There was no handbook or instruction sheet published for tuning the aerial coils at a Decca transmitting station. It was picked up by watching others, discussions with design engineering staff and by some amount of experimentation. As a craft, it was passed down by study and memory. Each person had their own way of carrying this out and I’m sure that many fine technical points of the process were skipped over by some or carried to extremes by others. This is my recollection from the mid-1970’s to about 1984.
It must be remembered that on many occasions, this tuning (or-retuning) was carried out under arduous conditions, often at night or in bad weather and the station had to be on the air as quickly as possible. In the case of a routine shutdown, there was a fixed time allocated for the station or chain to be off the air. The coil house sat at the base of a 300 ft tall mast, remote from the main equipment building and was often as not, quite in-hospitable. Some masts were perched on a hill top or bluff and subject to winter gales, driving rain or just constant winds. For chains in hot countries, the coil house was a hot, oppressive concrete box that with very little ventilation and certainly no air-conditioning. Hence the work performed there was often carried out under very uncomfortable strained conditions. With the main equipment building some 1,000 ft feet away, it could be a very lonely place with limited communication with those back at the transmitters. It was not unknown for an engineer to be seen sprinting down the coil house path in the midst of this exercise in order to get some just-remembered piece of wire or tool. Space inside the coil-house was cramped so tools and test equipment had to be carefully placed to avoid damage and flash-over from the exposed coil windings and band spread capacitors. The inside walls of the coil-house were lined with copper sheets so any static charge or stray voltage had an immediate path to ground from your body.Occasionally, in the case of a mast failure, this emergency would require that a temporary mast or even a mobile coil house be brought to the site and set up. In the mid 70’s, Decca made a trailer mounted full-size coil house that could be moved onto the UK stations and allow for mast maintenance. This trailer was fitted with feed cables that could be connected to the existing transmitter feeds and allow for a fast change over if needed. Some of the UK sites had a second antenna mast, (normally a T-array), either permanently set up or else the anchor blocks were in place ready for the array to be set up. The mobile coil house is still in storage as part of the Decca Navigator legacy equipment being held by Trinity House in the UK.
During the life of the Decca System, I believe that the design of coil system only changed once and that was when the systems moved from Mark-V to Mark-X (multi-pulse) mode. The only coil-house variation after that were the double-tuned stations of which there were only a few. Up until the end of the system’s life, the coil house equipment remained essentially unchanged. When correctly installed and tuned, the aerial coil set would give years of continuous operation without any need for servicing or repair. It was only re-adjusted if it was found that site conditions or the mast characteristics had changed over time. Typically, this would be the case if a mast insulator had broken down or become covered in a contaminant such as salt or sand. By checking the aerial current readings on a regular basis it was possible to see any trend developing that would signify that the system was off-tune and radiated power was down. The later development of the Automatic Tuning and Matching system, (ATAM) provided the coil system with a full feedback motor control system to keep the coil set on-tune.
From all the sites that I had been to, one classic coil-house memory was from Gujarat in western India. In 1976, I and another UK engineer were wiring up the aerial coils of the Purple station as part of an upgrade to Mark-X Multi-pulse and in the heat of the mid-afternoon sun, our Indian cook, his hair slicked and white shirt gleaming, was to be seen walking up the long flat dusty path to the coil-house carrying a tray of tea. His years as a steward on merchant freighters were now transposed into this desolate corner of north-west India where he was still keen to perform this quintessentially English custom. It was welcome relief. (As of late 2008, the station is still there and now serves as a DGPS transmitter. As for the tea service, I have no recent news).
For a brand-new site. - Summary of steps performed.
• Load test the transmitters inside the equipment room.
• Load test the transmitters at the coil house feeders.
• Load test the transmitters at the coil frame and adjust matching units.
• Set coils to initial trestle spacing values.
• Use signal generator to set coil resonance point for all five chain frequencies.
• Use signal generator to set neutralizing of each coil for all five chain frequencies.
• Use transmitter feeds on low power to set coil resonance point for all five chain frequencies.
• Increase transmitter power and re-fine tuning point to set mid-range travel for C-coil.
• Adjust neutralizing wires on each coil set to obtain maximum rejection of each un-wanted frequency.
• Run full power individual transmitter test to record line currents and set matching units.
• Cut of excess wire from C-coil output and re-terminate to antenna feed pipe.
• Re-check tuning and neutralizing.
• Commence multi-pulse sequence test with each transmitter on low power.
• Increase transmitter output power and continue multi-pulse sequence test.
• Observe and listen for any corona, arcing or flashover as power is increased.
• Record all values of line current, (DC and thermal ammeter), Matching Unit dials and Matching unit capacitors. If applicable, record values of ATAM control unit and perform de-tune sequence tests.1.1 General
In a Multi-Pulse, Mark 10 coil set there are five coils set on a mahogany support frame, designated Coil A to Coil E. This frame stood about 4ft high and was about 6 ft wide and 15 ft long.
End Coil and closest to lead-out window: Coil A.
Next Coil along: B
Next Coil: C or antenna coil, (so called because this coil had one end bonded to the ground plane and the other connected to the antenna feed tube). The DC and thermal ammeters were connected to the end of this coil winding.
Next Coil and first coil on the movable trestle: D
Next Coil and second coil on the movable trestle: E (closest to matching units and main feeders).Coils D & E are mounted on a trestle that slides on top of the main frame. The trestle runs on glass balls that sit in a groove cut into the trestle and the top of the coil frame. A wheel and crank on the C coil rotor is linked to the trestle by a connecting rod so that as the C coil rotor turns the trestle moves forwards and back. This provides the main tuning adjustment for the antenna. The C-coil rotor could turn 180? allowing for 90 degrees of adjustment either side of the mid-point setting. All the materials used in the coil set and support frame are non-conductive and comprise Perspex sheet, nylon screws and a resin bonded composite that was made into tubes and machined sections. The coil frame and trestle are made of a hardwood (mahogany or teak) and are held together with dowels and pegs.
Each coil is about 48 inches in diameter and 15 inches wide. They are constructed in an open weave pattern with multiple turns of Litz wire on slotted Perspex acrylic spacers. Litz wire is made from numerous strands of very fine copper wire, perhaps 100 or so, all covered in a cotton wrap with a clear vinyl jacket over that. As the Litz wire is wound over the Perspex spacer strips, each layer is stacked on the acrylic spacer below. These spacer strips were held together with thin nylon twine but in the main, the winding was self–supporting. Bracket sections made of composite material extended out from the centre core through which passed the two coil support rods. The winding was actually about four or six strands of Litz wire run continuously in a spiral from the core out. Each end of the winding was terminated in a large flag terminal that was then bolted to one of the band spread capacitors placed on the floor below each coil.
These capacitors were large, oil-filled metal-cased mica units and were perhaps 12” x 18” x 18” tall and each had a single glass insulator post terminal on the top with a large nut and bolt connection.
The C-coil connection to the transmitting aerial used standard 1 inch copper pipe and fittings, suspended about 2ft from the roof of the coil-house by glass insulators. This tube ran straight to the lead-out insulator in the center of the single, coil-house window. From there it continued to the mast base, about 12-15 ft away. A copper tube mast connection was preferred but occasionally a heavy gauge wire was used. Constant winds could cause a fixed tube to resonate and vibrate while temperature extremes could cause a fixed tube to contract and expand excessively. Some sites were fitted with a large, air-spaced adjustable capacitor connected to the aerial lead-out tube and placed near the window. This capacitor took the form of a two large nickel-plate sections, each about 2ft square, held inside a metal frame with clear Perspex sides. These fixed and adjustable plates could presumably be adjusted by the hand wheel and drive shaft on the top. Hence, as the mast conditions changed, this capacitor shunt could be changed to compensate.Most sites had a mast lighting transformer which was placed inside the coil-house, adjacent to the window; (now it’s starting to get crowded in there). In the main, all coil houses were of a similar design and layout, the only differences being the original high-power Dectra sites where larger coils were used to compensate for the increased transmitter power and bigger mast. Local building design or weather conditions may have altered the basic design by adding a larger vestibule or in the case of some, a reinforced roof to resist ice falling from the mast.
1.2 Procedure
Use a dummy load and line current ammeter to load test each transmitter; first at the equipment room and then at the coil house. This checks that both the equipment and feeders are functioning. Line current into the 50 ohm load unit should be about 4 amps. Coil house load test is normally performed at the building’s feeder entry point junction box. For this purpose, Decca made a free-standing, resistive load unit with a connection cable and in-line ammeter.
Set coils on the trestle to initial spacing, using the C-coil end-faces as a datum. (The actual values were often kept on a notepad or handy scrap of paper and have thus been lost to the sands of time). Check that band spread capacitors are correctly terminated to their respective coils and that all wiring is secured. The Litz wire connection tails from each coil were covered with spiral wrap. Dress these wire tails so that they hang down straight and will not move easily or make accidental contact with ground. Later movement of these wire tails can cause the tuning conditions to change. Confirm that antenna spark gap is set and that the mast is not shorted. Set the matching unit control dials to mid-scale.
Use masking tape to place a temporary scale marker on the D & E coil trestle. Take a bent paper clip or similar and fashion a temporary pointer and scale indicator for the C rotor drive. Tape these in place on the C-rotor drive pulley. Set the C-coil rotor to the mid-point of its travel and remove the connecting arm that links the C-coil rotor crank to the D & E coil trestle. Like a steam locomotive, the C-coil rotor had a connecting arm that went from a bushing on the rotor wheel to a plate on the side of the trestle. Hence, as the C-coil rotor turned, the connecting arm crank moved the trestle forwards or back. Threaded adjustment of the connecting rod allowed the rate of movement to be changed as well as the trestle mid-tuning point. By starting the process with the connecting arm removed, it is easy to slide the trestle on its ball-bearings during tuning. Handy wedges were normally supplied to hold the trestle in place during this part of the process.
Connect the C coil antenna cable fitting to the overhead antenna feed tube and tie all the excess Litz wire slack into a folded loop, (a non-conductive loop). Temporarily secure this loop to the antenna feed tube. The C coil feed cable is initially terminated into the feed block fitting without cutting the length. The individual strands of the C-coil Litz wire are soldered into pre-dilled holes in the end of the tinned, solid copper rod that was the feed block fitting. This was in turn joined to the antenna feed pipe with either a solder or compression pipe fitting.
At the connection panel on the end of the trestle, disconnect the antenna feeders that come from the Matching units. These feeders were terminated on the panel studs with ring terminals making them easy to remove and replace. Use a Decca signal Generator, fitted with a chain crystal to inject each of the five frequencies into each of the antenna coil feeders. Decca made a small, rugged signal generator that could be fitted with a station frequency crystal and had two 807 valves (tubes) in the output stage. The unit provided individual pattern frequencies, selected via pushbuttons and had an adjustable output level with meter indication. These signal generators could be used on any chain by simply changing the plug-in crystal unit.
Use an oscilloscope on the C coil to show the resonated output and tune each coil for maximum signal on the scope. Normally, the ‘scope probe would be connected to the aerial feed tube, either above the coil set or near the window connection. If no oscilloscope was available (and in many places, this was the case), an RF voltmeter is used. When first built, Decca stations were normally supplied with some basic test equipment and a large, metal-cased RF voltmeter was normally included. Even after years of use in such inhospitable environments, these units would keep working. The station’s individual frequency was normally used as the reference signal and this was the first resonance point established. When this was found the individual slave frequencies (and 8.2f) would be injected one at a time and their resonance point also found. Each one would be marked on the rotor scale tape.
Start the process by injecting the station frequency into the coils and turn the C-coil rotor through its range to seek a resonance point. Often, a flat topped resonance response will be seen first. Mark this resonance point on the temporary tape scale. Continue on by injecting the other individual frequencies into their respective coupling windings and again turn the C-rotor to seek any resonance point. When the initial resonance spread is marked, it is now time to move the coils on the trestle to bring them to a common resonance point.
Tune the coil set by moving each individual coil either forward or backward along the support frame using the two tuning wands provided while observing the scope or meter for a peak. These tuning wands are 12” tubes with slots cut on one end which fit over the coil support rods. By twisting these tubes, the coil can be rolled on its support rods in either direction along the support frame. As each individual frequency is resonated, always switch back to station frequency and re-check that its resonance point is still in the same place. It was often required that pairs of coils would need to be adjusted to achieve a resonance for one frequency. Typically the A & B coils would affect the 9f and 8f resonance while 5f was most affected by the trestle coils.
The initial intention is to reach a point where the complete coil set resonates to all five injected frequencies without the need to move any coils or the C-rotor. This can be achieved by carefully adjusting the trestle pair D/E as well as the crank setting of the C coil rotor. It may take numerous incremental adjustments to reach a common tuning position. It is important to get to a point where both the C rotor and the D/E trestle are in the mid-point of their range of travel. By adding marks to the temporary masking tape scales, the effect of each small adjustment could be seen on the overall tuning point. Note that the only coil with an adjustable rotor is the antenna coil, the C-coil. If it was found that the overall tuning point left the C-rotor near the end of its range, it would be necessary to remove one or more complete winding turns from the C-coil to bring its rotor back to mid-range. Again, this excess Litz wire would be removed and held in a temporary loop on the feed tube while the main tuning point was re-established. The final cutting of the C-coil winding was not usually performed until the later power tuning stages.
When satisfied that a common tuning point has been reached, it will be necessary to adjust the neutralizing windings of the whole coil set to obtain maximum rejection of all other frequencies when a single feeder is active. This is done by injecting one frequency at a time into the coil set feeder and observing the values picked up on the scope on each of the other four feeders. A solid wire buck-turn section is placed into each of the main coil feeder windings and by reaching in and bending this section slightly the frequency rejection is tuned for a null. This process can be quite exhausting since there is no fixed value to work from; it is simply a matter of repeating the same sequence over and over and each time making a small bend or kink in one of the solid wire sections whilst observing the scope. Quite often, you would find yourself going round in a circle; as two or three frequencies were showing maximum rejection, one or to others would show an increase. At a time like this you tended to stop and re-check the initial tuning points before starting again. At this point the C-coil connecting rod can be replaced and its fittings tightened. Remove the wedges from the D & E trestle and allow the C-coil rotor to slide the trestle on its tracks.
After the signal generator resonance tests were complete, it’s time to connect the transmitter feeders and commence the on-air tuning. At this point, check again that a temporary masking tape scale is in place on both the trestle and C-rotor crank. Use a paper clip or piece of wire to make a pointer for the C-rotor so that the tuning positions can be observed and changes noted. This trick was invaluable when trying to keep track of all the coil adjustments and their overall effect. To perform the on-air tuning it is necessary to have all five signals being fed continuously equipment racks to the transmitters and be able to turn each transmitter on and off as needed. On stations with the 820 type equipment, the only way of feeding all the frequencies was to stop the duty rack at its identification point in the 20 second sequence. The 1880 and later equipments were fitted with a selector switch for this purpose.
The coil-house was fitted with a field telephone that linked it back to the transmitter room and this was the way that we communicated with the room and requested transmitter switching. If, as was sometime the case, we were working alone with local staff and poor language skills, we would set the phase control rack to stop on a multi-pulse indent point, pull out the coil-house “clanger relays” and switch on the transmitters. This meant that all five frequencies were now being continuously fed to the transmitters, which were themselves live but for the control switching signal coming from the “clanger relay” Then, when you required a single frequency from the transmitters you would only need to plug in the appropriate “clanger” at the coil-house and the circuit was made. These relays were large, open contact units, perhaps 6” by 4” with a loop handle and placed below the matching units; ideal for this task. I don’t actually know why they were called “clangers”, perhaps it was due to the metallic clang they made in operation.
With the coils set to their initial tuned position, check that the feeder cables are all connected and commence by switching on the station frequency transmitter at low power, (depending upon equipment type, this could be 100 – 200 watts). Tune for maximum antenna current on both the coil-house DC and thermal ammeter while, if possible observing the transmitter line current. Each coil-house had two large dial ammeters placed on a floor bracket near the center of the support frame. It was thus quite easy to stand beside the coils and adjust the C-coil rotor while looking down at the ammeter needle. Slowly increase power and re-tune, again monitoring line current. The dummy load line current meter can be inserted into the coil frame feeder to monitor line current during this stage. It should peak at about 4 amps when the coils are at resonance. Adjust the individual matching unit to obtain peak antenna current at mid-scale on the matching unit range. The dial has 180 degree of movement and should this peak occur on the high or low side of the mid-point, it would have been necessary to change the value of the fixed capacitance, soldered inside the unit. Such a change was calculated out on site and a selection of capacitors was provided for this purpose.
On those sites which used the 820 valve (tube) equipment, the transmitters comprised multiple plug-in modules of 807 valve PA units and two driver modules. These transmitters fed into individual tank units before connecting to the coil house feeder cables. Both the driver modules and the tank units had indicator dials that allowed the screen current and feeder current to be monitored. Thus we were able to see if the transmitter and feeder loading was correct during the tuning cycle. The valve transmitters could not take being over-driven and would blow fuses easily if pushed into this state. With the next generation of equipment, the type 1880, the transmitters were all solid state and could suffer this type of mis-tune without loss. No individual transmitter dials or meters were fitted on the 1880 transmitters, instead a single test meter box was designed that plugged into each individual transmitter rack and would show its major parameters. The final generation of station equipment was even more advanced and compact, replacing the five individual transmitter cabinets with a single cabinet and plug-in, wide-band transmitter modules.
Repeat the sequence with each of the other frequencies in turn and record their tuning points on the masking tape scale placed on the C-coil rotor. Use a pencil or pen to mark the tape with each frequencies tuning point. Make incremental changes to the individual coil tuning points to close the spread of resonance shown on the tape scale. It will often be possible to close the spread on three or four of the frequencies but one may stay stubbornly out of range. In this case, it may be necessary to re-adjust the D & E separation and/or the C-coil crank position to bring these together.
In some cases, it was found that the antenna capacitance was greater or less than expected and the C coil tuning position was skewed to one end of its adjustment. In this case, it would be required to take one or more turns off the C-coil in order to reduce its inductance and so move the tuning point back towards the mid-range. As can be imagined, this decision was not taken lightly since once cut, the Litz windings could not easily be replaced, especially at a lonely, remote site. Multiple re-checks were made and often confirmation was asked of others before this action was taken.
Once the overall resonance point was set and the tuning adjustment range was acceptable, it was time to re-check the neutralizing. This was done by injecting each of the five frequencies individually while measuring the level of signal at each of the four unused feeders at the end of the coil frame. An RF voltmeter or oscilloscope was used to measure the induced signal and the coil frame neutralizing links twisted and flexed in sequence to seek a null value. The transmitter would be shut off while you reached into the coil set and bent the neutralizing wire slightly. These neutralizing wires were pieces of heavy gauge, tinned solid copper wire, covered in a colored PVC sleeve and connected to screw lugs on the primary winding frame fixed to the face of each coil. All the primary windings were linked with loose wire Litz between the coils, (coils A, B, D & E) which then went down to the coax feeders at each end of the coil frame.
By incremental adjustments, the neutralizing was carried out, all the time checking that the overall tuning point was still in place. It was really quite remarkable how much fine adjustment could be made to a set of heavy, open weave coils, suspended by tubes on a mahogany frame. The tuning wands used to roll the coils forward and back on the frame could deliver a very precise movement in skilled hands. The marks on the temporary tape scale and the wire pointer on the C coil were the only true way of knowing how one adjustment affected the others.
Finally, when all seemed well, it was time to send out a full sequence transmission. This would be started at reduced power and built up slowly to observe for any corona or flashover. Prior to this, the C-coil antenna connection would be re-terminated with any unused slack removed. When this was done, it would be necessary to re-check the overall tuning since a hank of unwanted Litz wire on the C-coil could have some effect on overall resonance. In some cases, it was found necessary to re-adjust tuning and re-check neutralizing.
Typically, we would start with the station frequency and the zone (8.2f) transmitter since this gave a two second burst prior to ident which was handy for the first test run. Other frequencies were added to the pattern cycle with line currents monitored throughout. Finally, the full station cycle was run at full power and with great satisfaction, we realized that another station was “on the air”.
A final set of readings was recorded, these being the antenna current, (DC and thermal) of the individual frequencies at both full-power and half-power. In addition, antenna current was recorded for the multi-pulse ident, again at full and half power. These would be used to show the expected values of the system when correctly aligned with the antenna at full efficiency. It was quite interesting to see the improvement in antenna currents when mast cleaning and re-tuning were performed at operational sites. With many stations being sited near a coastline, salt deposits would build up on the mast insulators. Other sites in desert locations would suffer from sand sticking to the protective grease often applied to insulator fittings. All such conditions required routine cleaning and removal and this event was often combined with a system re-tune while the station was shut down.
If the station was fitted with Automatic Tuning and Matching, (ATAM), it would now be necessary to connect the drive motor linkage to the C coil rotor and check the range of operation of the ATAM. This unit also controlled drive motors on the matching units so that any system de-tune could be compensated for by tuning and matching adjustments. The ATAM had built-in range settings and these could be pushed to first offset and then correct the tuning position. There was very little to adjust in the ATAM unit; it was merely a question of seeing that the overall resonance point of the coils corresponded with the mid-range zero on the meter of the ATAM control unit. The ATAM used the measure the phase shift between voltage and current on each of the transmitter feeders and it was critical that they were all moving in the same direction from the zero phase point.
Typical aerial current, as measured on the thermal ammeter fitted to the C-coil ground tail were as follows: - Green 9f - approx: 8-12 amps, Red 8f - approx: 10-15 amps, Master 6f - approx: 18-24 amps, Purple 5f - approx: 24 amps. These were based on my personal recollection and were likely to change due site conditions such as ground conductivity, mast efficiency and overall state of tune. Any corrections and comments are welcome.
For a site re-tune.
This process was normally carried out during routine mast maintenance, (insulator replacement, guy line replacement, guy line tensioning, band spread capacitor replacement, mast lighting transformer change, etc). Shutting down a Decca station for any short period was not a decision that was taken lightly.
Summary of steps performed. – details generally as outlined above.
• Use transmitter feeds on low power to set coil resonance point for all five chain frequencies.
• Check condition of main antenna feed and re-make any sections if needed.
• Increase transmitter power and re-fine tuning point to set mid-range travel for C-coil.
• Adjust neutralizing wires on each coil set to obtain maximum rejection of each un-wanted frequency.
• Run full power individual transmitter test to record line currents and set matching units.
• Commence multi-pulse sequence test with each transmitter on low power.
• Increase transmitter output power and continue multi-pulse sequence test.
• Observe and listen for any corona, arcing or flashover as power is increased.
• Check antenna spark-gap units are correctly set.
• Check condition of mast light transformer and fittings, (if provided).
• Record antenna current for each individual frequency and at multi-pulse identification.At a site re-tune, the coil set should not have to be moved much at all to achieve resonance. Any dramatic change that may be seen would indicate perhaps a change in the aerial characteristics, ground plane changes or some other error. Occasionally, some of the oil-filled band spread capacitors would degrade due to moisture contamination and these would need to be replaced. At other times the connection to the aerial mast was found to have degraded over time and after replacing or re-soldering pipe joints the system efficiency improved immensely. With a site re-tune, the operation was performed using the station’s transmitters since all the feeders and related systems had already been checked. Typically only small coil adjustments would need to be made to bring the coil set to peak resonance.
Nov 7/08