way valve for RF signal flow.  They are indeed used to prevent signals that reflect from various points along the path from leaking back into active gain stages.  That sort of leakage is where a lot of intermodulation products are created and retransmitted.

The way they work is that they're a magnetic (usually ferrite) semicircle, with an input and two outputs.  That first part is known as a circulator.  To make a circulator into an isolator, the second output is terminated with a resistive load.  Because of the electromagnetic relationship between the circle's flux and the signal's h-field, It can only travel in one direction along the ferrite.  Signal enters the input, travels around the circle, and exits the first output.  Any signals reflecting back into the first output can only travel in the direction of the second output, where they are dissipated as heat in the load.

It's a bit like a crazy kind of traffic roundabout, with exit 2 being a turn onto a one-way dead end street.

To expand on Jason's explanation as it pertains to the real world^© of RF (LMR and cellular infrastructure), isolators also serve the purpose of protecting the transmitter or amplifier output stage from damage due to a catastrophic fault resulting in a large power reflection. This is critical when dealing with tens to scores of thousands of watts.

And on a pedantic note, isolator ports are generally referred to by port number, as opposed to input or output, even though the higher quality, and higher power, units are in fact built around Port 1 being the TX input, Port 2 the TX output, Port 3 the load, and if a 4 port design, Port 4 being the secondary load. But in many cases, circulators are fully symmetrical devices.

[Russell Ault]

I still keep having a recurring dream about using a four-port circulator to duplex an IEM TX and (half a) microphone RX onto a single antenna. With a similarly-equipped custom beltpack and a very clever diversity regime you could almost get away with putting a saxophone (for example) player's mic and IEMs on the same frequency (how's that for spectral efficiency!). But I digress.

-Russ

[Jason Glass]

I still keep having a recurring dream about using a four-port circulator to duplex an IEM TX and (half a) microphone RX onto a single antenna. With a similarly-equipped custom beltpack and a very clever diversity regime you could almost get away with putting a saxophone (for example) player's mic and IEMs on the same frequency (how's that for spectral efficiency!). But I digress.

-Russ

Nope. Not same frequency, under any imaginable circumstance under today's understanding of physics with FM analog or PSK digital audio transports.  Same antenna is plausible.  But that's a giant difference in definition.  The only way what you propose works is with time domain management synchronized between all devices.  Meaning nothing ever transmits on the same frequency at the same time, but data packets are sequentially organized and transmitted in two directions with great precision. And also meaning latency while the system waits for complete transmission of each packet.

But please keep dreaming!  Mine have been getting weirder every night for the last 5 months.  Hopefully one of us will have one just weird enough to work.  ;-)

[Jason Glass]

I still keep having a recurring dream about using a four-port circulator to duplex an IEM TX and (half a) microphone RX onto a single antenna. With a similarly-equipped custom beltpack and a very clever diversity regime you could almost get away with putting a saxophone (for example) player's mic and IEMs on the same frequency (how's that for spectral efficiency!). But I digress.

-Russ

And, FWIW, I have successfully diplexed RADcom using only filter type diplexers on several occasions.  The extreme frequency spacing between TX and RX with that system's utilization of VHF and UHF makes it possible.

[Russell Ault]

Nope. Not same frequency, under any imaginable circumstance under today's understanding of physics with FM analog or PSK digital audio transports.[...]

In a rack, you've got a TX and an RX (both FM, say), and a four-port circulator. Port one is the TX, port three is the RX. If you load ports two and four, no(-ish) signal should go from the TX to the RX. If you replace the load on port two with an antenna, the only signal that the RX will receive from the TX will be whatever is being reflected.

If you create a second rack with the same configuration, you can operate the TX in both racks on the same frequency and each RX should capture onto the other rack's signal so long as the other signal at the local RX input is more powerful than the local TX's reflections.

This would require very well-matched components and careful antenna placement (and probably environmental design). It probably wouldn't work for portable applications, either, and the whole idea is likely just beyond the realm of practicality. That being said, maybe I'm missing something, but on a purely theoretical level I feel like it works.

And, FWIW, I have successfully diplexed RADcom using only filter type diplexers on several occasions.  The extreme frequency spacing between TX and RX with that system's utilization of VHF and UHF makes it possible.

You make a good point about diplexing. Should't be too hard to do that with a VHF/UHF split IEM RX and betlpack TX, either (although designing the antenna would be...interesting). Cheaper than a circulator, too.

-Russ

[Jason Glass]

(although designing the antenna would be...interesting).

-Russ

The beauty of the diplexer scenario is that we can use off-the-shelf antennas!  I stumbled into this almost by accident when I saw that the bandwidths of the components freakishly lined up.  We insert identical VHF/UHF diplexers (<1dB measured pass bands' insertion loss) on each end of our low-loss coax transmission line, connected to the diplexers' common ports.  We connect the base station RX to the VHF pass port, and the base station TX to the UHF pass port (I also insert a wideband 470-698 MHz isolator here, just to be safe), then at the antenna end we simply connect a RAD VHF LPDA or whip to the VHF pass port, and a PWS UHF helical 8089 or Sennheiser A1031-U, depending of course on the application regarding pattern coverage, to the UHF pass port.  It worked perfectly each time.

[Russell Ault]

The beauty of the diplexer scenario is that we can use off-the-shelf antennas!  I stumbled into this almost by accident when I saw that the bandwidths of the components freakishly lined up.  We insert identical VHF/UHF diplexers (<1dB measured pass bands' insertion loss) on each end of our low-loss coax transmission line, connected to the diplexers' common ports.  We connect the base station RX to the VHF pass port, and the base station TX to the UHF pass port (I also insert a wideband 470-698 MHz isolator here, just to be safe), then at the antenna end we simply connect a RAD VHF LPDA or whip to the VHF pass port, and a PWS UHF helical 8089 or Sennheiser A1031-U, depending of course on the application regarding pattern coverage, to the UHF pass port.  It worked perfectly each time.

Ah, gotcha! I was angling for a single diplexer and a dual-band antenna (especially for the on-person variation). According to my math, any (simple) antenna tuned for high-band VHF should also work just as well for the upper three-fifths of the remaining UHF spectrum. This would definitely work with a simple monopole/dipole; I'm not as sure how it would play out for a LPDA (and even less certain for something like a helical). Of course, a dipole/counterpoised monopole centred on 153 MHz would be just shy of a metre long, so the costume department would really have to be willing to play ball.

Conversely, anyone got a a RAD VF-1 (or a CP-1V) and a VNA kicking around? I'd love to see the SWR between 522 and 608 MHz (assuming they don't have extra built-in filtering, that is).

-Russ

[ProSoundWeb Community]