Observations and thoughts on 26 MHz test transmissions 25 November 2007 Glen English, Pacific Media Technologies Pty Ltd, Cooma, NSW Trevor Harwood, TJH Systems Pty Ltd, Mt Barker, South Australia. email :glen@pacificmedia.com.au Contents : Introduction Expected Performance Vehicle Test Setup Receiver Deficiencies Results and Conclusions Suggested technical operating conditions Further Receiver thoughts. Introduction A 26 MHz DRM transmitter has been operated by TJH Systems Pty Ltd from time to time under a test licence granted by the Australian Communications and Media Authority. Rather than use a theoretical approach to the trial, these series of tests highlight a practical approach, to show real world operation. It is hoped that others can use the experience gained. The transmitter is designed by TJH Systems Pty Ltd, and is operated at 14 watts RMS. Transmitter in Rack same as at SMPTE show in Sydney 2007.
The antenna is a quarter-wave ground plane at 9 meters above ground, approx 3 meters above the factory roof. There are numerous factory buildings that surround the transmitter site and it is by no means a ideal site. The antenna view is obstructed by about half of its azimuth. For vehicle testing : The DRM mode used was Mode A and 16 QAM, protection level 1, providing 18400bps in the 10kHz RF bandwidth. The service area is hilly with numerous valleys, farmland and built up small suburban zones. Propagation was as expected for this band, that is, similar to 27 meg CB. Good coverage in hilly terrain but still of VHF character.
Approximated coverage in yellow areas by TJH Systems Pty Ltd using EDX SHDMAP TX at 35 03 18.31 S; 138 50 47.03 E Line of site areas are in green.
Expected performance : The conversion from the received power using an isotropic antenna to dbuv/m, at 26 MHz is : dbuv/m = dbm + 97 From this point onward, dbuv/m will be abbreviated to dbu. The thermal noise in the received bandwidth is -134dBm @30 MHz : (CCIR322) The mean atmospheric and background noise level is 20dB above the thermal noise (ITU R.372-6) man made noise levels are specified as +7dBu for residential. Considering the electronics installed in modern homes, we believe the residential man made noise level figure to be out of date. Assuming 16QAM is used and 10kHz bandwidth, the minimum usable signal will require 12dB C/N. Therefore the MUS (minimum usable signal ), with an ideal outdoor location is approx 7dBu + 12 = +21dBu In bench testing, using the external antenna connection, the receiver operated to less than -115dBm., we believe the ultimate sensitivity to be sufficient for the tests. The use of the internal antenna suffers due to the self-generated interference. Vehicle Test Setup The receiver used for testing was a Morphy Richards that uses the Radioscape Tuner, powered from a 12V isolated battery.. The latest software upgrade that improves the Radioscape tuner s RF IMD performance by some 50dB was installed. An 27MHz helically loaded CB whip 100cm (30 degrees) was modified and tuned to 25.795 MHz and attached to a magbase, which was attached to a late model utility vehicle. The antenna base height was approx 150cm above ground level. The estimated as-installed vehicle antenna efficiency is -10dBi. The directivity (pattern distortion) due to the irregular shape of the vehicle was not measured. In order to isolate the antenna coax from the rest of the system, approx. 5 turns of coax was wound around a high permeability moderate loss 4 toroid, and two low loss high perm coax clamp-on ferrite chokes were used. There were placed about 1 m form the magbase. Due to the super small ground plane, and also the lack of a direct connection to ground at the magnetic base contact (a Magnetic base relies on capacitive coupling to earth or ground plane ), the current flowing in the coaxial cable is high.
The photo shows the toroid and ferrite chokes on the roof for clarity. They are 'Gaffer' taped down to the side of the vehicle. An external antenna socket was fitted to the radio. Deficiencies in receiver performance were the big surprise of this test. The receiver can display MER and received signal configuration. Receiver deficiencies found/known were : a) Non existent power supply noise rejection. These radios pick up noise from their 9V DC power lead and seemingly inject it into the front end. We have always needed to wind the power cable through a few turns of a toroid in order to prevent power line noise, and further unwanted RF pickup (interference) via the power lead. In our vehicle test setup the radio was found to have very poor sensitivity, and this was traced to low level computer noise inside the vehicle getting into the tuner via the power lead.
Suppression beads in series with power lead. b) Antenna. We fitted a BNC socket with a direct coax connection to the tuner. The tuner impedance was measured at approx 130-j220 with the receiver tuned to 25.795 MHz. It was assumed that the DUT was a conventional input stage/device and not an A to D Converter as the sampling characteristic would have made measuring impedance difficult. After construction of a matching network, the receivers performance was significantly improved when used with the 50 ohm antenna. Additionally, the low pass nature of the network gave the receiver more VHF and UHF frequency rejection. c) Steady state performance. Signal level dependent, when connected to the transmitter directly, the observed MER tended to wander from 16 to 29 db MER in a quasi- cyclical manner. Road Test Receiver Issues: Numerous drops in strong signal areas did occur when the signal strength changed abruptly. The AGC loop needs work! If the vehicle was rolled slowly through the area where the dropout was observed, no dropout occurred. This leads us to think the signal quality was OK and the receiver lost the ability to track the signal level change. In defense of the radio, it is a 'portable' not a 'mobile' radio. However, we felt the performance could have been better. The radio had difficulty handling strong signal levels. Signals beyond approx -30dBm reduced the reported MER.
Results and conclusions The road test dropout locations were used to tune the EDX propagation prediction model. The vehicle setup was useful down to a level of approximately 5 dbu. Power line interference played a significant part in reducing the coverage. Built up areas and shopping areas reduced the coverage marginally compared with deep gullies and noisy power lines. When driving slowly through a town centre, with reasonably expected constant field strength, the signal to noise reduced approximately 3 to 6 db. The transmitter site covered well in moderately hilly terrain, however its performance in deep gullies and pockets was poor and a better transmitter site would be needed for a proper service. The performance was very similar to the difference between TV channel 10 and TV channel 1. Due to the excellent propagation features of this band, and in the interest of re-use, it is suggested that efficient sites be used. This equates to those transmitter sites that have a good general view of the built up areas and valleys but not views of the entire countryside. 26 MHz is not the band to put transmitters on mountain tops with commanding views, that is, sites with positive horizons are more suitable.. Doing this will stifle channel reuse and exacerbate the expected Sporadic-E skywave problems. Sites that have a poor siting will require greater power and generate greater skywave spillover, which in turn causes a fringe coverage problem for another station. Deep nulls (flat fades) were observed where the entire channel experienced a fade. The channels in a suburban environment are flat fading, not frequency selective channels. It is suggested that polarization diversity be used to reduce the depth of these fades and reduce the variability caused by user antenna orientation. Compact cross-polarized and circular polarized antennas are quite practical. This needs to be tested. Delay Diversity (Two or more transmitters operating a prescribed distance apart and timing offset) is also suggested and the forthcoming transmitter design will offer this. This changes the channel from a flat fading channel to a frequency selective channel. The downside is the moderately large antenna spacings required to make this work efficiently. It is the authors opinion that a combination of delay and polarization diversity be used to mitigate the flat fades. Indoor levels in the test house were found to be approx 12dB below the outdoor vehicle level, and approx 6dB noisier than immediately outside the test house. The internal radio antenna was found not to work inside the house, despite a FS of 40dBu outside. This was due to poor sensitivity when using the internal antenna, and the high amount of self generated interference by the receivers circuitry. Using the vehicle antenna inside yielded marginal results as the antenna was without its ground plane and the mismatch losses were very high.
Precisely determining building penetration attenuation will require more work. In particular, steel frames houses and apartments need attention. If the radio design was improved, and some form of diversity transmission was used, I expect 40dBu to be a good start for the planned minimum FS using 16QAM/ 10kHz transmission. Levels would be approx 6 to 8 db greater for 64QAM planning. A first solution to the required FS would be to increase power. BUT this is a dangerous idea, because skywave propagation will be a limiting factor, especially for 64QAM transmissions that require high levels of C/I ratio. Propagation via Es mode can be close to free space, or even less than free space loss. Increasing TX site height dramatically is also a poor solution because of the hard to control spillover causing co-channel interference, and reducing the number of channels and ease of band planning. Therefore, coverage areas should be limited to that which can be covered by 'moderate sites' with 'moderate power' There are also good practical reasons to keep the power moderate. It is expected that 26 MHz transmitters will go inside towns or on 2 way radio sites, rather than broadcast sites, and high powers in these locations are not welcome. Small scale SFN's can also be used. Most 26 MHz DRM transmitters have SFN capability and due to the low bit rate of the DRM system, SFN 's can be fed from ADSL and other telco style lines. Suggested technical operating conditions : Planned FS Urban: (Longley Rice), 10m, 50/50/90 = 40dBu (16QAM), 46dBu (64QAM) Planned FS Rural: (Longley Rice), 10m, 50/50/90 = 34dBu (16QAM). ERPd : 50W to 100W RMS, site dependent (ERP per polarization plane) Maximum TX antenna height above local terrain : 30 to 100m Service Area Radii : Typical 10km, 20km max Channelling : 16QAM services require less C/I and less FS, therefore I believe, at first sight, 16QAM channels should be grouped in their own band segment. Far more 16QAM channels could be packed into an area than 64QAM channels. The greater FS required by 64QAM channels, and hence greater C/I required dictates that fewer channels will be able to be used in a given area and therefore they should be in their own band segment as to avoid reducing the available 16QAM channels. An alternative is to hold the same nominal power for a 64QAM as a 16QAM channel and operate the service at reduced range (coverage). Use of diversity transmission, especially polarization diversity will substantially increase the reliability for a given power output, and hence improves reuse and band capacity.
Further receiver comments : Internal radio antenna design : A short 50cm long whip at 26 MHz appears as a small resistor in series with a small capacitor, approx 1 ohm in series with 1.5pF Therefore, to avoid enormous losses of sensitivity, great care has to be given to stray and input capacitance. Ideally the radio input would be a high impedance such as a FET gate with the FET operating in a high linearity bias mode. The receiver should not overload at non-frequencies (that is, those out of band) of 1V/Meter. In the Morphy Richards, the use of a coax cable between the radio input and the antenna is suboptimal as the 15 cm of coaxial cable, some 15pF cable capacitance shunts the input signal (a capacitive divider in formed). A prototype hiz preamp board will be constructed for the MR radios for future indoor tests Bandwidth : 20kHz was not tested as it is expected that receiver support will be much better at least in the short term if the 20kHz support requirement is dropped. I have spoken with numerous receiver manufacturers and their comment is they can make an inexpensive 10kHz BW radio for LW/MF/SW DRM using off the shelf IF bits, however, the switchable 20kHz bandwidth requirement puts it in the 'new' or 'hard' basket. Sure, modern digital IF radios are flexible in this regard, but such a radio is more expensive than an traditional superhet as this time. This item needs discussion.