Up one level
Back to Home
Some Thoughts on Repeater Performance and the Isolation Between
Receiver and Transmitter
By Mike Morris WA6ILQ
Repeater performance is defined by RF performance as well as audio performance. You can have the best sounding repeater on the planet, but if the users can't get into it or hear it then it's worthless. Audio performance is covered elsewhere on this web site. Good RF perfomance is a balance between receiver and transmitter... you want neither an "elephant system" (all ears and no mouth) or an "alligator system" (all mouth and no ears). It gets complicated when some users use flea-powered handhelds, some use 10w mobiles, some use 50w mobiles, and others use hilltop remote bases. What set of users do you use as a reference for "balanced"? You have to consider the demographics of your users.
In many cases fixing a transmitter problem is easier than fixing a receiver problem. Good receiver performance depends on being able to hear the users... and the receiver performance is divided into two major areas: sensitivity and selectivity. Sensitivity is obvious - how weak a signal can you hear clearly, but selectivity is a more difficult parameter to measure. A good analogy to selectivity is this: imagine trying to focus on only the person sitting next to you whispering during a LOUD rock concert - the more surrounding noise you can tune out the better you can hear the signal you are trying to hear.
To continue the rock concert analogy, it's not how well the receiver works on the bench, it's how well it works in real life, at the site. You may have good ears, but when the noise surrounds you just how well can you hear the person next to you? And if you can't accurately measure the before and after performance, you can't really tell if you've made any improvements, and if so, quantify how much. You need to be able to measure the Effective Sensitivity of the receiver, and the methodology must be repeatable.
Receiver performance in a duplex environment (doesn't matter if it's repeating or not) depends on isolation between the repeater receiver and not just the repeater transmitter but all other emitters within earshot. I used the term "emitters" instead of "transmitters" deliberately - you have to consider every RF source that might affect the receiver. Most folks think that you need some magic amount of isolation for good performance, and buy a duplexer with more than that magic number, and end up disappointed. In fact any more than “enough" (plus a reasonable amount of headroom) does nothing but drain your wallet.
In reality you have to isolate your receiver from all of the other RF sources at the site that your receiver can hear - the local oscillators, the transmitters in the building (and the next one over, and the one down the road), the RF-noisy tower strobe light blinker, even that intermod-creating rusty tower bolt or the internal hidden crack in the antenna. Isolation starts at the antenna jack(s) of the equipment and includes everything - the best duplexer in the world won't help if the jumper cable between it and the receiver was made from thirty year old Radio Shack CB coax with 75% shield. Actually, all it takes is one nickel-plated connector, or even one good quality connector that was improperly installed.
Isolation is inherent in the design of the radio equipment, for example minimizing the off channel energy (noise) created by a transmitter exciter (sometimes called "grass" after the pattern it makes on a spectrum analyzer screen), and while it's better if it is designed in, sometimes it can be added on. As one example, I've seen a pass cavity inserted into the line between an exciter and a PA deck. If necessary you can add additional front end selectivity with an additional pass cavity in front of a receiver (at some cost of insertion loss), but sometimes the need for increased add-on isolation can be reduced by using careful tuning (with a spectrum analyzer) of both the transmitter and the receiver. Yes, the receiver also. I say this because I watched a gentleman fine-tune a receiver local oscillator / multiplier chain with a spectrum analyzer on the injection point of the mixer, and he was able to reduce a mix in the receiver front end by about 6 dB (there were nice smooth peaks on the test set meter at each test point, but on the spectrum analyzer there was a very narrow peak where the desired signal was at maximum and the "grass" markedly dropped in level). On the receiver side, you can take advantage of the fact that a repeater is a single channel device and you can ignore the tuning instructions that tell you to stagger tune the front end and multiplier chain, and instead precision tune the receiver for one frequency.
An increase in receiver isolation can be achieved by looking at the receiver front end by sweeping it with a sweep generator. In most cases you will see that the further you move the transmitter away (in frequency) the better. The 6 meter repeaters in some areas have 1 MHz offsets where others have 0.5 or 0.6 MHz, and those that have 1 MHz have it easier. Many others have found that the 5 MHz offset on UHF makes it much easier than the 0.6 MHz on 2m. And crossband is even easier, but watch the math. Don't try to listen to 443.425 MHz and talk on a 147.81 MHz transmitter at the same time.
Sometimes you come up against a brick wall of equipment performance: If you are using a modern mobile radio as a transmitter (or as an exciter), meaning synthesized with a broad-tuned PA deck you may find that the broad-tuned or lumped-LC front ends have reduced selectivity and front end rejection. That's a polite way of saying that the super-wide front ends (sometimes referred to as wide eneough for DC to Daylight) are a total joke when it comes to trying run full duplex. You need some serious selectivity to run duplex. Then you have the low level transmitted noise from the receiver and the PA deck - that is another limiting factor. Take this as a strong hint: for years crystal-based exciters were cleaner than synthesized ones, and tube-based transmitter power amplifiers are cleaner than solid-state ones. For years Motorola specified 4-cavity duplexers on 90-100 watt tube based high band and UHF repeaters and had to switch to 6-cavity designs to get equal performance with solid state repeaters of the same power level and sensitivity at the same sites.
Modern amateur and commercial radios are designed to hear and talk over multiple megahertz of bandwidth, therefore the front end filtering and the interstage filtering is almost nonexistent, meaning that they are prone to in-band desense. Time after time I have seen people start out building a repeater by taking two synthesized mobile radios and connecting one up as a receiver and the other as a transmitter, and wonder why they have desense or overload. Or they burn up the transmitter (google the term "duty cycle"). Then they find an "elmer" and a few months later they are looking for a second-hand MICOR, MASTR II or MSF5000 station - i.e. something with a tight front end and a tuned exciter, and designed for continuous duty from the outset. It's interesting, but 1960s-1970s tube based radios were inherently tighter and much cleaner, therefore requiring less isolation in the antenna system. The 150-170 MHz receivers of the day covered roughly 1 MHz before you saw degredation, and maybe 1.5 MHz if you "stagger tuned" them. The transmitters covered 1.5 MHz, maybe 2 MHz if you stretched them. If you peak-tuned them on one channel (like a repeater) you could get by with a minimal duplexer. Radios on low band (10 meters, 6 meters or the commercial frequencies in between) covered a similar percentage... if you had two low band channels more than 275-300 kHz apart you probably had two separate radios (base or mobile), feedline and antennas.
At some point you can't get any better because of other parameters outside the radio chassis - maybe the noise floor at your site is higher than the effective sensitivity of the receiver (and as said on the linked page above, the effective sensitivity is NOT the bench sensitivity). I suggest you read the "effective sensitivity" articles on this and other web sites because improving that number is the single best thing you can do to improve the ability of the repeater to hear your users.
Or maybe it's the cheap duplexer. Mobile duplexers are tempting for the newbie repeater builder since they are half the price of a "real" duplexer, and he's happy until he discovers that what he bought is nothing but notches (sometimes called a "notchplexer"). Yes, it notches the transmitter out of the receiver, and the receiver out of the transmitter, but offers nothing in keeping any other radio out of his receiver - the notches are very narrow and the unit might as well not be there at all for any frequency except what the notch covers. I saw a dramatic example of this many years ago when a friend's UHF two-mobile-radios-and-a-mobile-duplexer repeater was receiving interference a week after it was installed at a site. The notchplexer was offering zero protection to the receiver from the UHF paging transmitter in the next building down the road, and he was certain that the paging transmitter was "dirty". A quick look with some proper test equipment showed that the paging system was clean, and his duplexer was doing absolutely nothing to protect his recever from a simple front end overload condition from a transmitter 20 MHz and three hundred feet (100m) away. Adding a pass cavity to the receive line cured that problem, and a second one in the transmit line insured that his transmitter would not cause any problems to any one else. The new repeater owner learned a lesson about purchasing the wrong equipment for the job. A few weeks later a real 6-cavity pass-notch ended up on the repeater, and the mobile duplexer was shelved (and later ended up in a portable repeater system).
Or maybe it's that the feedlines are not enough for the site: one six meter repeater I worked on was limited by the fact that the transmit and receive antenna feedlines were coupling into each other... Two parallel runs of over 150 feet (50m) of 97% braided shield coax tie-wrapped to each other and to the tower leg means that there is 150 feet of 3% holes facing each other... made a nice lossy coupler for a 100w transmitter into a sensitive receiver. Only after switching to heliax on both runs with it's 100% shield was that problem cured.
Another limiting factor might a missing shield on an exciter, a PA deck, or on a receiver board... remember, the RF design engineers added those shields for a reason, not for cosmetic looks.... and someone may have lost a shield before you got the radio and you don't know it's missing... (the voice of experience from the first MICOR station I worked on)
Or it might be one defective jumper... I saw a situation where the receive side of a six-cavity highband duplexer fed a preamp which fed a receiver. The preamp was physically adjacent to the receiver. The jumpers from the duplexer to the repeater cabinet was new, good looking coax but when you pulled back the jacket it had a loose weave braid. The system was desensed and the owner could not figure out where the grunge was coming from ("But I've got all new cabling"). Replacing the jumper from the duplexer to the preamp with Superflex (100% shielding) fixed 90% of the problem, replacing the one from the preamp to the receiver fixed the rest, but we replaced the jumper from the transmitter to the circulator and from the circulator to the duplexer on general principles. And the connectors have to be installed properly on the cable, especially on the superflex. The PL-259 connector is frequently found on older duplexers and cavities, and was never designed for superflex, but Tony King W4ZT developed an innovative way to make it work, and did a photo article on installing PL-259 connectors on FSJ1-50A 1/4" Superflex Heliax.
A comment by Eric Lemmon WB6FLY on the repeater-builder mailing list:
Be very cautious about buying any coaxial cable that has the word "TYPE" or "LIKE" printed on its side, even if the maker claims that is "military specification" cable. That one word can allow the maker to market a totally inferior product to unsuspecting buyers. Genuine, RG-214/U coaxial cable made to the MIL-C-17 specification (and it will be labeled as such) has double silver-plated copper shields. Several companies manufacture an RG-214 "TYPE" cable that is very similar, but without the silver plating. As you would expect, it's a lot cheaper than the genuine MIL-specification RG-214/U cable. Such junk cable may also have less dense braid coverage than the genuine cable.I don't know if the defective jumpers mentioned above were "TYPE" or "LIKE" cable (it's been several years), but they could have been.
And while Eric specifically mentions RG-214, the problem exists on many, many types of cable. I've seen "RG-174 LIKE" cable that had a very loose shield - the gentleman standing next to me commented that his window screening had a tighter weave.
Sometimes the isolation problems are just not possible to overcome due to the equipment in use. In the USA the UHF band is 420-450 MHz, with repeaters from 440-450 MHz... and most local band plans have some spectrum between 420 and 440 MHz allocated to point-to-point links. The Motorola Mitrek is an older crystal controlled radio and is popular as a UHF link radio - but some design oversights can severely limit the performance (and the Mitrek is not alone in having design oversights).... when you duplex a 406-420 MHz Mitrek mobile (which will get up to 430-431 MHz without a problem) you will find that you have spurs on the transmitter. The spurs happen because the receiver local oscillator multipliers just happen to couple into the transmitter multipliers due to a poor choice of multiplying factors, too-close physical spacing in the original design (i.e. across-the-PC-board coupling), and the fact that the original design had the receiver multipliers powered down while it was transmitting... (note that the 450-512 MHz models will do the same thing but most of the time the frequencies are separated far enough to not cause a problem)...
Another limiting factor can be when the transmitter antenna couples RF energy into the tower itself, and the tower couples it back into the receiver antenna. This is very visible on 10 and 6 meters, but can affect systems as high as UHF. Grounding the feedline to the tower can help... Feedline manufacturers such as Andrew typically recommend a ground strap at the top and bottom of the support structure, physical clamping every so often, and an additional grounding kit where the feedline enters the building. Good practice suggests additional ground kits at specific intervals on really high towers. The basic idea is not RF but lightning: the grounding kits keep the feedline and tower at the same potential so there isn't any arcing between them during lightning strikes. With kiloamps of current you can get kilovolts difference with very small resistance differences between the tower leg and the feedline - and you will have those differences between them due to the differing electrical resistance of steel versus copper. In addition I have seen feedlines stripped from the tower by the intense magnetic fields caused by lightning hits... don't forget that current through a conductor (the feedline shield) generates a magnetic field, and the parallel current through a steel tower leg (or an adjacent feedeline) generates a like magnetic field, and like magnetic fields repel each other.
Back to antennas... I've seen interfeedline coupling when a receive antenna is mounted above the transmit antenna and the transmitter RF power couples into the outer shield of the receive antenna feedline... grounding the shield at multiple places helps (especially if you can do it at a voltage peak). So when you order the grounding kits mentioned above you may want to order an extra one or two... This appears more on low band than on high band or UHF, but it does happen. I've also seen it on another low band repeater when they added an amplifer and the transmitter RF coupled down the receive feedline shell and overloaded the receiver. On ANY decent system (no matter if it's low band, high band, UHF or 900 MHz) you need 100% shielding on your feedline(s), and you really want to keep the attenuation in the receiver feedline under 1-2 dB from antenna to receiver front end. If the noise floor at the site is low enough, a tower-top RF preamplifier may be appropriate, but tower-top preamps are another can of worms. And if they overload, they can create more problems than they solve.
Sometimes it's just outside of your control. For example, the signal strength difference due to trees having leaves (middle of summer) and having none (middle of winter) can be as little as 2 dB or as much as 10 dB on UHF and even higher on 900 or 1200 MHz. And that's going both ways - the UHF users have 10 dB less arriving at the repeater receiver, and have 10 dB less arriving at the users mobile or handheld receiver. The foliage absorption factor was a big surprise to public safety agencies that were conned into switching to 800 MHz or 900 MHz from 150 MHz or from 30-50 MHz... And once the users signal into the receiver is degraded by that 10 dB (or more) all the preamps in the world won't restore the reduced signal to noise level (a preamplifier just amplifies what it is fed - both the signal and the noise). It's all about noise figure and noise temperature and you really need a working grasp of the concept...
Some systems have adequate duplexers until they add a preamp to the system receiver. The problem turns out to be that they put the system in several years ago, with an adequate duplexer that had several dB of extra isolation ("headroom"). Then more radios were added to the site. Then they added the preamp - and the amount of isolation isn't enough. They forgot that you need the same amount of extra isolation for the amount of gain the preamp provides, since it raises the apparent noise floor as well as the signal of interest. In most cases you will have to fight with desense when you add a preamp. A top-quality preamp with a huge dynamic range like an AngleLinear will help.
Some repeaters get away without using a duplexer at all - one excellently performing local UHF system I have used and worked on simply uses two large (8" diameter) pass cans - one on the transmitter, the other on the receiver, with heliax feedlines and two antennas about 100 feet apart vertically (the shared UHF receive antenna is on the top of the tower at 160-170 feet, the transmit antenna is at 60 feet). The repeater itself is a 100w UHF MASTR II.
Or you can go split site (the design of the MASTR II makes that very, very easy)... One six meter system near here uses a 200-300 milliwatt 420 MHz "cross-link" (with home-made 6 dB beam antennas on each end) between the two sites (which are about a mile apart). They took a 420 MHz MASTR II mobile radio with a dead power amp (really cheap on eBay) and jumpered the exciter to the transmitter output filter. They then swapped the 6m receiver into it (making a 6m to low-power 420 MHz repeater), then put the 420 MHz receiver board into the 6 meter base station radio, then duplexed both of them. Note that to do it right a split site repeater needs two controllers since it is actually two repeaters back-to-back. The receiver-to-link site has the primary contoller, the link-to-transmitter site has the secondary controller. Split site repeatrs are not limited to 10 meters and 6 meters, one local 2m system started out as a split site, with about 300 feet between sites, and after the club gained enough members they did a funds drive and bought a real duplexer, feedline and antenna. Once they were installed the repeater became a single site, and the "real" antenna gave them much better coverage than the same site had provided before. Later on the old second site became a diversity receiver site - when it came online the received mobile flutter dropped by 50-70%.
In summary, isolation is acquired from multiple sources, all of which add up - a tuned front end instead of a lumped LC front end, cavities on the receiver or transmitter (or both), a 6-cavity duplexer instead of a 4-cavity, and even proper antenna placement on the tower (careful positioning can get you an extra 10-15 dB for free - just put the receive antenna in a voltage null of the transmit antenna - and nulls exist both vertically and horizontally). Each piece of the puzzle contributes some isolation by itself and when added together makes up the total. When the total exceeds what you need you have a properly performing (and therefore useable and useful) system. Many people think that the more duplexer isolation you have the better. And that's true - up to a point.
The web page Vertical and Horizontal Antenna Separation Charts provided by Kevin Custer W3KKC may be of interest.
The author can be contacted at: his-callsign // at // repeater-builder // dot // com.
Back to the top of the page
Up one level
Back to Home
This web page originally created and posted June 2000.
Text, hand-coded HTML and layout © Copyright 2003 and date of last update by Mike Morris WA6ILQ
This web page, this web site, the information presented in and on its pages and in these modifications and conversions is © Copyrighted 1995 and (date of last update) by Kevin Custer W3KKC and multiple originating authors. All Rights Reserved, including that of paper and web publication elsewhere.