The HAL RTTY-1 Tuning Display

              The RTTY Tune is a device marketed by HAL Communications to displays standard split teletype tones in a traditional cross display that mimics the “Crossed banana” displays of oscilloscopes.  In its standard configuration, the RTTY-1 takes raw audio from the receiver or other source and filters out all but signals near the 2125 Mark and 2295 Space tones used for old school baudot teletype.  Of course, ASCII and AMTOR/SITOR tones near these same frequency pairs will also be indicated.  The RTTY-1 does not decode anything, but rather just indicates the amplitude of the signal found within its passbands.

              When properly tuned, the arms of the cross will be equal length and the stronger the signal, the longer the arms.  They will alternate back and forth at the baud rate indicating proper reception of the signal.  If the signal is not centered on the standard tones, the arms will be of differing lengths and a slow, but sure alteration of the receive frequency should bring the arms into alignment and decoding should begin.

              The RTTY-1 is small and a far easier display to use than its predecessor the RS2100 oscilloscope.  The solid state RTTY-1 is much smaller and more power efficient while giving a much larger display as well.  The RS2100 was an actual CRT oscilloscope requiring 120V input in a large cabinet and substantial current draw.  The RTTY-1 is about an inch thick and runs off 12-13V DC.  The RTTY-1 also only requires a single audio feed via a phono plug, while the RS2100 requires decoded mark and space inputs from a terminal unit, a feature missing from most modern equipment, and difficult to pull from a DB25 port on other units.

              The standard RTTY high tones are the default of the RTTY-1.  However, if you change certain resistors, alternative audio frequencies can be selected.  Unfortunately, this is more of a permanent modification and does not allow for rapid changes.  The idea is to be able to use the indicator with marine radios or other equipment that uses lower frequency pairs.  This limitation makes the unit less useful for modern amateur RTTY operation where signal selection is usually made on a waterfall display by clicking the mouse and the tones can be anywhere from 500 Hz to 2500 Hz. 

              Where the unit holds the most promise is older hardware that relied on small LED indicators of various quality.  I can see this unit being very helpful for PK-232 or older KAM and SCS units that used bouncing LED’s to indicate tuning.  Of course, old terminal units such as the HAL ST-6 would benefit greatly from this display if a vintage RS2100 isn’t available.   For display purposes, it also makes a nice light show for when you are listening to RTTY.

              Typical of most HAL products, the power connector isn’t standard so if you get one missing the HAL specific coaxial power plug, it is worth a note to Urbana to acquire a new one.

              HAL did publish a paper on how to modify the tones of the RTTY-1 to either Low tones or Marine tones.  The former centers at 1,360 Hz (1275/1445) and the latter around 1,700 Hz (1615/1785).  Unfortunately, both modified tones still don’t help with most waterfall displays which like to center around 1500 Hz.  Still, the low tones are far closer to the center of the audio passband than the standard high tones.  I have not modified mine, since I tend to listen to internet teletype which is still sent on the high tones and all the scopes in my shack are calibrated to these tones.  However, if you would like to modify an RTTY-1 to the low tones you must replace some resisters.  The following is the table


Low Tones


Change     from   to

R1         6,980    11,000

R2         6,490    10,500

R12       7,500    12,400

R13       6,980    11,820

R6 adj 2295 Hz to 1445 Hz

R20 adj 2125 Hz to 1275 Hz



Marine Tones


R1         6,980    8,870

R2         6,490    8,450

R12       7,500    10,000

R13       6,980    9,310

 R6 adj 2295 Hz to 1785 Hz

R20 adj 2125 Hz to 1615 Hz


              Of course, my mind always goes to the place where you could add 6 potentiometers and have a variable resistance for the six resistors and then you can adjust the tones to whatever you’d like.  Of course, you’d have to mark each control to ensure that each is at the proper value for the tone set you’d like to see.  You would also be able to adjust it so that you can have different tone spreads for commercial/military teletype.  This would take some engineering of course as well as a plan to wire up a control panel with the new controls.



                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      

Front Panel controls of the HAL ST-8000A

              The ST-8000A was designed for the military and as such, the controls are simple to use and robust in construction.  There is a single toggle switch that powers the unit on, and the rest of the controls are membrane buttons.  There is a “2^nd” key that allows for most buttons to be used for two purposes.  There are also LED’s used for tuning, status indicators, and data output for settings such as tone and baud rate that have been selected.  Once quirk of the controls is that once the selection has been typed in, a green LED will blink until the enter key is selected to finalize the input.

              The 8000A is a complex modem that can be configured for a wide variety of inputs.  However, I use it only as an audio filter between the radio and the decoding modem, currently a HAL DSP-4100.  In this case, the important settings for operation are the setting up the filters so that the signal gets processed, and ensuring that the audio path is correct.

              There are two methods to tune the filters of the 8000A, both get to the same place, varying only in the parameters that are being utilized.  The traditional way to set the filters is to have the unit display the mark and space frequencies along with the baud rate.  First select the M/S display and enter each frequency in turn.  The difference between the mark and space will be the shift and the average of the two numbers will be the center frequency.  Since I am feeding a DSP-4100, I set the M/S at the standard Amateur RTTY tones of 2125 and 2295 at 45 baud.  This gives the standard shift of 170 Hz.  Since I am feeding the audio to a DSP-4100 that has it’s RTTY filters set at these tones, it is easy to match the settings.  The Radio RTTY mode outputs are also set at these tones, so the M/S display and setting method makes tuning amateur RTTY easy.

              The second method is to have the 8000A display the center frequency, shift, and baud.  This method is more used when monitoring commercial or military RTTY that use alternative shift and baud rates.  Since commercial and military stations are not as worried about bandwidth, the shift is often around 850 Hz, even with a relatively slow baud rate of 50.  In this case, the radio will be set to USB and to more easily utilize the radio passbands, the 8000A and DSP-4100 tone selections need to be changed.  To make things simple, I change the Cf to 1500 Hz, and select the shift  to the appropriate amount and change the baud rate, all by direct entry.  I then switch the display to M/S where the tone frequency are automatically displayed.   I then go to the terminal software and use the configuration commands to align the DSP-4100 to the same M/S settings.

              For an example of copying commercial RTTY over the air, German weather station DDK broadcasts weather over RTTY using a 425 Hz shift and 50 baud.  I dial in the frequency on the radio set to USB to 10.099.315.  This centers the tones at 1500Hz in the passband.  I can then use the various radio filters to reduce noise and interference.  I set the 8000A to 1500 Cf, 425 Shift, and 50 baud.  Switching to M/S display, I see the tones at 1287.5 Hz for Mark and 1712.5 Hz for Space.  I then go into the Hal Software and adjust the baud on the front panel, and go to the configuration page to change the M/S frequencies.  Using these settings, audio passes through the radio filters into the 8000A where it is filtered and then passed to the 4100 where the audio passes through the DSP board and finally to the decoder where it is converted to serial data for display on the screen.

              Traditional RTTY is sent on LSB but new amateur convention, which is nearly all soundcard modems now, uses USB for digital work.  To correct for this problem when using legacy equipment, everything has to be set to Reverse.  Since I am using FSK settings on the radio, the 8000A is set to REV as well.  I keep this setting for commercial RTTY and reverse it again in the software rather than change the modems around.

              There are some articles on the internet about using the HAL ST-8000 and ST-8000A as regeneration units to supply audio to a second decoder for conversion to serial data that is fed to a computer for display.  This is known as regeneration.  Basically, audio from the radio is filtered and decoded by the primary modem.  Instead of outputting serial date, the modem generates high quality Mark and Space audio tones that are fed to a second decoder for conversion to serial data.  The second decoder can be pedestrian in performance because the quality of the output from the regenerating modem is extremely high, even if that output is full of errors.  In effect the signal is decoded twice, but any errors in the first decoding are carried through to the input of the second decoder.  This is a stop-gap method to use the older TU with newer equipment.  My setup does not work like this.  In my system, the regeneration is on, but the pin that has the output signal is not connected to anything.  In my set up, the audio is filtered in the 8000A but never decoded.  The filtered audio is extracted before the demodulator and fed to the DSP-4100, where it is further processed and then decoded using higher end computers rather than just the analog decoders in the 8000A.  I have a ST-8000 in line that directly decodes the same audio that is fed to the 8000A.  The decoding of the 8000A/DSP-4100 is superior by a significant margin than the old “gold standard” that was the ST-8000.

              The final article in this series will see how radio teletype is utilized at N1ZZZ, and it certainly is, if not unique, at least highly unusual.

HAL ST-8000A Powering up and Internal Testing

Before you purchase a HAL ST-8000A, it would be best for you or the seller to run the internal test procedures to check your firmware and to ensure that everything is still running correctly.  These units are over 20 years old now, and the electronics could start to fail if they have been sitting in storage for too long.  The good news is that this testing is very easy, and only takes the power cable to complete.  So even a NOS unit can be powered up and tested with no technical knowledge about the unit.

              Check on the back panel to make sure that the frequency and voltage match your power mains.  Unlike most TNC’s, the 8000A uses about 30W of AC power and not 13.8 VDC.  In the US this would be 115V and 50/60 Hz.  Europe and Asia require it to be set to 220V and 50/60 Hz.  The frequency switch can also be set to 400Hz for aircraft use in accordance with MIL-STD-704.  If your 8000A has the accessory package, you should be supplied with two power cables, one that is a standard desktop computer cable, and a short adapter that allows the unit to be plugged into military power systems.  The male plug is for the US standard, but adaptors or other cables with the proper female socket will work.

              Once the unit is plugged in, flip the front power switch and the 8000A will go through the first diagnostic check.  During this check the firmware version will be displayed.  The newest firmware is version 1.9.  Once the start up has ended, the unit will show mark and space frequency as well as baud rate, and you should get a slight flicker in the LED bar graphs.  A few of the display LED’s might also be illuminated, depending on the settings.  This is the first test, and the one most ignorant sellers will claim is all they can do.

              The second test is called the BIT test, and this tests the internals of the ST-8000A by routing the modulator through the demodulator to check that the circuits are working correctly.  To perform this test, turn the unit it on, and upon completion of the start up test, hit the “2^nd” key, then hit “BIT” and finally “Enter.”  This will run the machine through a 13 step self-diagnostics test, including testing all the LED’s and testing the demodulator at several audio levels.  If all is well, the unit will end the test by putting “Pass” on the LED display and finally returning to standby mode.  If there is a failure at any point, it will display the step of failure and say “FAIL” on the screen.  The test will not proceed beyond the failure point.

              Once the ST-8000A is verified to be functioning correctly by these tests, it is time to move onto the wiring to be able to get audio into the unit from the source, and then intelligence back out.

HAL Communications ST-8000A Overview

              This will be the first of a series of blog posts on the HAL ST-8000A.  I have had quite an adventure getting one of these terminal units working in my station and there is very little information on the internet to help with the process.  I hope to provide an overview as well as some hints and tricks I learned, or discovered that helped me along the way.  If someone else wants to use this unit in their station, I hope that they can use this information to make the process more painless than mine was.

              The HAL ST-8000A is a teletype modem that was designed for the military to provide a drop-in replacement to the Frederick Electronic 1280A modem.  It can be used in both radio and hardline applications and converts FSK Teletype at a wide range of baud rates to/from either RS-232 or MIL-188 protocols.  Since it was designed for the US Air Force, it uses US Department of Defense Bendix plugs and is built to military specifications.  This means that it is very well build, EMP hardened, and subsequently very costly.  This was HAL Communication’s last terminal unit that did not have a DSP circuit.

              The flow diagram of the unit is as follows: audio is fed through a balanced line into the unit where it passes through a series of filters.  The filters are automatically selected based on the programed front-end settings.  Filter width varies directly with the selected tones and baud rate.  The unit can handle tones from 300 Hz to 3000 Hz, which will handle most radio SSB bandwidths.  There are a series of low pass filters that cut out noise and then the audio is routed both the I/O port (more on this later), and one of two individual discriminators. The first discriminator is used at baud between 30 and 600 baud, and the second for baud rates from 601 to 1200.  The discriminators decode the audio input and uses a microprocessor to convert it to one of three digital signal outputs: 1) RS-232 data, 2) MIL-188 data, or 3) audio tones (Regeneration).

              On the transmit side, the ST-8000A will take data input (RS-232 or MIL-188) and convert it to AFSK tones which will drive the SSB radio or pass it over hard lines.  This is very similar to how most sound card and many TNC’s are utilized to send RTTY.  The unfortunately part of this is that you cannot use the FSK mode on the radio which is disadvantage on many radios due to the filter limitations while in SSB mode instead of RTTY mode.

              Using the ST-8000A as designed will require the appropriate software to decode the RS-232 or MIL-188 data streams in to text.  Given the age of these units, the terminal programs will be difficult to find, and will need to use older operating system emulators to run the software.  Add to this the lack of DSP and the use of AFSK on the transmit side, you may wonder why you would bother to use an ST-8000A in the first place.  The answer to this is the front-end filters.

              When people still used the ST-8000A, or it’s civilian cousin the ST-8000, the operators were looking for a way to use the superior decoders on the ST-8000’s while using their favorite software, which often did not interface with the terminal units.  The stop gap measure was to use the “regeneration” feature of the 8000’s.  This had the 8000 decode the audio with it’s superior filters and discriminators, and then convert them to pure, high S/N ratio audio signals to be converted to data by a secondary TNC which fed data to the software.  Any TNC was able handle the high-quality signals from the regenerator, and then the secondary TNC could be wired to send FSK signals to the rig, overcoming the second issue of the 8000’s.

              With the utilization of DSP, this stop-gap use of terminal units was rendered obsolete.  While some newer TNC’s use DSP, it was pointless to input regenerated tones into them, as the regenerated audio did not need further processing.  In standard soundcard interfaces, software was able to take audio directly from the radio and DSP algorithms were superior to the discriminators to the 8000’s.  Some software and interfaces could even overcome the FSK keying issue and software took over from the terminal units in most applications.  The use of external TNC’s and expensive terminal units essentially stopped.

              My idea in getting the 8000A was to see if I could use the TU as an audio filter and then feed the processed audio into a DSP TNC to take advantage of the filtering of the ST-8000A and then use a secondary TNC to use DSP on the signals, and decode them.  This also allows the secondary TNC to use FSK keying.  This plan could also have audio sent to a sound card interface and so the same thing, but onboard the computer.  This was the challenge that the internet had not seemed to meet as I rooted around for information on the HAL ST-8000A.

              The 8000A was never as popular as the 8000.  A few reasons for this are the fact that the 8000 used standard 25 and 9 pin serial connectors which were easily and cheaply available.  The 8000A uses Bendix 37-pin plugs that are harder to find, and far more expensive.  The other big issue for hams was the lack of dials and the replacement of an analogue tuning oscilloscope with the less popular MARK/SPACE LED bars.  The fact that the Mil-spec 8000A was also considerably more expensive sealed the deal.  Today you can find 8000A’s inexpensively, but the 8000’s are quite rare.  While the interface issue is a matter of preference, there is a way to pull filter audio out of the 8000A and feed it into an oscilloscope for a similar display.  The plugs, while expensive, can be bypassed and refitted because the jumpers between the plugs and the board go into a Molex plug.  Just be sure if you go this way, to ground all of the unused pins or you will get errors. 

                I was able to find a ST-8000A online for a price that I could justify, and had it shipped to me.  The unit had never been wired up, so all the original accessories including the Bendix plugs and pins were included.  Unfortunately, the instruction manual was not in the box.  Fortunately for me, the customer service at HAL is excellent and I had a pdf version of the manual in my Dropbox account which was worth it’s weight in gold.  Now that I had the ST-8000A, I had to research to see if it could do what I wanted it to do, and if it could, how I needed to set it up.  Once I had the ST-8000A in my shack, the real adventure could begin.

Operating Clover

Clover is one of my favorite high frequency (HF) modes.  It was developed in the early 90's and was brought out as a proprietary digital mode in the early 1990's by Hal Communications.  There are currently 3 variations of the mode in current use, Clover II, Clover 2000, and Clover 2500.  A good history and technical discussion can still be found at Hal's website http://www.halcomm.com  I won't try to rehash all of the ins and outs of the mode, but will instead give a fairly brief overview of the mode.

Clover II was the first commercially available iteration.  It consisted of 4-tones sent sequentially in a 500 Hz pass band.  Each tone shifted phase in order to pass binary characters. The number of phases increased with better band conditions until it reached 16 phase shifts.  In addition, in the highest modulation schemes, amplitude modulation was added to pass more data.  All of this was done at the low symbol rate of 31 Hz.  Clover II was targeted at both the amateur and commercial interests.  Unfortunately the cost of the mode was too much for the amateur community and only caught on with some, primarily US based, commercial interests.  Data rates for Clover II range from 125 bps to 750 bps and was available new until recently in several modems.  The last Hal unit to carry Clover II was the DSP4100

In 1999 Clover 2000 was introduced.  The basic modulation was the same, but the number of tones was increased to 8 and the bandwidth spread to 2 KHz.  Additionally the symbol rate was doubled for more throughput.  Hal decided that this mode was a bit of overkill for the amateur community and marketed it almost exclusively to commercial interests.  The modems remain expensive, but the increase in throughput was impressive at 500 to 3,000 bps.

In 2011, Clover 2500 was introduced in the late model DSP4100/2K and DSP4200 modems.  The symbol rate was increased even more to 71.125 baud which lead to an increase of bandwidth to 2.5 KHz.  This shows the practical limit of Clover in standard SSB transceivers and seems to be marketed in response to a rival's introduction of a new mode at approximately the same time.  The data rate for Clover 2500 is 625 to 3750 bps.

While there are published reports about the technical specification of Clover on the internet, there are few, if any on-air reports available except those found in relatively old equipment reviews of clover II.

I have fairly extensive experience with Clover II over the last 10 years or so, primarily with my DXP-38 modem.  This was the last modem marketed to amateurs with Clover in it.  It featured a tuning indicator which was omitted on commercial modems which primarily operate on fixed frequency channels and should not require manual tuning.

Clover II operation requires precise tuning, to within 20 Hz for link establishment.  The modem does not have the ability to compensate for frequency drift, but with statistics passed between the modems on a regular basis, the radio can be slowly (no more than 10 Hz at a time) zeroed in to each other.  Amateurs using the mode tend to have a set frequency to meet on (14.065.5 LSB dial is the most common) and fine tuning for rig errors can be done once the link is established. 

Once the link is established, Clover is pretty much on auto pilot.  All the operator has to do is type or send files.  The mode works on automatic overs after certain amounts of data is passed.  The length of these blocks between overs is determined by the amount of data to be sent and can last several seconds.  Overhead and a small amount of keyboard data are sent via short blocks with a very robust waveform while data blocks are longer and will be sent at the highest speed possible as determined by conditions on the other end of the link.

I have also experimented with Clover 2000 several times.  This mode is much more difficult to operate than Clover II, especially with lower power systems such as amateurs often use, and me especially.  The same amount of power is now spread over 2 KHz and 8 tones instead of 4 tones in 500 Hz.  This means that there is a lower average power on the other end of the link requiring more ERP and better conditions.  Clover 2000 is also set up to run with fewer error retries and is very prone to link failure if struck by fading.  Still, when the conditions are right, Clover 2000 will move data very quickly, even at it's more robust waveforms.

Clover is unique in current modes in that statistics for both sides of the link are readily displayed.  Waveform , amount of online error correction used, S/N ratio, frequency offset, throughput, and phase dispersion are all shown to the operator.  The most important to judge the link condition are S/N and PHS (phase dispersion).  The former shows the relative strength of the signal and higher numbers are desirable.  The second shows how much phase distortion is being caused as the waveform is propagated.  lower numbers are better in this case.  Since Clover relies on being able to determine phase shift, higher amounts of dispersion require fewer shifts per tone, and that will mean less throughput.  SN around 30 and PHS in the teens indicate a good link and will allow speedy transfer of data.  PHS in the 40's and S/N in the teens will expect low throughput.  Any worse than those numbers, you can expect link failure or at least numerous error signals. 

For Keyboard chats, using Clover 2000 is both frustrating and wasteful of spectrum.  The enhanced signal conditions that are required coupled with the relatively slow rate typing and reading make the mode difficult to use.  Save this mode for passing traffic.

Clover II is a very nice keyboard mode due to it's speed and semi-duplex nature.  Also the link status panel is of interest to radio enthusiasts.  It can also serve as a relatively fast file transfer mode provided the file isn't very large.

There is a yahoo group dedicated to Clover operating with a small but dedicated following.  So if you see a Hal modem online that has the mode for cheap, have a go at it.  If you think that clover alone isn't worth even the modest used prices of these modems, you can still use them as great RTTY or Pactor I modems.

Pactor II verses Clover II

In the early 1990's DSP chips finally became financially feasible and two DSP based, phase-shifted modulation modes became available to hams.  Ham developers in the USA and Germany separately developed two HF digital modes that combined high data transfer speeds while remaining fairly robust and capable of maintaining point-to-point links over the HF radio spectrum.

The two modes were Clover and Pactor II.  Implementation of the phase shift modulation was slightly, different, but these two modes were generally the same speed and claimed nearly equal robustness in the face of low S/N ratios.

I have always been a big fan of both of these modes and have used both since about 2003.  These modes have be upgraded over the years, but the version two of these modes have always been very good for keyboard links. 

Recently I was on the air with Pactor II with VE1XL.  Conditions were poor with large fades in the path.  Even the robust Pactor II was having some trouble moving data.  It was still good enough for keyboard work, but the fades caused several error messages going back and forth as the data packets did not make it.  Given that these were adverse conditions, I suggested to VE1XL that it would be a nice opportunity to test Clover and Pactor head to head. 

We ended our Pactor link and switched to our DXP-38 Clover modems.  These are the latest of the Hal Clover II modems and gave the mode the best chance at keeping a link.

We were able to establish the link, but we had trouble with the data blocks.  Unlike Pactor, Clover uses what it calls Clover Control Blocks.  These are sent at a modulation scheme that is the most robust it can muster.  Unfortunately the CCB's modulation is not available for data packets.  This proved to be the deal breaker on this link.  Despite some error packets, the CCB's managed to get through, but as soon as the data packet tried to make it through, it was not able to get there.

Sadly eventually the link failed and we had to go back to Pactor.

So what shall we make of this?  While I still find Clover more fun to operate due primarily to the fact that it has a more effective semi-duplex paradigm, and the ability to exchange station data on both ends of the link, it still needs more power to keep the S/N ratios up to the point where data can make it through.  Pactor is a superior poor link mode, and this is a common consensus.

73
N1ZZZ