Friday, November 4, 2016

Adventures on Digital Voice Using the ARD9000MK2


An ARD 9000Mk2 was purchased some years ago at the same time as the Yaesu FTDX5000MP; a present to myself on retirement. Early experiments included recording the transmitted output onto a tape recorder and playing this back through the unit to hear the results. It worked fine. A second unit had been found at a rally and it was snapped up for a mere 20 Euro - a bargain. Until recently, nobody had ever been worked or heard on digital voice irrespective of the number of times the rig was left monitoring a supposed digital channel on any of the HF bands.


The ADR 9000MK2

The AOR Corporation were the first ham manufacturer to introduce a Digital voice "modem" in 2004. Their first unit was the ARD9800 which had the ability to transfer files. The ARD9000 was produced without the ability to transfer files and was a voice only unit and a little cheaper to purchase. The design of this unit was based on a VOCODER (voice encode and decode) protocol designed by Charles Brain G4GUO. His protocol involved the use of Advanced Mulit-Band Excitation (AMBE), a propriety speech coding standard developed by Digital Voice Systems.

The specifications of the ARD 9000Mk2  are as follows:

Modulation method:   
                                   Orthogonal Frequency Division Multiplexing (OFDM)
Bandwidth:                   300Hz - 2500Hz, 36 carriers
Symbol Rate:               20mS (50 baud)
Guard interval:              4mS
Tone Steps                   62.5Hz
Modulation method:
                                   36 carriers: DQPSK (3.6K)
AFSK                          +/- 125Hz
Error Correction:           Voice: Golay + Hamming
Hearder                       1 Sec. 3 Tones + BPSK training pattern for synchronisation
Digital Voice:                AMBE coder, decoder
Signal Detection:          Automatic Digital detect. Automatic switching
                                   between analogue and digital mode.

Description and Theory 

There are a few common elements to a digital voice system. The Modem converts a signal from the microphone, or another analogue source, to a digital bit stream. This is referred to as an A/D converter. This is driven by an algorithm referred to as a CODEC. The signal is then processed into a modulation waveform that can be applied to a radio transmitter. The received signal has to be converted to a digital stream e.g. demodulated. The digital bit stream is then demodulated to form an analogue waveform via a digital to analogue converter by a matching algorithm or CODEC to that applied to the transmitter. The resultant signal can then drive a speaker. This is the basic function of the ADR 9000MK2.  See diagram below:


 The basic blocks of a digital voice modem

OFDM is a form of multicarrier modulation. An OFDM signal consists of a number of closely spaced modulated carriers. When modulation of any form - voice data etc. is applied to the carrier, the sidebands spread out either side.

On analogue modes, it is necessary for a receiver to be able to receive the whole signal to successfully demodulate the data. When signals are transmitted close to one another, they must be spaced so that the receiver can separate them using a filter and there must be a guard band between them. 

This is not the case with OFDM. Although the sidebands from each carrier overlap, they can still be received without the interference that might be expected because they are orthogonal to each other. This is achieved having the carrier spacing to the reciprocal of the symbol period.



Traditional view of receiving signals carrying modulation


To see how OFDM works, it is necessary to look at the receiver. This acts as a bank of demodulators, translating each carrier down to DC. The resultant signal is integrated over the symbol period to regenerate the data from that carrier. The same demodulator also demodulates the other carriers. As the carrier spacing equal to the reciprocal of the symbol period means that they will have a whole number of cycles in the symbol period and their contribution will sum ot zero - in other words there is no interference contribution.


One requirement of OFDM transmitting sytems is that they must be linear. Any non-linearity will cause distortion amongst carriers as a result of inter-modulation distortion.  This will introduce unwanted signals that would cause interference and impair the othogonality of the system.

Advantages:

Immunity to Selective Fading: One of the main advantages of OFDM is that it is more resistant to selective fading than single carrier systems as it divides the overall channel into multiple signals affected individually as flat fading channels 

Resilience to Interference: Interference appearing on channel may be bandwidth limited and will not affect the sub-channels. This means data will not be lost.

Spectrum Efficiency: using close spaced overlapping sub-carriers, a significant OFDM advantage is that makes efficient use of available spectrum.

Disadvantages: 

Hight peak to average power ratio: An OFDM signal has a noise figure like amplitude variation, and has a relatively large dynamic range, or peak to average power ratio. This impacts the RF amplifier efficiency as the amplifiers need to be linear and accommodate the large amplitude variations and these factors mean that the amplifier cannot operate with a high efficiency level.

Sensitive to  carrier offset and drift:  Another disadvantage of OFDM is that is sensitive to carrier frequency offset and drift. Single carrier systems are less sensitive.

In practice

A number of settings need to be performed before actually transmitting on air.

Ensure that the receiver is actually 100% on the same frequency as the other user. IF necessary, once the QSO is established, use the RIT.

The I.F. filters on the transceiver need to be set to 3KHz

Ensure that Speech compression and equalisation is switched off

Bear in mind that the duty cycle of the transmitted signal is around 100% so set the carrier level to a suitable power output in accordance with the radio specifications. 

Establish the initial contact on SSB and ensure that the resolution of the voice is perfect and the TX and RX are on the same frequency. This is essential. 

Ensure that the ALC meter shows no movement. If there is movement, back off the input from the modem using the mic gain from the transceiver. If the modem is a bit "lively" reduce the pre-set mic input on the mic pot on the underside of the modem. 

When transmitting pay attention to the overdrive light on the modem. Back away from the microphone or reduce the input to the modem.The overload light should seldom flash and occasional flash will not cause a problem.

It is a good idea to run the ARD9000 off a different supply to that of the rig. A small Gel Battery was employed to be on the safe side. This eliminates the possibility of a hum loop. Ferrite beads in the power line and on the mic lead will prevent any RF getting into the box. Again all precautionary. This will ensure that any poor decode is only down to mic gain on TX or input levels at the RX end.

The Receive settings:

Ensure that the modem is not overloaded watch the Overload light on the Modem. If it is flashing, the input to the modem is too low. If it remains off the input is optimal. If it remains on the AF input level is too high. These levels are dependant on the AF gain from the transceiver. The unit has its AF gain control for personal preference either through the speaker mic or via a lousdpeaker.

Results to Date

Initial work was carried out on 2 metres FM. Both myself and EI7GMB had never used the units before so this was a steep learning curve. Following successful results on 2 metres FM, SSB was tried with surprisingting results.

Using the Kenwood TM-D710, the following settings were applied. 

FM modulation:    Wide
Mic Sensitivity:     Low
AF gain:              Set as required
 
The mic output pre-set found underneath the unit was reduced in gain slightly. 

Operation on FM

Surprisingly, this process was not painful. In the initial stages, everything broke up and the decode was erratic until levels were correctly set at each end. Everything came together quickly and 2 way communication was possible with comfortable decode. The reproduction of the voice was excellent and both sides achieved good results. A reduction of power between the stations was disappointing. When the carrier was not fully quieting, the noise was too great for the unit to overcome. This resulted in almost incoherent signals. Back up to a fully quieting FM carrier gave good results once again. 

A trip onto the Limerick voice repeater was made. The signal of the repeater was good on both sides of the transmission. The results were excellent from the digital signal transmitted through the Limerick repeater. The received signal was flawless. 

It was concluded that Digital Voice would be excellent over a good path on FM but not at fringe levels where the carrier is noisy. Mobile operation has not yet been tried but will be the next test on the agenda.

The problem with the FM is that the digital carriers are superimposed on the one FM carrier. If the carrier received over a distance is noisy then the whole signal breaks up. Conclusion that digital voice needs a quiet carrier to be received well. Digital is either there or not there. How does D-Star cope or any of the other digital modes? This would be interesting to observe the difference.

Operation on SSB

Connection to the Icom 756 tranceiver was simple and the operation required a few pre-sets to be made, for example, the microphone gain, the power level and above all observation of the ALC meter.

It was necessary to initiate the QSO on SSB and ensure that both tranceivers were on the exact same frequency.  If adjustment has to be made, once in the QSO,  use the RIT. This having been done the mic gain was increased to give about 30% power output and with care to observe the ALC meter. Whilst transmitting the overload LED was observed, ensuring that the signal did not overload the modem. 

The results were impressive and feedback via 2 metres proved that the signals were being decoded properly at each end. Reduction of power levels made no difference with a 5 watt level emanating from my own radio. The path was not too distant so this probably would have made no difference. This may be reduced further to see how the decode occurs with a weaker signal.

The results overall were excellent and the mode did everything it was supposed to do. Voice reproduction was similar in quality to FM and overall via SSB, the quality as good. 

SSB obviously allows this mode to come into its own. The 32 carriers were independent of a carrier as used in the FM mode. The modem captured the signal so much better and indeed as it was designed to do. 

There is no doubt that with care and attention, this mode is far better suited to SSB than for use on FM. 

The AMBE codec is the same as that used in D-Star, C4FM, and the commercial DMR. The unfortunate side, is that each mode employs a different protocol and none are compatible with each other. This results in a divided community. All the modes work perfectly within their own community. 

The fact that there are so many different digital modes will never allow one to take off in any great numbers. In Ireland, the ham community is small and there would have to be some standard to suit all. Would it be D-Star, C4FM or DMR at the end of the day. It is not going to be viable to buy one of each! With the AOR system, it can be interfaced to any radio whether on HF or VHF as it is literally plug and play. The Chinese manufacturers are beginning to take an interest in the digital market and will no doubt gravitate toward the DMR commercial standard as this would appear to be a more viable market with cheaper equipment. Hopefully they will not introduce yet another protocol.

This summarises the work to date, but if there are any other major developments, they will be added to this text.

Further information: It is wise to ensure that RF does not get into the set up. Ferrite beads of clip-ons around the mic lead and also the power lead may help. Experiments performed to date have used an alternative power supply to that used by the transceiver. This could be a small gel cell or a bank of NiMH batteries. The literature suggests that this is a good idea to prevent loops and any distortion being added to the signal.

Power levels: These have been reduced to 0.5W with no problem. Signals decoded well and the comparison to analogue was way different. The Digital still came out with 100% clarity and no background noise compared with the analogue signal  which was noisy.




Tests currently take place on 28.330MHz or 144.600MHz

References:
http://www.areg.org.au/activities-old/hf-digital-voice
http://www.radio-electronics.com/info/rf-technology-design/ofdm/ofdm-basics-tutorial.php

If there are any EI or UK stations using the AOR fast modem please drop me an E-mail as I would like to try a few contacts outside of Galway City.