Repeater - No Duplexer - Receiver Desense Testing

[marcspaz]

 

Sorry, this is a little long.  Hopefully not TL;DR long.

This is somewhat Emergency Communications related, but will hopefully answer some questions a lot of new people have about needing a duplexer for a repeater.  I have to say that you need a duplexer if you want to use one antenna.  There is no way around that.  However, you can run a repeater with two antennas if you have the physical space to work with.

 

I have been playing with repeaters for some time.  I have never run a duplexer with any of my repeaters, but rather using two antennas physically separated horizontally.  This is because most of my repeaters are intended to be used in the field during an EmComm event.  So, rather than re-tuning a duplexer while trying to resolve frequency coordination issues in the field, we can simply pick our new frequency pair and carry on.

 

I never really had any desense issues, because I have historically been using two cross-band repeaters, occasionally with as much as 1,000 feet of physical separation and simply RF-linked the two systems.  The space between the two systems resolved any desense issues through natural atmospheric attenuation.  

 

This weekend, I tried doing some testing to see how vertical separation would work, with both antennas on the same mast.  Based on advice from several reliable techs and engineers in the commercial RF world, I figured I would give it a shot.  Unfortunately, my results were not as good as what my research implied it would be.

 

The idea behind stacking the transmit and receive antennas is that there are nulls above and below the axis of the antennas. The more separation, the wider the null in the field gets and the more attenuation you get.  My problem is, with a portable mast, I can’t separate them enough.  I tried clocking the antennas on the mast and had no change in performance worth mentioning, regardless of where they were. 

 

First, only using 5w on 145 Mhz, I stack the receive antenna on the mid-section of the mast and the transmit antenna on the top of the mast.  Both antennas have ground plans.  The initial separation was 40 inches.  The transmitter was splashing the receiver with -4dBm / 0.14 volts.  I moved them to 80 inches apart and it dropped to -57dBm… still not great at an additional 17 dB of attenuation.  Finally, I moved the antennas to 360 inches, a full 30 feet apart.  The splash only dropped to -74dBm, about an additional 23dB compared to 40 inches.

 

As a quick note, I was using LMR400 cable for both antennas.  Swapping the antenna positions made it so the repeater was completely unusable.

 

The next test was using horizontal separation but wired instead of RF linking.  I put the two masts up at 35 feet with the antennas at the top, and same elevation.  The masts were only 130 feet apart from each other.  I placed the repeater in between the two masts for general testing.

 

Testing in this configuration showed that at 20w the transmitter was splashing the receiver at -120dBm… a shade over 38dB of attenuation compared to stacking the antennas at 40 inches of separation and almost 47dB of atmospheric attenuation in total.  At 5w, it was attenuated to the degree that splash was no longer measurable on my SA at the offset frequency.  This is pretty good for what it is, as my repeater’s receive sensitivity is 0.2uv / -121dBm.  This means that at 20w the transmitter is preventing only the very weakest of signals that the repeater could possibly hear from actually being heard.

 

The cable I am using has a loss of 1.5dB per/100’ on 150 MHz.  So, I could increase the physical separation to 340 feet (two 200 foot cables on each feed) and still get 50% of my power to the transmit antenna.  Or, I could set the repeater off-center, closer to the receive antenna to minimize line loss on the receive side, turn the power up to 50W on the transmit side and still have plenty of natural atmospheric attenuation for good performance. 

 

I haven’t performed any testing on the UHF spectrum yet.  However, you need less space for UHF than VHF.  I would suspect that you would see at least another 10dB of attenuation at the distance I tested (130’) or the antennas could be placed as little as 40 feet apart with similar performance.  I will be testing again on 440 MHz, possibly on Thursday.  So, I should be able to confirm.

 

Obviously, having the transmitter and receiver tied together with cables will not provide the low loss and optimal atmospheric attenuation that can be obtained with split-band RF linked systems, but it reduces the amount of hardware you need for radios, power sources and physical security.  As long as you have the 200 feet to separate your two antennas, you should be fine.

[tweiss3]

This tracks. I never thought about the crossband as links to eliminate duplexers though, that's a good back pocket solution.

 

By the numbers, to get 50db isolation:

@145MHz, you need either 700' horizontal separation or 24' vertical separation

@440MHz, you need either 22' horizontal separation or 8' vertical separation

50db isolation isn't great, but a repeater will probably "operate" at that level.

 

A decent duplexer will exceed 60db isolation:

@145MHz, you need either 2100' horizontal separation or 43' vertical separation.

@440MHz, you need either 650' horizontal separation or 14' vertical separation.

 

When discussing separation, it is from the closest part of each antenna, usually clamp to clamp for horizontal separation and top of the bottom antenna to the bottom of the top antenna.

[marcspaz]

43' vertical separation....  We're going to need a bigger tower.  LOL

 

There is some serious benefit to the commercial duplexers, for sure.  I have seen some better units with 90dB notches.

[WRKC935]

Well, the commercial radio method of conducting this test is with a signal generator and an isolation Tee and something that will measure 12dB sinad.

You inject signal into the isolation Tee that is connected to the receiver of the repeater.  You adjust the signal generator so the test equipment indicates 12dB Sinad which is a 12 dB signal to noise ratio.  Once the initial number for signal generation is reached you turn on the transmitter and then increase the output of the signal generator until you have again achieved 12 dB Sinad.  The difference is the loss of receive sensitivity that the repeater is experiencing. 

The actual signal level numbers are NOT applied to the effect of the loss, you are ONLY looking at the difference.

You can also before or after do a signal test for 12 dB Sinad directly into the receiver, which will give you a real world number of what the receiver is capable of with out the interfering signal present from the repeaters transmitter.  This number typically is going to be between '117 and -119dBm of signal level for most repeaters.  Some are better some aren't quiet that good. 

Now you apply the change in signal level required to bring the receiver back to 12dB Sinad and ADD (remember it's a negative number) that to the direct receiver test. 

Not running a duplexer and getting that level of isolation (typically 75 to 95 dB) and instead applying your isolation numbers of only 38 to 47 dB, you are giving up 30 dB of signal sensitivity more or less.  Meaning instead of -115dBm of required signal, you need -85 dBm of signal level to achieve the same level of receiver performance.  Putting that into perspective, a typical subscriber receiving a signal from a transmitter at -95dBm signal level will indicate a full signal (four bars) on the front of it.  It's PAST the level needed for a typical receiver to have a FULL QUIETING signal.  And you are going to need 10dB MORE signal to achieve 12dB sinal which has a degree of noise in the receive signal but is fully copyible but is NOT full quieting.

And while I sort of understand your stance on using this solution for emergency repeaters during an activation, there might still be a better way.  Now for VHF, you are hung.  There is no simple way around a duplexer on VHF.  And if we are talking about HAM radio with the 600Khz frequency split, for a standard repeater, your really hung.  And tuning is going to be needed.  But there are even options here.  The first being a set of high Q pass cans that are tuned for each frequency (single can for each) and dual antenna's.  This is a far better option than just relying on horizontal separation for isolation.  If you can work that out, you will see a marked improvement in repeater performance.  It's not going to be a good as a duplexer, but it's FAR superior to what you are doing now. 

GMRS isn't that way depending on the type of duplexer you choose. 

A Notch duplexer (small mobile types) can be tuned for the middle of the repeater frequency range and be used for that full range with some degradation on the band edges.  A pass / notch duplexer, especially a good one isn't going to allow for that.  The SWR the transmitter see's will climb too fast for it to work. 

Now, one thing you can do with the ham allocation is look at the probable repeater pair assignments that will be offered and do something similar with those.  Tuning up a couple duplexers, as long as the frequencies are close enough together for them to fit in the notch duplexer. 

And please understand, I am not trying to poo poo on what you are doing.  I am only trying to give you options and point out the math on what you can gain by taking a different path.

 

[marcspaz]

@WRKC935 understood and all good information. I have a UHF duplexer, so no issues on that end. I am trying to solve for an amateur VHF EmComm issue if we need to use a frequency other than what any duplexer may be tuned for, without having to re-tune the duplexer on the spot in the field.  

 

In the past, horizontal separation has worked extremely well for us with my original repeaters, as the up-link and down-link are to separate radios a significant distance apart and RF linked.  I was hoping to accomplish the same results, but with a single repeater... though, it's not promising. There is a good possibility that I will tune a duplexer for the single repeater solution and tote my original split system in case it's needed.

 

The main reason I shared what I did is because several people over the years have askes about if a duplexer is needed for a repeater or not.  This was a real-world result rather than a theoretical bench test. The bench test is great, but for me, actually hearing and experiencing the results translates to something more meaningful and easier to digest. Plus, it helps me explain in some odd way that no duplexer is needed in limited circumstances... but it's far from perfect. 

 

Something I did find interesting is, while there is noticeable desense in the test I did, with a 35 foot portable mast, I actually ran out of LOS before the desense led to loss of comms.