Decca 5501 Transmitter Conversion.                                                                   7/July/2000

Here are some details of the conversion of the Decca transmitter from 127kHz ( 9f designator) to the 136kHz amateur band. My Racal Decca TX is now useable. However a 3 hour period of testing at 600W out has shown that some further refinements are necessary.

The standard transmitter comprises 3 x  400 Watt modules which are combined to give 1200 Watts.
The suggested DC supply is 27V @ 700mA for the drive and 67.5V at 1500 Watts for the PA.

The 9f tank circuits on my transmitter can be retuned to 137k by adding 200nF.  I used 2 x 100nF 1kV capacitors in parallel which can be mounted neatly in series with 8 similar capacitors on the tank circuit capacitor board. This re-tuning process is only applicable to the 9f (127kHz) units. Other types will need turns removed from the Litz wound tank coils.

Designators 5f 6f are low band while 8f  8.2f and 9f  are high band.
The Tank inductors have a few extra turns of wire which are used to trim the circuit to resonance. The extra turns are either just bundled up with a tywrap or are wound either in-phase or in antiphase with the fixed winding.
The loaded Q of the circuits is about 7.

Protection Circuits
The 5501 has a number of monitoring circuits which are designed to alert the operators at Decca when a unit goes off-air. In common with other professional transmitter there are no auto shut down circuits. Instead the 5501 is designed to operate with a "guard circuit" that protects the unit in the event of a short or open circuit. This circuit shunts power back to the PSU in the event of a fault. It appears to work well but the thought of several hundred Watts with no-where to go made me decide to add some conventional protection. 

If you're brave you can just construct 2 DC supplies, retune the PA's and you're on-air.
I'm not brave, so I added the following:

1)  VSWR protection.  More accurately, an adjustable reflected power trip.
2) An internal low voltage 24V supply for the drive  (78S24) - 24V works OK
3) A 12V supply for my added CMOS circuits          (7812)
4) An input buffer for an external logic level drive    (4013 divide by two from my DDS and 4426 FET driver chip)
5) A keying circuit - PNP BD136 transistor to the 4426 supply
6) Front panel BNC sockets for monitoring Voltage and Current phase (tuning aid)
7) Transmit receive relay
8) Transmit / Receive / Net  switch
9) A simple 70R in/out Low Pass Filter. ( The 5501 has no filtering other than the tank circuits)

The list is quite a long one, but most of the required electronics was already part of the class D TX design on my web site.  I took one of the old prototype PCBs and cut out the necessary parts of the board.  Not a neat solution, but quick and hopefully complete. Details of the class D circuit is at:
 www.g0mrf.freeserve.co.uk/300w.htm

I've mounted a  TX / RX switch on the front panel. It replaces the original Decca 'Alarm' reset button. I've wired it and the adjacent LED with 4 core cable and routed this to the electronics on the new rear panel. There's plenty of room to run the multicore cable via the grommets which separate each compartment within the transmitter. The switch is a 2 pole single throw with a center off position. The center off  provides a netting function on receive.
.
The adjacent  front panel LED has been replaced with an ultrabright type which is driven from the CMOS via a 12k resistor. No drilling is necessary as Decca have used a standard sized holder. This LED indicates when the unit has shut down due to a VSWR fault. The shutdown circuitry then resets automatically (G3YXM monostable timer circuit).

The extra electronic circuits are accommodated on a 3mm thick rear panel. This also adds a little extra rigidity to the TX.  The TX / RX relay is arranged to short the receive line during transmit periods.

I think I'm 90% there now.
I ran the transmitter at 600Watts out for three hours into the long wire antenna at the Whitton Club. My only criticism is that the output filter is not efficient enough and was running at apx 80 -90 degrees C. - This would probably fail at 1200 Watts.  - All ideas welcome! All other components ran without problem.

Low Pass Filter Details  ( Not recommended due to high insertion loss)-Now solved see bottom of page

Design impedance  70R                              Type.    Single PI   LPF
Frequency             196kHz
Capacitor (2)         11.6nF   ( 4.7 + 4.7 +2.2 )
Inductor                 58uH   - 36 turns of 1.5mm dia wire on Amidon T225-3  Iron dust.
The inductor is held apx 1/8 inch above the aluminium panel by two pieces of fibreglass PCB material.

Above:  Rear panel prior to wiring.  78S24 and 7812 regulators are mounted above IEC connector.

Above.    Close up of the filter components and directional coupler.

Above.   Tx on the bench.   2 BNC connectors for Voltage and Current phase measurement.
TX/RX/Net switch and VSWR shutdown LED fit neatly into original Decca panel.

Rear Panel.  Connectors from top left:
24V supply output (4 pin XLR) to drive an external DDS source.  Antenna (N-type).  RX out (BNC)  RF input to divider and buffer amp. (BNC)
Lower:  IEC filtered  mains connector and on far right  Keying input  (6.25mm mono Jack)
This picture also shows the two 100nF added on the left side of each bank of capacitors (slightly smaller than the originals)

UPDATE:   12/7/2000
Changed the overheating T225-3 core in the low pass filter.
I decided to use a -2 material but did not have a 225 size available. Instead I stacked two T200-2 cores.
I expected to have over 60 turns for the required 58uH but surprisingly the stacked cores only needed 44 turns. I wish I used a wire with a larger diameter now, but the 1mm dia appears to work well with minimal heating at 1000 Watts.

**** Since building all the protection circuits Jim, M0BMU, has tried full power short and open circuits. The guard circuit worked well, so maybe all you need is sine wave drive at 7-10V P-P and a big PSU after all!****

UPDATE 12/Sept/2000
Used the decca with a large PSU in QRSS mode at about 850W and then 325W output.. Signals received in Canada as part of first transatlantic crossband QSO to 20m. Also received 136kHz signals from Jack VE1ZZ. However no QRSS at VE1ZZ so did not receive call sign, just letter S which was confirmed by phone.