Land Mobile Radio, or LMR (the FCC's name for commercial 2-way radio) has been around since the late 1930s. Even back then there were problems with channel sharing. Motorola came up with a way to get more than one Land Mobile user on the same frequency - they determined that different customers could coexist on the same frequency (called co-channel users) if they did not have to listen to each other. In the early 1950s the geniuses in Moto R&D developed a system that added a low frequency tone to the existing transmitter audio that was present as long as the transmitter was keyed. The receiver was modified by adding a tone detector circuit to the receiver and then the receiver squelch was modified so that it wouldn’t open unless the tone was there. They also modified the receiver audio circuits to strip the audio tone out of the speaker audio. The engineering folks patented the idea and the technique then the legal folks trademarked the term "Private Line". Sales literature of the time compared the system to the receiver having a front door lock on it, and the transmitter having a key ring full of keys, and only the signal with the correct "key" could open the lock. Note that all the system consisted of was additional tone modulation and tone decoding - nothing that would actually reduce or eliminate RF interference - more on this in the "CTCSS Doesn't Fix Anything" article.

Other manufacturers, finding that the system was absolutely necessary to stay competitive, copied the technology but couldn't use a trademarked term, so they came up with "Channel Guard" or "CG" from GE, "Quiet Channel" or "QC" from RCA, "Call Guard" from E. F. Johnson, "Quiet Tone" from Kenwood, and "Electronic Tone Squelch" or "ETS" from Canadian Marconi Company. Other manufacturers had / have other terms. Over the years, the Motorola trademark "Private Line" (or "PL") has become a generic term despite the best efforts of their marketing and legal staffs.

Once the other manufacturers came out with their systems, the EIA decided to write a standard for it (RS-220 that defines the tones, tolerance, levels, and more). They needed a generic name and came up with "Continuous Tone Coded Squelch System" or "CTCSS". If you think that's a mouthful, be glad that the second choice wasn't selected: "Continuous Subaudible Tone Coded Squelch System" or "CSTCSS". Later on Moto engineering developed a digital bit stream based system and they called it "Digital Private Line", or DPL. The generic was called "Continuous Digital Coded Squelch System" or CDCSS. Many people have shortened that to "Digital Code Squelch", "Digital Coded Squelch", or "DCS" (which is a term that Yaesu/Vertex uses in their sales literature). More on digital squelch later. By the way, when "DPL" became common usage many people started using the term "TPL" to precisely specify a tone PL system.

The Technical Side:

The Private Line (PL) system was first introduced in Motorola tube based radios in the early 1950s, was copied by GE, RCA, Link, Johnson, Kaar, and others. The generic name CTCSS and the EIA RS-220 standard followed much, much later.

The overall system is designed around a specific set of low frequency tones ranging from about 65 Hz to about 250 Hz. The oldest list that I am aware of (from November 1952) is ten tones: 100.0 cps, 110.9 cps, 123.0 cps, 136.5 cps, 151.4 cps, 167.9 cps, 186.2 cps, 206.5 cps, 229.1 cps and 254.1 cps, identical to tones 1Z through 0Z in the Motorola standard tone list (and just for information, the confusing tone descriptor "0Z" (zero-Z) came to be when the descriptor "10Z" got truncated to 2 digits). The list of 10 does, however have a note that the highest three are not recommended due to inadequate filtering of the receiver audio. Over the years the available tone list has been expanded - by 1965 Motorola was using 26 tones from 82.5 to 192.8 Hz, and by 1983 the MSF5000 station offered 42 tones in its list. These days, and depending on which industry "standard" set you chose to use, there are 32, 37, 38, 41, 42, 47 or 50 tones available, and the U. S. Military has their own unique tone of 150.0 Hz that doesn't appear on any list of standard Land Mobile tones.

No matter what tone you use, the deviation is typically 500 Hz on a 5 kHz channel, and on a 2.5 kHz channel (common on 900 MHz) it's half that. The EIA/TIA-603 international standard for Land Mobile Radio performance calls for 500Hz of tone deviation on a 5kHz deviation channel, but some radios run as high as 800Hz. I've seen public safety grade commercial radios that reliably decoded only 100 Hz of PL deviation, and the audio filters were so good that the hum was still inaudible at 1 kHz PL deviation.

For fast operation the encode tones must be a clean sine wave and the frequency tolerance is very tight. ANY frequency error in the tone encoder or decoder, or distortion in the tone encoder output will slow the decoder down, and can sometimes prevent it from decoding at all (and thereby causing the receiver to never open the squelch).

CTCSS (no matter what the manufacturer / trademarked name) is a continuous and subaudible tone. In fact, "sub" means "below", as in this case "lower than 300 Hz". That doesn't mean that it's INaudible, meaning "not hearable". That's a big difference, and the tone was audible in cheaper receivers that didn't have a good high pass audio filter. Users of receivers with poorly designed (or otherwise inadequate) audio filters frequently mistake the tone for a power supply hum. When the system is properly implemented it is almost impossible to hear with the un-aided ear.

The circuitry used in the CTCSS system, when you actually sat down and analyzed it was pretty simple and was made up of four, sometimes five, separate sections:

1. A very clean and frequency-precise audio tone was generated and was mixed at a low level into the transmitter audio. The modulator was refined to minimize the distortion on the tone. Most early designs had two separate modulation level adjustments, one for the voice deviation pot and the second for the tone deviation. To minimize any distortion on the tone some radios actually implemented a separate modulator optimized just for the tone encoder.
2. A tone decoder was connected to the receiver demodulator. The demodulator components were refined to minimize any added distortion to the tone.
3. The receiver audio mute section (normally driven by the squelch circuit) was modified so that a toggle switch selected the mode - in one position the audio mute circuit (switching the speaker on or off) was driven by the presence or absence of the incoming carrier, the other way it was driven by the presence or absence of the tone. In most cases the tone decoder just forced the audio mute section open (which in some cases resulted in a receiver that was more sensitive to tone-modulated carriers than to plain carriers). Later designs offered a mode where both the carrier AND the tone-present signal from the decoder was required to unmute the audio. This is called "AND squelch" and there are a couple of articles on it at this site. Changing the tone decode mode from tone-only squelch to "AND" squelch was done by jumpers (or sometimes by component changes) inside the radio.
4. A high-pass audio filter was added to the receiver audio path so that only the audio above the tone range was passed on to the audio amplifier and the speaker. This got rid of the tone hum under the incoming voice.
5. Better radio designs added a high-pass filter to the microphone audio circuits to keep low-frequency speech artifacts out of the tone region in the transmitter audio. Without this filter it was possible to "talk off" the transmitted signal with loud modulation as it would distort the encode tone waveform, and the decoder on the other end would assume the tone had gone away and mute the received audio. See the "Talk Off" section below.

For years the component that determined the tone was a single plug-in unit. Motorola, GE and RCA used plug in vibrating reeds, E. F. Johnson and others used plug in electronic modules. Since the average radio only had one tone socket, changing the tone in any given radio was as simple as unplugging and plugging in a reed or a module (or using a relay (or some relays) to switch between two (or more) plug-in reeds or modules). This held true up to the days of programmable radios.

Radios in the late 1950s and throughout the 1960s usually could be had in up to four RF frequencies, but 99% of the time only one tone frequency, if the radio had tone at all. As such it was not unusual to have most agencies in an area (especially in rural areas) to use a common tone throughout the area. When the California Highway Patrol wrote the RFQ (Request for Quote) specifications for a new statewide fleet of radios in 1966 they included dual PL tones (both encode and decode). Motorola took an existing 4-frequency design (the LLT series Motran) and modified it to meet the specs - the selection of encode and decode tones was slaved to the RF frequency selector switch by way of a diode matrix. As far as I can tell this was the first factory radio that was capable of more then one tone.

Early Technology:

When it was invented in the early 1950s the tone was generated and detected with the only technology of the day that was frequency stable across a wide temperature range and in a mobile (high vibration) environment: vibrating reeds - one-legged tuning forks in a protective housing. The first implementation in the Motorola tube based base and mobile radios had large copper-encased encoder (the "Vibrasender") and decode (the "Vibrasponder") reeds.

From the 1952 Motorola Manual:
The Vibrasender is referenced as a P-8511 and is "packaged as a small plug-in unit, approximately 1" wide x 3 3/8" long x 2" high."

...and...

"The P8611 Vibrasender is similar to the P-8511 unit except for the internal mounting of the tuning fork. The P-8611 is packaged as a small plug-in unit, approximately 1" deep x 2" wide x 3-3/4" high."
The Vibrasponder is referenced in the manual as a P-7810.

The Vibrasponders had four pins: two went to the drive coil and two went to a pair of internal contacts that closed and opened with each cycle when the reed vibrated at the proper frequency - if you had a 100 Hz reed you had 200 pulses per second from the contacts. An integrator circuit filtered this pulse train contact closure and the developed voltage turned on an audio gating circuit that allowed the audio to pass to the speaker.

A side view of the TU333 Vibrasponder
Except for the Vibrasender name stamped in the side of the unit, the TU217 Vibrasender was packaged identically.

The package is 1/2 inch high, 7/8 inch wide, 3-1/8 inches of body length, 3-1/2 inches including the pins. One common name for the reed package was the "copper banana". Some paging tone reeds were made in a package that was the same height and width, but about 2/3 the length.

The Vibrasender encode reed used the same housing and connector. It had two pins for the drive coil and an internal jumper from the third pin to the fourth pin. When plugged into the circuit, the jumper would activate the oscillator and the coil would be energized such that it resonated and developed the desired tone.

You couldn't swap the copper Sponders and Senders in a stock circuit due to the second pair of pins being shorted in the Sender but not in the Sponder. You could put a short across the encoder enable pins on the socket and use a Sponder in a Sender socket (and on my first PL mobile that's just what I did until I found a 127.3 Hz Sender). Using a Sender in a receiver would put the jumper in place of the contacts, and cause the squelch to be open all the time, thinking it was receiving a PL signal. So you were screwed if you swapped a Sender into a Sponder socket, but it was a fun trick to pull on the new shop tech. My first multi-tone mobile was made by slaving a relay to the frequency select line in the radio and using the contacts to select one or the other Vibrasender coil. I had 131.8 Hz on one RF frequency and 127.3 Hz on a second frequency. Later on a second relay allowed 100 Hz on a third frequency.

The older copper reeds had different vibrating characteristics, i.e. a "Sender" without the contacts produced a smoother output waveform and took longer to start or stop vibrating. The "Sponder" didn't care about the waveform and it was often overdriven just so the contacts would make a good connection as quickly as possible. They also were weighted such that the reverse-burst would stop them without causing them to start oscillating with the reverse phase (before the transmitter shut off). More on that later.

The big copper reeds worked and worked pretty well, but they had their problems. One was that on a rough or washboard road you could get a false decode if the contacts in the Sponder closed at a repetitive rate. Another was that if the audio coupling cap was leaky the DC could "burn" the fragile contacts and kill the functionality of the decode reed. And the radios were getting smaller, and the copper reeds were just too big.

Enter Bramco Corporation of Piqua Ohio, and their contactless mini-reed.

Motorola needed to add PL to the design of their handhelds and had a problem... there was no room for the reed. So they went to the reed experts. Bramco Corporation took the "copper banana" and shrunk it (to the size of the KLN6210 reed in the photo below), and eliminated the contacts. Internally the Bramco reed was a tiny one-legged tuning fork with a magnet mounted crosswise in the end of the tine. The magnet moved inside two tiny coils and formed a very precise and very stable frequency resonant transformer. A clever circuit design allowed the single reed to work as both an encode and as a decode reed. Both Motorola and GE marketed the Bramco design for years and years under a wide variety of model numbers. Bramco was eventually sold and became the Bramco Controls Division of Ledex Inc. Bramco published several books on tone control, their 1966 book, "Control Techniques with Resonant Reed Relays" was a bible of the radio controlled model community. I had a copy, loaned it out, and the borrower evaporated along with my book.

A side view of the Bramco-manufactured mini-reeds. The top reed is a Micor decode reed, the lower reed is a standard sized mini-reed.
The TLN8381 Micor decode reed was the only one to be made in the oversize package.

A top view of the small reeds. The left reed is a standard sized mini-reed, the right reed is a Micor decode reed.

The Bramco mini-reed specifications were:
  Frequency Range: 66 to 3000 Hz (CTCSS used 67-251 Hz, the higher frequencies were used as paging tones)
    Temperature Stability: Frequency varies less than 0.002% per degree C
    Impedance: approximately 1000 ohms at resonance (input and output)
        (but different model numbers used impedances from 500 ohms to 5 K ohms).
    Drive Level: 66-300 Hz - 0.05 to 0.2 VRMS
    Drive Level: 300-3000 Hz - 0.10 to 0.3 VRMS
    Encoder Freq. Tolerance +/- 0.3% (- 40 to 85 degrees C)
    Dimensions (Height, Width, Length of body, Length with pins) in inches:
        Standard small reed (the KLN6210A in the photo): 3/8, 5/8, 1-1/8, 1-1/4
        The large TLN8381A decode reed package: 1/2, 3/4, 1-1/8, 1-1/4

Many commercial 2-way radio systems needed "split tones" - where the encode tone was different than the decode tone, so the mini-reed was made in both Sender and Sponder designs. Split tones were difficult in hand-helds due to size and space, but Bramco had an answer to that as well... a half-size reed.

I don't have the model number, or a photo, but I've seen a "sugar cube" reed (yes, that is what it was called) - it was half the length of the KLN6210 above. The first time I saw them was in a special products PL board made for a handheld. The board held two reeds, and it allowed either two different tones on the six frequencies, selectable by a diode matrix, or different encode/decode tones. The sugar cube reed worked on a different principle: instead of a one-legged tuning fork with a magnet in the end, it used a taut band with an S-shaped weight mounted at right angles in the middle of the band, with tiny magnets at each end of the S-shaped weight, and coils surrounding the magnets. As the coils were driven the S-shaped piece twisted the taut band back and forth and that action formed the resonant transformer.

The use of plug-in reeds (large and small) continued through the 1950s, the 1960s 1970s and into the 1980s: the Motrac (hybrid tube and transistor), Motran (all transistor), Micor (all transistor), Mitrek, and other product lines of the period. At some point the production was switched from a metal case for the reed to a plastic case to save weight and lower the cost. In addition many specially packaged reeds were developed to fit them into unique spaces in specific radios (especially handhelds). Some reeds were wide and very thin, or very long and short, or other weird shaped.

Roll forward to the current era and the ham radio manufacturers seem to have settled on "tone" as the term to use in the operators manual to describe the encode side and "tone squelch" for the decode side. Most of the amateur VHF and UHF equipment manufactured since 1990 has at least encode capability (standard or optional) and many have decode capability (standard or optional). If it isn't built in, it is simply a plug in optional circuit board. Most radios oriented towards amateur radio do not offer split tones. Aftermarket encoders and encoder/decoders (such as those from Communications Specialists, better known as Com-Spec) can be added to almost any radio, as they are smaller than a large format postage stamp. And it's not that hard to build your own tone encoder.

Applications of CTCSS:

The initial design specs at Moto was to enable multiple user groups share a simplex dispatch channel - maybe have the county dogcatcher share with the county road crews and the school busses. But the first few years of reality outpaced anything that was envisioned at the designer's workbench.

In the commercial two-way radio world, there is what is called a "community repeater". This is a regular repeater with a "tone panel" attached. Picture a number of tone decoders on the receiver, each linked to a separate tone encoder on the transmitter. User group A (perhaps a diaper service) uses 67 Hz, User group B (perhaps a landscaping service) uses 82.5 Hz, user group C (perhaps a drain cleaning service) uses 94.7 Hz, user group "D" (the roofing company) uses 103.5 Hz, and so forth. Note that in the example above there are only four co-channel user groups. In dense two-way radio environments the number of groups can be higher. As long as nobody "hogs" the channel and each waits their turn, each user group can use the same repeater without hearing the transmissions of other groups (but a receiver without a tone decoder, or a disabled tone decoder, would hear all transmissions from all groups). The tone panels also remove (filter out of the receiver audio) the incoming tone and replace it with an internally generated tone (into the transmitter) so that the outgoing signal has a clean tone. Communications Specialists (Com-Spec), Connect Systems and Zetron are just three manufacturers of tone panels, there are and have been others.

The use of a different CTCSS tone for each group of users on a channel allows several different groups of users to be on one frequency without hearing each other. They will, however, cause interference to each other if they just key down and talk. Another use of the tone is to allow several different repeaters to share the same RF frequency - just use different tone frequencies. And there are tricks that can be used there - I have seen two community repeaters on the same RF channel with overlapping coverage areas, and one user group on both repeaters. The radios have channel 1 and hcannel 2 set up with the same RF frequencies and different encode tones and the same decode tone. They might encode, for example, 127.3 on channel 1 and 131.8 on channel 2. and decode 131.8 on both. Repeater one is on the south end of town and listens for 127.3 and talks with 131.8, and repeater two is on the north end and listens and talks on 131.8.... If the dispatcher needs to find a user he/she can find them no matter where they are...

In a commercial mobile installation, the mobile microphone hanger is wired into the tone decoder and when the microphone is hung up, the decoder is turned on, thus muting the receiver. When the user picks up the microphone, the decoder is disabled, the receiver goes into carrier squelch mode and the user hears any activity on the channel. If nothing is heard, the call is made and every radio using that tone unsquelches. If, on the other hand, the channel is in use, they are supposed to monitor until the traffic clears and then make their call. Base station microphones have a "monitor" button next to the PTT button to disable the decoder (and most are interlocked so that you can't PTT without holding down the monitor button), forcing the operator to check for traffic first. Commercial handheld radios have a monitor button, usually located near the PTT bar for the same purpose. Some radios have a feature called "Busy Channel Lockout", which will not allow the user to transmit as long as the radio is receiving another signal.

The same microphone clip trick is still used today, but some other techniques are used to make it simpler - see this article on hangup methods. Police dispatch systems mix patrol, vice, detectives, and more all on the same channel. Rather than have everyone listening to everyone, the each group of mobile users hides behind the tone decoder and does not hear anyone else unless they lift their microphone out of the clip. Then they wait for an open channel.

Modern amateur radios do not have this microphone hanger feature since the CTCSS system is used to allow users to restrict what they want to listen to, not to allow several fleets of radios to operate on the same frequency.

Some radios may have a "monitor" button, which will momentarily open the squelch. Other amateur mobile users have to manually turn off the tone to monitor the frequency in the carrier squelch mode. Some folks program adjacent memory positions (channels) so that bumping the channel button up one position switches to carrier squelch mode and bumping and bumping the channel button down goes back to CTCSS decode mode.

From an email to repeater-builder:
One cute trick that has been used since the 1970s was / is called "PL Paging" (it's called that since it has been around since long before the term CTCSS was established). Picture a repeater that may or may not require a tone to key it up, but the system transmitter does not use a tone encoder. The "PL Paging" method adds a tone encoder to the system transmitter, but it is normally off. The repeater system operates as usual. All that is needed is a DTMF decoder than can pull in a relay, and that functionality is in almost any modern repeater controller. Simply set it up so that when a user sends the proper DTMF digit (or sequence of digits) a relay closes and enables the system transmitter PL encoder - but that relay drops out when the carrier on the input goes away.

Note that Ralph does not have a DTMF decoder - he has a plain-jane radio that happens to have a tone decoder feature.

The text above mentions a DTMF decoder pulling in a relay, because that's how the first one was implemented - a Speedcall brand DTMF horn-honker decoder (that had been retired from IMTS service) drove one coil of a two-coil latching relay. The COR pulsed the other coil. The contacts operated the tone encoder. These days it can all be implemented in the repeater controller - and while any modern controller can do PL Paging, the Scom 7K (no longer in production) actually has a tone encoder audio gate built into it and you'd think it was designed just for the job.

From another email to repeater-builder:
...most commercial and public-safety repeaters and base stations automatically strip (or just don't encode) the tone whenever the station ID is transmitted. The users of such systems would quickly find a Morse code ID every N number of minutes to be very annoying. All of my repeaters (no matter what service) strip the encode tone during ID, and nobody ever hears an ID. Many of the Hams who use my repeaters have commercial-quality radios with full-time PL decode, and they are very happy to not hear the ID or any squelch crashes.

A Few Notes on Tones and Tone Selection:

• Many manufacturers use a number to select the tone in their radio, for example, tone number 12 might be 100 Hz. If you are setting up radios from multiple manufacturers on a common channel (like for a neighborhood CERT team) do not use the tone numbers from the radio (like "use channel 14 and tone 10"). Use the tone frequency instead. Why? The problem is that what is tone 12 on one brand might be tone 13 on another, or tone 43 on a third, or tone 100.0 on a fourth and sometimes it even changes from model to model within the same manufacturer. Because of this you ALWAYS SPECIFY THE TONE FREQUENCY in any list or documentation - let the end user be responsible for keeping a cheat sheet of the tone table from the manufacturers literature for his or her own radio with them (maybe a laminated copy of chart from the critical manual page, and in the wallet?).
  Look at this page for some examples of non-compatible tone numbers (there are six different tone #38s just on that page). Even the FRS manufacturers can't get it right, and while the information on this spreadsheet page is dated (several models mentioned are no longer produced or sold), it does makes the point... and if anyone has additional radios to add in additional columns, please send the information to the author.
  
  
• Many older books recommend using every other tone. This was especially true in the 1950s and 1960s due to three reasons - the close spacing of the tones, tone encoder start-up drift and the early decoders had a wider bandwidth and would false on adjacent tones. Many cheaper radios still have that falsing problem today, mainly because the lower frequency tones are spaced closer together (percentage) than the higher tones.
  
  
• Many phase-modulated transmitters have trouble sending the tone at the proper level across the entire tone range. The lower frequency tones are often lower in level as the modulator components are selected to prevent the higher tones from being over-deviated. The lower frequency tones end up being under-deviated and hence difficult to decode reliably at the other end. True FM transmitters don't have this problem.
  
  
• Lower frequencies are easier to filter out of the audio stream.
  
  
• Higher frequencies decode faster. Why? Basically the decoder, be it a vibrating reed or a microprocessor, needs some minimum number of cycles of tone to assure accurate decoding. If, say, it requires 20 cycles, then a 100 Hz tone could be decoded in 200 mSec, while a 200 Hz tone could be decoded in 100 mSec. EIA/TIA-603-C states that CTCSS decoder response times may vary between 224 milliseconds at 67.0 Hz to no more than 150 milliseconds at 100.0 Hz and above. In reality, many real hardware decoders are very close to 200 milliseconds at 67.0 Hz and as little as 80 milliseconds at 250.3 Hz.
  
  
• Another reason to avoid low frequency tones is if you are using a toned repeater and tone decoders in the mobiles. Think about it - the user presses the PTT button and as much as 2/10 of a second later the repeater receiver tone decoder activates. It may take 1/10 of a second for the repeater transmitter to comes up to full power. It takes as much as 2/10 of a second for the users receivers to unsquelch. So 4//10 of the 5/10 of a second delay is due to the chained decoders - this can cause clipping of the first word of each transmission until the users train themselves to press the button, WAIT a half-second, then talk. It gets progressively worse if you have toned links between repeaters creating longer and longer decoder chains.
  
  
• 100 Hz is the most popular tone in the USA, 103.5 Hz the second most, and a tossup between 107.2 Hz, 110.9 Hz 114.8 Hz, 127.3 Hz and 131.8 Hz the third.
  
  
• In areas with 60 Hz mains power you really want to move 179.9 Hz to the bottom of your available tone list, and both 118.8 Hz and 123.0 Hz your second last choice. Why? 179.9 Hz is 1/10 Hz off the third harmonic of the power line and while it isn't the problem that it was in the 1960s and 1970s (due to better designed decoders) but even so, a user with a hummy power supply can false a 179.9 Hz decoder. The other two tones, 118.8 Hz and 123.0 Hz are far enough away from the second harmonic that they don't cause as many problems (it takes a really poor quality decoder and a really distorted encoder), but if you have a choice, why ask for problems? Just put all three at the bottom of your list of choices.
  
  
• Don't use 100 Hz in areas that use a mains power frequency of 50 Hz. In fact, the British OFCOM (Office of Communications, their version of the USA FCC) deliberately skips over 100 Hz in their list of standard tones (part of the UK MPT1306 Specification which applies to business radio and amateur repeaters).
  
  
• If you have a collection of radios made by a variety of manufacturers on your system then you want to avoid tones above 203.5 Hz as many radios don't have them.
  
  
• Likewise some lower pitched voices on radios with inadequate transmit audio filters will false a decoder above 200 Hz.
  
  
• Don't use 131.8 Hz or 136.5 Hz on a mixed mode channel (both CTCSS and CDCSS). More on that later in the CDCSS section below.
  
  
• The "gold standard" in PL decoders remains the Motorola Micor as far as its ability to decode a tone out of on-channel noise.
  
  

Summary:   The above concerns causes some system engineers to restrict the tone selection to the 10 tones from 127.3 Hz to 173.8 Hz - as this range balances fast decoding with keeping the tones out of the audible part of the receive audio. If CDCSS will ever be used on the channel then that lowers the selection to 8 tones as that eliminates both 131.8 Hz and 136.5 Hz. If your system users have radios with better quality transmit audio filters (to keep the speech out of the encoder) and receiver audio filters (to keep the hum out of the speaker) then you can skip over 179.9 Hz and raise the upper end to 186.2 Hz or even 192.8 Hz.

Political Considerations, Open, Closed, and Private Systems:

To quote Bill Pasternak WA6ITF in the 1980s book on repeaters, "PL does NOT make a repeater closed or private!"

To quote another friend of mine, "'Private Line' has nothing to do with operational status of a system, being open or closed. PL does not mean 'please leave'".

Thirty to forty years ago adding a tone encoder to a radio was a technical challenge - making it frequency and amplitude stable across the peak of summer to the depth of winter temperatures, making it fit into the leftover space in the handheld or mobile, and making it modulate cleanly. Today, it's a menu selection.

The classification of an amateur repeater (in any list or directory) is purely the system owner's choice. In many areas there are three choices: open, closed (i.e. club), and private. Open is just that - any licensed ham can use it, closed requires you to join the group and pay dues (but membership is open to anyone with a ham license and a fat wallet), and private allows the system owner to pick and chose who uses the system. Some areas have additional, sometimes informal, categories: "Open system but private remote base and autopatch" is one, "Private but Friendly" is another. I know a few system owners that run their repeater(s) as Open, but list their systems in the directory as Private just to preserve their legal option of de-inviting any troublemakers.

However, as you can see from the above, CTCSS is purely a technical system; the repeater access classification depends on the wishes of the system owner (or trustee) and the co-operation of the local coordinating council. An open repeater can use carrier squelch or CTCSS decode, a closed repeater can use carrier squelch or CTCSS decode, a private repeater can use carrier squelch or CTCSS decode. In all cases, it's politics, not technology. Setting a tone encoder to the proper frequency is a minor inconvenience when you consider how many potential problems it can eliminate. Most coordinating groups strongly recommend the use of tone squelch on repeater receivers. Some areas mandate tone squelch as part of an individual coordination grant due to such situations as proximity to a co-channel repeater, or in an area where band openings frequently aggravate co-channel interference problems.

Some open repeaters use a CTCSS decoder to prevent key-ups by users of other repeaters on the same channel. Many of these announce the tone frequency on the voice IDer (You might hear "Monitoring one-forty-six-point-two-two with eighty-two-point-five Hertz" on a 146.82 MHz repeater), as well as list the tone frequency in the Repeater Directory. Traveling hams really appreciate the voice announcement, especially on an odd-split repeater. Some closed repeaters use CTCSS or the digital version to limit access to only the club members. Some private systems do the same. Some private systems run a strange offset in addition to tone or digital squelch.

Some folks try to use weird nonstandard tones to resolve people problems. Sure, the receiver won't unsquelch if it is requiring a weird tone and the individual only has the standard tones, but he can still sit there and transmit on the channel causing a heterodyne or capturing / blocking the receiver, so what real security (or increase in usability) have you really gained? And what prevents a determined and resourceful person from listening to your input frequency with an OptoElectronics tone grabber (or another similar piece of equipment) and finding your weird magic tone (or tone sequence)? It takes maybe 3 to 5 seconds of signal, and then it's no longer a secret. And the Tone Grabber unit is simply a portable luxury as a carrier squelch receiver plugged into the computer sound card and a "waterfall display" program is all they need.

In some geographic areas specific tones (such as 100.0 Hz) are indicators of open repeaters. No, we are not saying that 100 Hz is an open tone in YOUR area, you need to check the open repeater list published by your local coordination group (or the ARRL repeater directory - the League gets the information it publishes from the local coordinators).

Just remember, the requirement or use of a subaudible tone to access a repeater does NOT mean it is not an open repeater. The addition of a tone decoder to a repeater receiver does NOT change the status of a system to closed or private.

"Talk Off" and other problems:

Some radios have very poor CTCSS performance and have distorted encoders, slow-to-open or slow-to-close decoders, audio filters that don't remove the higher tones, audio filters that add distortion to the receiver audio, etc. Sometimes you can abandon the internal CTCSS sections (encoder and decoder) and use a Com-Spec, other times you just have to use a different radio.

As an example of extreme problems, some Alinco models, including the DR-135, DR-235 and DR-435, are (in)famous for having the speaker continue to talk for as much as two full seconds after the incoming tone goes away (one eHam review says FOUR seconds). Or how about the Yaesu VX-1 - in the tone squelch mode once the tone decoder opens the squelch it stays open - forever - until the carrier squelch closes.

Another example, from an email to repeater-builder:
One manufacturers current batch of two-band and dual-band radios are also notorious for having CTCSS encode problems. It's because of several outright design screwups:

1. The designer(s) strive to get very flat, wide-band microphone audio through the entire transmit audio system, starting with the microphone element itself. They seem to think that broadcast quality audio should come out of their equipment. I don't agree. Time after time it's been proven that the communications range is from (about) 300Hz to (about) 4000 Hz!
2. The designer(s) introduce the CTCSS tone at a fixed (non-adjustable) level before the deviation control.
3. There is NO high-pass filtering of the mike audio before mixing with the CTCSS tone, so all that low-frequency mike audio mixes and interferes with the CTCSS tone.

Or how about this problem, interestingly enough with the same brand of radio mentioned above:
A friend had a vehicle with rather aggressive tread tires. Whenever he drove at a certain speed, or over a metal-grate bridge, the growl from those tires, and vibration within the vehicle itself, would be picked up by the microphone and severely interfere with the radio's internally-generated CTCSS tone. This audio mix would cause the repeater receiver to close the squelch on a perfectly quiet RF signal just because the CTCSS decoder got confused by the impure signal and the interference in the sub-audible tone band. Apparently his radio had NO sub-audible band filtering in the microphone circuit, it seemed like it passed anything that came into the microphone onto the modulator.

Sometimes problems in one part of the radio show up as tone problems. A friends old Ford F250 had a strange problem with a Kenwood 721... The radio had very poor DC power filtering, and coupled with one bad diode in the alternator resulted in a weird symptom... certain engine speeds would generate alternator whine that would mix with the tone encoder and cause repeater dropouts. The whine frequency never dropped as low as 100 Hz, but there were some "magic" frequencies that would mix... You could hear him talking on the repeater with a full quieting signal, and as he decelerated and braked to a stop the whine pitch would drop and at two or three points of whine frequency you'd hear the repeater receiver lose CTCSS decode and a few moments later regain it. The problem was solved when the F250 was sold and went to Montana, and the radio was given to a new ham (that used it as a base station).

Other radios have another problem due to bad design, as this email to repeater-builder shows:
...Can anyone suggest ways to reduce CTCSS dropout when someone is talking loud on a radio? The user radios are set up for 600 HZ of CTCSS encode and a 1khz tone gives 3kHz modulation, but I have one user that talks louder than the other users and he seems to breakup all the time where other users seem to have no problems. Could the problem be more tied up with the repeater receiver?

The reply was:
...I suspect that the problem is "talk off" caused by overdeviation of the mike audio. When the CTCSS tone is mixed with the mike audio prior to the deviation limiter stage, an excessive voice signal will be hard limited along with the CTCSS tone, causing a distorted signal. This problem occurs quite often in amateur-grade equipment, since the makers often have very "hot" tone deviation that is set with a fixed resistor rather than adjustable as in most commercial-grade radios. Alinco, Icom, and Yaesu handhelds seem to have this problem all the time. Every Alinco handheld I've tested has a PL deviation above 1 kHz, and a few samples were up to 1.5 kHz- about three times the necessary amount. Needless to say, a tone deviation of 1.5 kHz doesn't leave much room for voice deviation, and all it takes is a loud talker to cause the limiter to "squash" the modulating signal and distort the CTCSS tone enough for the repeater to drop. It's not easy, but I have padded the CTCSS tone circuit in several Alinco and Yaesu radios to reduce the level to less than 500 Hz deviation, and that cures the problem. Most modern repeaters can detect a tone as low as 100 Hz deviation, so setting the tone between 400 and 500 Hz should work fine.

Of course, there are "diddle-stick artists" who can't resist cranking up the mike gain pot so that they will "sound better" (they think) and make the problem worse. In order to stay within a 16K0F3E emission envelope, the total deviation should be limited to around 4.8 kHz. Most commercial-grade radios reduce the voice gain when a tone is added, but cheap radios aren't that smart. If the total deviation is limited to, say, 4.8 kHz, but a tone is not used, the voice deviation will sound softer. I suspect that the diddle-stick artist cranks up the gain to offset that low level.

I suggest that you set the transmitter CTCSS deviation to 500 Hz, and ensure that the deviation limiter is set no higher than 5.0 kHz or a hair less. See if that reduces or eliminates the talk-off problems. The radio service manual should have a procedure for mike gain and limiter setting, which should be followed. The fact that only some users have the problem strongly points to the user radios, rather than the repeater. Also, ensure that the problem radios are exactly on frequency - any distortion will only be made worse if not.

Reverse Burst™ "Chicken Burst", Squelch Tail Elimination, Squelch Tail Eliminator, or STE:

Some other common terms you might hear used in conjunction with CTCSS are "reverse burst", "squelch tail eliminator" or "squelch tail elimination", usually appreviated as "STE".

An annoying problem showed up not long after PL was introduced. In a carrier squelch environment the squelch circuitry unmutes the speaker as long as the on-channel noise is gone. In other words, the user unkeys, and the receiver mutes as soon as the squelch circuitry seens the on-channel white noise. The amount of time it takes for the squelch circuit to decide that the carrier is gone determines how long the extended white noise burst (the "squelch tail") is present.

In a CTCSS / PL environment the receiver squelch is held open as long as the received PL tone is there. The problem was that the squelch was being held open due to the fact that a carrier squelch circuit took less time to determine that the carrier was gone than the PL decoder took to decide that the tone was gone. This is due to the fact that the PL decoder circuit on the receiving end continued to give a valid signal until the receiving reed coasted to a stop. Using "AND squelch" as mentioned above helped, but it is not a 100% cure for this problem.

This "coasting" time lasts from 1/10 to 3/10 of a second (depends on the frequency) and is due to the "flywheel effect" that is present in any very sharply resonant audio filter (and the decoder reed is just that kind of filter).

In order to eliminate this squelch tail at the end of a transmission, the engineers developed (and legal both trademarked and patented) the "reverse burst" technique. The original design used an audio transformer on the encoder output, with the center tap of the secondary grounded. The two end points of the winding supplied the audio signal in phase and 180 degrees out of phase. The two contacts of a SPDT relay selected one phase or the other. When the user released the radio PTT button the relay immediately dropped out and the opposite phase from the transformer secondary instantly flipped the phase of the PL encoder audio tone by 180 degrees. About 150-200 milliseconds later a timed-release relay actually unkeyed the transmitter.

This "reverse-burst" circuitry created a burst of phase-reversed tone that on the receiving end caused the decoding reed to slam to a stop, and just about when it was ready to start vibrating again, the transmitter shut off (causing the input carrier to disappear). During that slowing-up time, the squelch circuit would close (shutting off the speaker), hence you'd never hear the noise burst when the transmitter shut off. Note that the entire reverse burst circuitry was part of the transmitter - the only critical part of the receiver was that it used a decoder that had the flywheel effect. Later decoder developments allowed a reedless design to support reverse burst.

Over the years Motorola used three different circuits to generate the phase shift used in the reverse-burst encoder. The first one, as mentioned above, used the center-tapped transformer and the delayed drop-out relay selected one phase or the other before the transmitter actually dropped off the air. A later circuit used the same relay but dropped the transformer - they picked up the signal from either the input or the output of an inverting amplifier stage (tube or transistor). The third dropped the amplifier stage and used an R-C phase shifter that took the tone signal as input and let the relay select the input or output of that shifter. The amount of phase shift was frequency-dependent, so I suspect they didn't use this circuit for very long.

The 180 degree shift was used initially because it was the simplest to generate - a transformer, an audio changeover relay and a timed-release relay - but it was abandoned, not because of how reeds respond, but because it is guaranteed to make the worst "pop" in the receiver audio when the tone phase change begins, especially if it happens to change at the peak amplitude of the sine wave. These days 180, 120 degrees and 240 degrees are the most often used. Industry Standard EIA/TIA-603C covers STE systems (including Reverse Burst) and describes two reverse-burst formats, 180 degrees and 240 (AKA 120) degrees, without recommending one over the other. It also specifies 250 ms as the maximum time for audio cutoff upon removal of the CTCSS tone without reverse burst STE, and 50 ms with reverse burst STE. Last I checked, Motorola is the only major manufacturer that uses 240-degree shift; Kenwood, Icom, Vertex, and most others still use 180-degree shift.

Here's a classic photo of a 120 degree phase shift. The top trace is the PTT signal releasing; the bottom trace is the encoder output. The 2/100 of a second time difference between them is the latency in the station control electronics.

Nowadays, most radios use reverse burst for eliminating squelch tails. Cheaper radios either do not have any kind of squelch tail eliminator or they just pinch off their PL encode tone about 300 milliseconds before transmitter shut-down (this was very popular in tube-era GE radios to avoid infringing on the still-active Motorola patent, and was jokingly called "chicken burst"). The GE Progress Line and Mastr-Pro series used the tone-pinch-off method; the later Mastr-II was designed after the reverse-burst patent ran out, and used phase-changing STE. Overall, the phase changing system is preferred; some of the tone decoders used in the cheaper radios can take as much as 7/10 of a second to decide that the tone is gone.

For those that have a non-reverse-burst radio, and want to add it, look at the data sheet and schematic for the RB-1 module on the Communications Specialists page at this web site. The unit is long out of production, but is easy to duplicate. All it takes is a quad op-amp and a handful of components. All Com-Spec products that generate reverse-burst use 180 degrees.

More details on this topic in the article titled "Explanation of Reverse Burst & "And Squelch" at this web site.

A final comment on reverse burst:
Reverse Burst was invented over 50 years ago. The patent has long run out. Kenwood, Icom, Yaesu, et al all have it in their commercial radios, yet the various radio manufacturers haven't put it in their amateur market products, even as a menu option (and since the tones are all CPU-generated, it wouldn't take ANY new hardware, just a few lines of programming code).
Why ?

Continuous Digital Coded Squelch System (CDCSS), Digital Private Line (DPL), Digital Channel Guard (DCG), etc.

The "Continuous Digital Coded Squelch System" (CDCSS), was the follow-on to CTCSS (but some books use "Code" instead of "Coded" in the name). Instead of a constant low frequency tone, CDCSS superimposes a continuous stream of 23-bit square wave digital data words on the transmitted signal. The system is referred to as Digital Private Line (or DPL) by Motorola, and likewise, General Electric's implementation of CDCSS is referred to a Digital Channel Guard (or DCG).

In the same way that a single CTCSS tone would be used on an entire group of radios, the same CDCSS code is used in a group of radios. The code that is actually transmitted is actually a Golay (23,12) code. The (23,12) notation tells us it is a 23-bit Golay word with 12 data bits included in the total 23 bits. There is lots of information on Golay codes on the web, so I will not go into those details here. The number of codes varies with the manufacturer and the average is about 100 codes, but generally only about 50 are used due to the "complement" or "inversion" problem. Nine of the 12 bits are programmable in the DCS environment, and they are written as three octal digits (Octal means that the legal values are from zero to seven). For more technical details see this article. There is another good article here.

CDCSS systems don't need any tone-pinch-off or reverse-burst / STE games, since the system was designed from the start with a "close the squelch now" message at the end of every transmission: a 1/5 second burst of 1010101... (i.e. square waves) at the main clock frequency (134.4 Hz). Every manufacturer whose radios have digital-coded squelch also have implemented this turn-off code. It's part of the EIA/TIA standard.

Some radio systems use both a CTCSS and a CDCSS decoder on the same receiver. They use (and publish) the CTCSS tone for user access and reserve the digital code for control operator functions. If you ever expect to use both tone and DCS on the same channel, then DON'T use 136.5 Hz or 131.8 Hz !

Why? Due to the well-known CDCSS "kerchunk" problem. The designers of DCS made a poor choice of data rate, especially of the DCS turn-off signal. When the user unkeys the CDCSS radio transmits the standard 1010101... turn-off code for about 200ms. This looks like a 2/10 of a second burst of 134.4 Hz square wave on a scope and the frequency is close enough to 136.5 Hz and 131.8 Hz to cause any on-channel receiver using either tone frequency to falsely decode the turnoff code as a 200ms "kerchunk" - and 136.5 Hz gets hit more often.

From an email to repeater-builder:
...although a good decoder will have an acceptance bandwidth of about 1.1%, some are a bit wider in order to improve decode time. A 136.5 Hz decoder with a 1.54% bandwidth would accept 134.4 Hz, whereas a 131.8 Hz decoder would need a 1.97% bandwidth to accept 134.4 Hz. All other things being equal, I would expect 136.5 Hz to exhibit a higher degree of false decodes.... And a really good decoder would recognize that a square wave CDCSS turn-off code had been received and not false decode on the 1010101... pattern.

Digital squelch takes just a bit longer to initially decode than tone (CTCSS) squelch. Digital squelch is also a lot less prone to false decoding than tone squelch, as the digital data contains error correction information and the 23-bit words must be detected error-free (at least by the older Motorola hardware based decoders, but some of the software based implementations are more tolerant). However, once a digital squelch circuit is open the requirements for continuous error-free data are relaxed for the duration of the carrier, allowing the input signal to momentarily drop out without requiring the longer decode time when it returns.

Technical issues:

You will find that many radios that were made pre-DCS are not worth the effort to adapt for DCS service, due to the audio bandwidth. Due to the square edges of the digital modulation the response has to be to almost DC in both the receiver (for decoding) and the exciter (for encoding).

A comment on DCS code selection, from an email sent to repeater-builder:
We ran into a situation where the local Motorola shop sold a local entity a repeater and several portables programmed with a DPL code that only certain models of Motorola have. We checked every Kenwood, Vertex, and even some other Motorola models and none could do this particular code.
The list of 104 codes from the Motorola MSF5000 station is here: Motorola PL and DPL codes.

A few words on "Private Line" (and "Digital Private Line")...

Back in the early 1950s Motorola chose the unfortunate word "Private" as the first word in their trademark. This has caused confusion for over 50 years, and some unethical salespersons have capitalized on the confusion. FRS radio manufacturers do it even today - look at their use of the words "Privacy Codes" on the packaging and in the user manuals. Using CTCSS (or CDCSS) does not give you privacy from being heard, it simply gives you the ability to switch your decoder on and to not hear ANY OTHER activity on the channel (as long as the other users are using no tones / codes, or different ones).

Anyone using tones (or codes) is simply adding an extra signal to their otherwise normal FM modulation. This gives them no privacy as such (anyone can listen in with a carrier squelch radio if they want to), but merely means that they are not bothered by any other signals unless those signals also carry the same tone.

In closing, a couple of paragraphs from Wikipedia:   (but Wikipedia pages change, what you see if you go there today may be different)

Family Radio Service (FRS), PMR446 and other "bubble pack" radios often use from 10 to 38 different CTCSS tones (the number depends on the manufacturer), usually erroneously called "sub-channels", or "privacy codes" in the sales literature. While these do not add to the available number of conversations which can take place at once in a given area, they do reduce annoying interference experienced by users that enable the use. However they do NOT afford any privacy, no matter what the sales literature says. A receiver with the tone squelch turned off (i.e. in carrier squelch mode) hears everything.

It is a bad idea to use any coded squelch system to hide interference issues in systems with life-safety or public-safety uses such as police, fire, search and rescue or ambulance company dispatching. Adding tone or digital squelch to a radio system doesn't solve interference issues, it just covers them up. The presence of interfering signals should be corrected rather than masked. Interfering signals masked by tone squelch will produce apparently random missed messages. The intermittent nature of interfering signals will make the problem difficult to reproduce and troubleshoot. Users will not understand why they cannot hear a call, and will lose confidence in their radio system. In a worst-case scenario in a life safety environment a missed message, or a misunderstood message, will result in one or more deaths.

Back to the top of the page
Up one level   Tech Index
Back to Home

Credits / References:
The author would like to acknowledge the contributions of the following fine folks to this article: Eric Lemmon WB6FLY, Nate Duehr WYØX, Will Martin KA6LSD, Bob Meister WA1MIK, Lou Sturm N6LHS, Doug Marston WB6JCD, Jeff Kincaid W6JK, Don Best N6ALD, Dave Kaar KA9FUR, Kevin Custer W3KKC, Robin McCray W3RSM, Eric Lowell W1EL and Bruce Carpenter W3YVV.