JT9 (English)




Weak signal communication on the HF-Bands

By SM7VRZ, Anders Rhodin



JT65 have since it´s introduction to the HAM-radio society grown very popular. What started as a digital mode for EME in the VHF and UHF bands, now also have conquered the HF-band with many users around the globe. With a ordinary PC with a soundcard, cables to the transceiver, simple antennas and low power output JT65 have shown that you don´t need much to work DX. The wide usage on the HF-bands contributed to the development of JT9 which is more optimized for use in LF, MF and HF bands.


JT9, overview

JT9 is as JT65 designed for communication with weak signals and low output power. Commonly, output powers as low as 1-20W is not unusual as there is no need for more to work DX. The mode is using the same basic idea as JT65 and is, for the operator, much the same with a similar message structure and usage.

One of the advantages with JT9 is it’s bandwidth. Instead of using about 177Hz, as for a JT65 signal, it only uses about 15Hz. This means that, in theory, ten JT9 signals can be using the same bandwidth as one JT65 transmission. The narrow bandwidth is reached by using FSK (Frequency Shift Keying) modulation containing nine tones with a tonespacing of about 1,7Hz. The mode allows only 13 symbols to be transferred with each transmission, wich is the reason for using a number of short standard messages for communication. There is also a possibility for sending free text messages containg maximum of 13 symbols.

Figur 1       To the right, a JT65 signal and to the left a JT9 transmission.

A TX or RX period takes one minute to complete and starts every egen or Odd minute. Of 60 seconds, 50 seconds is used to send or receive the message itself. The remaing 10 seconds is used for decoding and operator input. A complete QSO with standard protocol messages requries six minuters.


Error correction and coding

JT9, just like JT65, is a weak signal communication mode and can, thanks to a powerfull error correction, decode messages from signals with low signal to noise ratio (SNR). In an ideal noise enviroment the decoder can work down to -24 dB SNR and manage to decode 100% of the received messages. Down to -27 dB SNR the decoder can work out about 50% of the received messages.

The message you want to transmitt ges through a number of steps from the message is entetered until its presented as a readable text message at the receiving station. A message that is to be sent is encoded with a error correcting coding known as a “Convolutional code”. After the encoding with the error correction, its translated to a tone sequence containing nine different tones in 85 intervals. These intervals is distributed over the 50 seconds a transmission is made. Eight of the tones is reserved as data carrying tones, containg the bit information you wish to transfer. The ninth tone is a synchronization tone and is always the lowest in the pass band. The tones are then transmited from the computer soundcard into the transceivers audio input.

When the decoding of a received signal is made, the decoder try to find suitable signal candidates in the pass band of the received freuency. It does this by trying to identify the synchronization tone. When a signal is succesfully found, it calculates the signals time and frequency offset, identifying the transmitted data tones and converts them into a bit coded message. The message is then inputed to the error correcting convolutional code wich decodes the messages to the original operator message. Thanks to the error corrective properties of the convolutional code, the decoder is able to recreate missing bit information in the operator message.


Message structure

As mentioned earlier short operator messages is used for the communication. The protocol uses a special message structure to get the best use of the available 13 symbols and is because of this divided into a number of standard messages. In the standard messages you transmitt callsigns, grid locator, transmission receipts and 73 in a predefined sequence. Also, there is a possibility to transmitt a free text message with a 13 symbol length.

Below is an example of a standard QSO:



The signal report is automaticaly generated by the decoder which measures the signals average signal to noise ratio (SNR) value during reception. A report between -50 to +49 dB is possible to be reported by the software.


Time synchronization

As this mode is dependent of transmissions and receiving periods are started and ended at a precise time, its very important that the software is in synch with the opposit stations time reference. A time drift around 2,5 seconds makes your transmitted signal undecodable, even If the quality is good. The software uses the computer system clock as the local time reference wich normaly isn’t stable enough to deliver the correct second based reference for a longer period of time. To solve this problem its mostly necessary to instal a NTP-client software inte the computer used for operating JT9. An NTP-client software uses reliable time references distributed over the internet to set the system clock to a correct time at shorter intervals, reducing time drift.



JT65 has a number of standard frequencies on the most of the amatureradio bands from 2200m and up. These are automaticaly set by the software that’s used for the mode or in other cases indicated by the software. As the passband on these frequencies is already occupied by JT65 traffic, JT9 users is alocated about 2kHz above these frequencies. The table below shows the currently used frequencies for JT9 on HF and 6m. All is in USB mode.

jt9 frequency


JT9 software

For the time being, there is only one software who supports the JT9 mode. The software is called WSJT-X and is created by the inventor of the JT9 and JT65 mode, K1JT, and a number of engaged radio amateurs in an open source project. Except for JT9, the software is also capable of using JT65 with a optional “multi mode” function that allows both modes to be decoded in the passband.

WSJT-X main window is divided into three sections. The “Band activity” list shows the active traffic in the passband, “RX frequency” list shows the traffic on the selected frequency in passband and the lower section which holds the input and output audio levels and operator messages.

Figur 2

WSJT-X main window. I´ve just finished a QSO with EA4BPN on10m.

You can also activate a second window, “Wide Graph”, which shows a waterfall view of the incoming signals. The waterfall shows the incoming signals as a function of time and frequency and a signal as a change in color. The color of the signal is also an indication of itˋs strength.

Figur 3

“Wide graph” window. At 800Hz and 1300Hz there is two JT9 signals visible.

By default, the software is set to decode both JT9 and JT65 signals simultaneously. JT9 signals is however only decoded over 2600 Hz in the passband with this setting. To set the program to use only the JT9 mode, you can set the software to work only JT9 and adjust the transceiver frequency up by 2 kHz to shift the passband to the JT9 frequency passband.

The software also offers the possibility for CAT-control, which allows the program to automatically adjust the transceiver frequency and PTT signal. I you donˋt have the hardware for using CAT-control, itˋs also possible to use a serial port for PTT or VOX. There is also a software interface towards Hamradio Deluxe. For logging purposes, there is a basic ADIF compatible logging function.

The software is available for Windows, Mac and Linux and is free of use. You can find the software at: http://physics.princeton.edu/pulsar/K1JT/wsjtx.html


JT9 submodes

JT9 is also divided in to submodes. All submodes are much the same with nine tones, synchronization and data tones, but the differences is transmission time, signal bandwidth and tone duration. JT9-1 is the submode which is used in the later versions of WSJT-X, the other ones are only available in previous versions of the software. The table below shows some of the differences between the submodes and it´s properties.


AS the table shows, there are many differences between JT9-1 and the other submodes. An interesting thing is the S/N threshold for JT9-30 which is down to -42 dB S/N. However, the cost for this is a total QSO time for 180 minutes.

From WSJT-X version 1.4.0 you can notice a further development on the JT9 mode, JT9W-1. This mode is developed for EME on microwave bands.


Working JT9

To work JT9 is in practise quite easy. If you are a new user of the mode it’s recommended to start with listening to the traffic, noticing how other users do. This also give you possibility to adjust the audio levels for optimum settings and to familiar yourself with the software. The online manual is easy to use and contain all info that you need to be up and running.

After you have been listening for a while it’s time to transmit a CQ. Before you do that you need to adjust the output level to the transmitter. Normally you can use 20-30W in JT9, then this is considered as a high power setting for this mode. It’s recommended to use 1-10 W as a starting point and adjust after your needs. As this is a weak signal mode, you might see it as a bigger accomplishment to use lower output power.

When you adjust the output level from the software you have to pay attention to the transceiver ALC meter. The meter shouldn’t show any reading while transmitting. Adjust the signal so that the setting is just below the level where the meter starts showing the readout.

When you have got started with JT9, you will notice that you are able to hear a lot of DX stations. To work DX is also quite common even with simple antennas. My own experience  have shown that you can work North America with a simple dipole antenna mounted approximately two meters above ground using 5W and still get reasonable signal levels.

To give you a more visual example of how to work JT9 in reality, you can watch this Youtube video. The video is made by N7JFP, Lawrence, and shows a JT9 QSO between Washington state and South Dakota using 4W. Lawrence is using an older version of WSJT-X and a accessory software which is given more information on further on in this article. The radio is also a SDR-transceiver which provides a more graphical possibility to monitor the transmitted signal.


What do you need to get started?

To get started you need, except for the software WSJT-X, also the possibility to connect your transceiver to a computer. Also, the computer must meet some demands regarding the hardware. According to the software manual, this is the minimal hardware required:

  • Windows XP or later, Linux or OS X
  • 1.5 GHz processor or faster
  • 100MB of available memory
  • A monitor with at least 1024 x 780 resolution
  • A soundcard with speaker output and microphone input, with at least 48 kHz sample rate.

Except for these demands you also need cabling to connect your radio to the computer and  a way to key the transceiver. The image below gives an example of how a connection between the transceiver and computer can look like.



Connection between computer and transceiver 

 The minimum required cabling that is needed is the following:

  • Audio cabling from the soundcards speaker output to the transceiver microphone input.
  • Audio cabling from the transceiver speaker output to the soundcards microphone input.

This solution requires you to key the transmitter manually unless you activate the transceivers VOX function, which will automatically key the transmitter when an audio signal is detected on the microphone input. Building your own connection cables is quite simple but you may have to consider preventing grounding currents when building. This is done using transformers, separating the computer from the transceiver in a galvanic sense. There is also another possibility controlling the transceivers transmitter by using a CAT control cable. This also enables the software to control the transceivers frequency setting, adjusting the required frequency automatically.

There’s also finished products in the form of USB-connected interface devices with built-in sound card and control circuits for keying and in some cases control of the transceiver. These are usually plugged into the radio’s accessory socket. Some examples of such interface is Tigertronics Signalink USB, Rigexpert TI-5, Rigexpert TI-7 or the Rigexpert Standard.

When it comes to further requirements of the transceiver and antenna, the ones you  already have will do just fine. probably will do what it has already just fine. An indoor antenna in the form of magnetic loop or tuned vertical antennas are good enough to get started. My own experience shows that dipole antennas, mounted i unfavorable heights in combination with QRP output power works great for working DX in certain cases.

As mentioned earlier, it is also a requirement that the computer time reference must be in sync. To meet this requirement you need to install a NTP client on your computer. An example of such is Meinberger NTP client software that is free to use and install.


Utilities and tools

Internet today offers many possibilities and resources for radio amateurs. Some of these are useful even when running digital modes such as JT9.

To get a good idea of what stations you have heard and worked, you can use the web service Pskreporter.info . The service collects data of which stations was heard by your software and also which stations heard you. The data is automatically sent from WSJT-X via the internet and presents the stations on a graphical map. Here you can also choose to show those who heard your signal.  In that way you can not only get a nice map with the stations heard, but also a decent idea about the band conditions.

Figur 4

Screenshot from Pskreporter.info. The map shows stations heard and how well they hear me.

To use the service you only need to enter your callsign in the settings box at the top of the page and press “Go”. After this, the stations that the software heard and reported as well as those heard you will be displayed. Also notice that you will have to activate the reporting in the software for this to work. If not activated, only reports from those who heard you will be displayed.

The service can be found at: http://pskreporter.info/pskmap.html

For WSJT-X there are also tools in the form of accessory software. A popular one is JTAlert written by VK3AMA which adds extra features to WSJT-X such as, keeping track of worked stations, worked DXCC countries, grid locators, US states, etc. Furthermore, there are also features to connect different sounds and alarms to desired countries, calls, etc. Also the program works as an interface to a variety of logging programs for automatic logging of QSOs. Additional useful features include band activity information which shows the number of active stations per band either via color coding of the main application window or in a table format as a poppup window.

Figur 5
Screenshot from JTAlert:s main window.


Figur 6

The band activity window in JTAlert.

The software is free to use and can be downloaded from: http://hamapps.com



JT9 is an exciting and perhaps different operating mode but is in the same time quite simple to work. As the mode is designed for weak signals, it becomes more exciting and challenging when using low output power into simple antennas. No matter how you choose to work JT9 I hope this article has given a small insight and perhaps even given you a to try out the mode.


Links and references

WSJT-X download page: http://physics.princeton.edu/pulsar/k1jt/wsjtx.html

WSJT-X User guide Version 1.4: http://physics.princeton.edu/pulsar/K1JT/wsjtx-doc/wsjtx-main.html

WSJT-X User guide Version 1.0: http://physics.princeton.edu/pulsar/K1JT/WSJT-X_Users_Guide.pdf

Meinberg NTP-client, download page: http://www.meinbergglobal.com/english/sw/ntp.htm

JTAlert download page: http://hamapps.com

PSKReporter.info homepage: http://pskreporter.info


Do you have any questions or comments? Use the form below to contact me.




2 thoughts on “JT9 (English)

  1. Thank you. Unlike other videos I have seen you made it very clear that you must enter both call signs at each step. I thought that these were entered automatically by the software.

    • Hi Thomas,

      Im not sure what you mean with entering the callsign in each step. I suppose you are referring to how the software is handling the callsigns along with the message generation. The software, at least in the current version of WSJT-X, handles the callsigns automatic when the standard message buttons are used. However you can override this and put the callsigns in manually if you wish. The only thing required is that you either dubble click on a received stations callsign or input it manually in the “DX-station” feild.

      However I hope you found the article interesting and hope to work you on the bands with JT9. 🙂

      73 de SM7VRZ

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s