A voting comparator can greatly enhance your repeater. By using multiple receivers and a s/n comparator (voter), you can expand your repeaters receive coverage in size, and in quality.

Basics-
The voting system is comprised of at least one... possibly several extra receivers placed at locations remote from the main site to extend quality coverage of the receive end of a repeater system.  These remote receiver sites are more properly called Satellite Receiver Sites.  At the satellite receiver site, a receiver on the repeater input frequency listens for a user.  If the user is getting into the satellite site, the remote receiver opens and turns on a link transmitter.  This transmitter sends the users audio back to the main repeater site where another receiver captures the link signal.  This linkback receiver is connected to the voter along with the main receiver so comparison of the two (or more) can be made.  The better of the two (or more) is selected by the voter and is transmitted out the main repeater transmitter.

A high profile transmitter is wasted if the receive coverage limits your range. Voting allows the tx and rx coverage areas to be balanced. The main receiver at the transmitter site may have reduced performance because of compromises made in the antenna system, or at a crowded site with a high noise floor or suffers from desense from other high powered transmitters at the site. The voter can help overcome the effects of these limitations because you don't have to rely on one great receiver. A few well placed low profile receivers can often be more effective. Even voting between two local receivers can get you some improvement.

Overlapping receivers allow the voter to remove the quick pops, fades, and picket fencing. As the user fades in and out of the various receivers, chances are good there is another receiver that is hearing the user fine at that moment. By continuously evaluating all the inputs, the voter can piece together quality audio to provide your controller, what looks to be one super receiver.

Hardware-
The Satellite Receiver site is made up of a quality receiver operating on the repeaters input frequency, and a linkback transmitter.  This set-up can be looked at like a crossband repeater.  The receiver captures a user and the resulting audio is relayed back to the main repeater site using a linkback transmitter.  The link-back is usually done on uhf or 900 megs and must be done on or above 222.15 MHz.  Many times the linkback transmitter is only on the order of a watt or two because the satellite site is quite possibly line-of-site to the main repeater.  At the main repeater site a receiver listens for the linkback transmitter to send the users audio.  This audio and COS is fed to the voter for comparison.

Some systems use more than one satellite receiver site.  There are different ways of getting the signal back to the main site depending on how much money you want to spend.  Some use different frequencies for the linkback of each satellite remote, while some use the same frequency.  The drawback of running different frequencies is the need for a separate linkback receiver for each remote site, at the main repeater site.  While this is the most expensive scenario, there are no drawbacks to its operation other than cost.  These extra sites can be configured "open" using 'carrier squelch', or in systems using CTCSS (commonly called PL) all sites can be configured to use (or pass)  the same PL tone.  The user needs to know nothing about its operation as it is transparent to them.

The alternative to the above is to use the same frequency for each linkback and then only one extra receiver is needed at the main repeater site and the voter can be a 2 channel model.  However,  since a user may be getting into more than one satellite site, each site must be PL'ed differently or a heterodyne will occur on the linkback frequency.  This makes the voting process more difficult for the user because they have to know what PL to use for the specific "zone of coverage" for a particular site.

It is possible to use phone lines to bring the audio to the voter, but leased lines are usually cost prohibitive. Clubs that are associated with universities, hospitals, or other facilities that include a large campus, can sometimes make use of the internal phone wiring for free. This conserves a link frequency and saves on linking equipment costs. By adding multiple receivers in these applications, improved in-building coverage can be obtained, as well as off campus coverage. When linking over a medium that only has an audio path, an idle tone (status tone) can be implemented to provide the COR status to the voter. Common frequencies for this are 1950 Hz (GE) and 2175 Hz (Motorola). When the receiver is squelched, the tone is sent down the line. When a user keys up, the tone is removed, and the audio of the radio traffic is sent down the line. A tone decoder at the voter end of the audio path provides the needed COR to the voter, based on the absence of this tone. This method can also be used with RF links, with the link transmitter on continuously.

Audio-
The hard part of a voting system is matching the audio's. When it votes, you don't want the audio characteristics to change. (other than s/n [quality])  That would allow you to hear the instant it switches between inputs. Also, if they have different tonal qualities, you are not getting a fair comparison. A shrill sounding link, compared to two regular sounding links, would appear to most voters as if it had more noise, and the voter may be biased away from that receiver. This would be unfortunate if it happened to be the better choice for intelligibility. Using all the same equipment is a short cut, and can help speed up the process of matching the audio's. The audio level from each rx is critical, but that is the easiest to match, simply by adjusting the input controls to the voter. Limiting on the link, or any compression on the audio, can cause a voter to misvote. The compression also compresses the noise floor, which is what the voter uses to make its decision. This can cause the voter to like that channel better.

Instead of modulating your link a full 5K, only use 3.5 or 4 K of deviation.  This will allow you to keep your limiter at 5 K to prevent problems, but will allow the audio to remain linear. You can boost the audio at the far end. When you listen to your link it may be a little low in level, but you will know why.

De emphasis and Pre emphasis-
De emphasis and Pre emphasis causes a lot of confusion. To simplify, a user transmits and his transmitter boosts the high end. This is pre emphasized audio. Normally a receiver does just the opposite and removes this boost. If you get the recovered audio straight from the discriminator, before the squelch gate, this is before the audio has that high end boost removed. The best way to deal with pre/de emphasis is don't. The user's tx will do the pre-emph. Let the users rx also do the de-emph. This means you need to get the non de-emph audio right from the discriminator when you interface to the remote receivers and link receivers. When connecting to the link transmitter, and main tx/controller, disable the pre emphasis, or insert the audio past this circuitry. Do not simply hook it up to your mic jack.  Transmitters (exciters) with a true FM modulator work best for this because they don't automatically pre-emphasize the applied audio like a phase modulator does.  If discriminator audio is used out of the satellite receiver, de-emphasis will need applied to the audio going to transmitters with a phase modulator.

Perfectly flat audio is not always possible, but the less you mess with it, usually the better. If the slope of the frequency response does not match exactly, each time you change this slope, you introduce error.  Some systems switch back and forth between pre/de/pre/de and they seem to work. You will have a better chance at success if you keep it simple, and keep it flat.  Definitely avoid double pre emphasis. This is a common mistake because the user transmits pre emphasis, but when you interface the remote receiver and get the audio from the discriminator, you are getting the audio before it has been de-emph'd. Then the audio is plugged into the mic jack on the link transmitter. Now you just added pre emphasis to audio that already had the high end boosted once before. When someone says there system audio is tinny or thin sounding, this is the first thing to suspect.

COR-
The COR presented to the voter, needs to accurately reflect the status and timing of the COR at the remote receiver. COR transitions should be fast, so don't have hangtime on the link. This would cause the voter to have a COR signal, without valid audio to compare. This dead audio during the link transmitter's hang time would look very quite to a voter..... and the voter would grab that channel. Gated audio from the receiver can also cause a similar problem.  If the squelch gate is delayed in opening, or ramps open, the selection process could be affected. If you get the audio directly from the discriminator, before the squelch gate, you avoid this problem. Once again, the less audio processing the better.

If you have to have hang time on the link transmitter, make sure the remote receiver is running open squelch audio to the linkback transmitter to allow noise to be sent to the voter. This should prevent the voter from selecting that link when its not really active, except at the end, when it is the only input present. The longer the hang time, the longer the squelch crash out of the voter.

CTCSS or PL-
If you will be using an access tone, here is one way to set up a system. The main/local receiver is connected directly to the voter, so it is handled the same as a link receiver. Decode the CTCSS and use this to signal the voter when it is active. Logically "And" the COR with Tone Detect. Read this explanation on the subject. There is no need for MICOR bi-level squelch on linkback receivers. The links rarely vary in strength. Stripping the CTCSS at this point is a good idea, but you may get by without worrying about it. The voter may see large differences in CTCSS level, and could be biased away from the input with tone. This is usually only a problem when one has the tone stripped off, and the other doesn't.  Try to keep things consistent.

At the remote site, you don't have to tone protect (PL) the receiver. Let the remote site pass everything, including the users CTCSS. Passing the tone at a sufficient level can be a challenge, depending on the equipment. The carrier squelch will allow any signal to bring up the link, but if it doesn't have the right tone, it will not get past the link receiver where the CTCSS decoder is. The reason for this is to improve the speed. With this method, there are only two decoders in line before the user hears it. The link rx, and the users rx both have a decoder that introduce a little lag.  This method avoids making this 3 decoders. I would also choose a tone frequency that is not too low. The higher tones decode slightly faster, and is also easier for the remote sites to pass. Remember, the higher the tone, the more you may hear through the high pass filter. (PL Hum on your audio) Avoid the extremes.

Many systems decode at the remote sites as well, but I wanted to show there are several ways to accomplish the same thing. An advantage of decoding at the remote site is it makes it easier, because you don't have to worry about getting the audio as flat at that crossband repeater at the remote site.  In other words, decoding the tone at the Satellite site may allow a lesser quality linkback to be used.

Implementation-
Installing a large voting system doesn't have to happen all at once, and probably won't. By letting the repeater grow and evolve, you can optimize the coverage area. Put on one remote receiver, then evaluate the change to your coverage. Strategically select your next Rx site, and gather equipment. Adding a 2nd receiver at the voter site can be a good stepping stone. Once you get your receive coverage up to par, raise the transmitter antenna or add a power amp. Then you can have an excuse to add even more receivers.

Things to Remember-
No hang time on the link tx.
No limiting on the link tx.
Match the audio's, the closer the better. Exact is nice, but close will work.
 

Questions ??-  Feel free to email either author below.
 

html 11-13-1999     Kevin K. Custer  W3KKC
Copyright 11-13-1999, 7-28-2000 Kevin Custer  W3KKC  and Duane Hall KA8GVH
Re-written 7-28-2000 by the above authors to include more detailed information.
All Rights Reserved.....