200W 136 kHz transmitter

By Claudio Pozzi, IK2PII – May 2001

1. Introduction
My first transmission experiments on 136 kHz band where based on a little TX build around surplus components,
particularly the output transformer was wound on a TV EAT transformer. In Italy is impossible to find Philips 3C85
cores.
For 200W transmitter I decided a more professional approach: buy a surplus switching power supply core and design
the output transformer according to ARRL Handbook suggestions.
The success was assured at the first try, so I publish my experience for all people interested in this band.

2. Design requirements
• Power supply: my biggest transformer was 200 W 30 V, coming from an old HI-FI amplifier
• Output transformer core: coming from flea market, light blue colour, marked as A-438281-2-9H9-3, OD = 47
mm, ID = 24 mm, height = 13 mm. No information about µ.
• Output impedance: variable, about 50 ohm.
• Configuration: class D single mosfet
• Mosfet to be used: coming from flea market, with the following characteristics. It’s an obsolete (no data sheet
could be retrieved) and cheap (1 euro) device and is similar to the well-known IRFP450.

• VDSS = 500 V max

• ID = 10 A max
• PD = 200 W max
• RDS(on) = 0,35 ohm (about)

3. Design steps
3.1. Measuring the toroid permeability
Wind the toroid with some turns; use any kind of insulated wire, according with the table 1 and measure the inductance
with a suitable instrument.
Then using the TOROID program (TNX G4FGQ, R. J. Edwards), input the mechanical dimensions, try some
permeability values ad find the µ value that give a computed inductance close to the measured one (if you are a
perfectionist do some form of regression). In my case a µ = 180 resulted in the values reported in the third line of the
table 1.

Table 1
Number of turns (N) 10 15 20 25 30 35
Measured inductance (L, µH) 31 67 115 175 247 351
Computed inductance, given a µ=180 (L, µH) 28,3 63,7 113 177 255 347

3.2. Dimensioning the output stage

Starting with my power supply characteristics I compiled the table 2 with five hypotheses based on my power supply
transformer.

Table 2
Hypothesis 1 2 3 4 5
Vdc (V) 35 35 32 30 33
Idc (A) 4 5 6 6 5
Pin (W) 140 175 192 180 175
Pout (W) 112 140 153 144 132
Zout (ohm) 5,5 4,4 3,3 3,1 4,1
XL (ohm) (= 6 x Zout) 33 26,4 20 18,6 24,6
L (µH) @ 136 kHz 38,6 31 23,4 21,8 28,8

Pout was computed estimating an efficiency of 80%.

Zout was computed using the ARRL handbook simplified formula:

Zout = Vdc2 / 2Pout (1)

The complete formula should consider the voltage drop on the mosfet:
Zout = (Vdc – V DS(on))2 / 2Pout (2)
If you are designing a push-pull stage you can apply the following formula: (TNX G0MRF, David Bowman)
Zout (drain to drain) = 2 (Vdc – V DS(on))2 / Pout (2bis)
The ARRL handbook suggests that RF non-resonant transformer must have an inductive reactance (XL) at least 4 times
the impedance. I decided to use XL= 6 times the impedance.

On the basis of the preceding experience I decided for the column 5 hypothesis, so for primary winding:

Z = 4.1 ohm and L = 28.8 µH

From table 1 the primary winding must have 10 turns (28,3 µH).
The secondary winding should match the antenna impedance, in my case about 50 ohms. Table 3 report the secondary
winding turns, computed with the following formula:

(Npri/Nsec)2 = Zpri/Zsec (3)

Table 3
Number of turns 15 20 25 30 35
Z secondary (ohm) 9,2 16,4 25,6 37 50

3.3. Measures on the running prototype

The transmitter was connected to the antenna and, starting from the first tap on the transformer (15 turns) I searched
what tap was better (maximum antenna current for Idc < 6A). If you have a dummy load you can do a better test. In my
case 30 turns tap resulted in the following table measured values:

• Vdc = 32 V
• Idc = 5.8 A
• Pin = 186 W
• Pout = 150 W (for an efficiency of 80%)
• Zout = 3.3 ohm (see formula 1)
• Npri = 10 turns
• Nsec= 30 turns (the better tap)
• Iant = 1.7 A (measured antenna current)
• Computed antenna resistance (Pout / Iant2) = 52 ohm
• Measured antenna resistance = 50 ohm (about one month later)

The transformer turn ratio (10 / 30) confirms the impedance matching from the 3.3 ohm of the output stage to the 50
ohm of the antenna system (see formula 3).
The running equipment seems closer to the column 3 of the table 2 hypotheses.

4. Circuit description
The TX was built on an unetched printed circuit board, a very good earth plane, using the dead bug technique.

The TX must be excited with a double frequency (272 kHz); this is better for the following motivations:
• The exciter is good also for a coming soon push-pull transmitter
• The flip-flop give a symmetrical square wave output
• During receiving time the flip-flop is unpowered, so no 136 kHz signal is injected into the receiver.

The input stage can handle a 5 to 12 V pp signal (TTL or CMOS, probably, but not tested, also sinusoidal signals). My
Xtal oscillator uses an ex CB Xtal 27.xxx MHz divided by 100.
The TC4426 must have very short wires bypass capacitors, soldered to the ground plane.
The BD136 transistor on a little heat sink handles the CW (or QRSS) manipulation. The circuit come from well-known
136 kHz transmitters.
The capacitor C15 between the TC4426 and the mosfet gate protect the mosfet from destruction in the case of carrier
absence.
The mosfet is mounted on a heat sink coming from an old Pentium 2 CPU, with the little fan running at 12 V. Another
fan in the box keeps the temperature of the transmitter comfortable; the source pin go to the ground plane directly.
The by-pass capacitors C2 e C3 must be of the best quality you can find, rated to 4 times Vdc, I use two 1 µF and one
2,2 µF polyester 250 volt in parallel. Don’t use electrolytic capacitors and keep the terminals as short as possible.
The primary winding of T1 is 1 mm2 Teflon insulated wire, the secondary 1 mm2 enamelled wire.
The Z1 varistor, in my case, is not mounted: no mosfet explosions at this time!
The low pass output filter is copied from other lowfers projects.

5. Possible variations
It’s possible to design the output stage for other power supply. Following are the maximum suggested limits for the
mosfet employed:
• Vdc 90 volt max
• Idc 6 – 7 ampere max
• Pin 250 watt
For lower voltages and higher currents consider a bigger heat sink.

6. References
This project was possible thanks the very good job done by other lowfers and published on the WEB.

http://www.alg.demon.co.uk/radio/136/home.htm Steve Rawlings, GW4ALG

http://www.picks.force9.co.uk Dave Pick, G3YXM
http://www.g0mrf.freeserve.co.uk David Bowman, G0MRF

Components
C1 10µF 35V Electrolytic
C2 2.2µF 250V Polyester
C3 2.2µF 250V Polyester
C4 0.1µF Ceramic
C5 2.2nF 1000V Polyester
C6 10nF 1000V Polyester
C7 0.47µF Ceramic
C8 47µF 35V Electrolytic
C9 4.7nF 1000V Polyester
C10 3.3nF 1000V Ceramic
C11 22nF 1000V Polyester
C12 10nF 1000V Polyester
C13 2.2nF 1000V Polyester
C14 4.7µF 35V Electrolytic
C15 0.47µF 250V Polyester
C16 10µF 35V Electrolytic
C17 47000µF 70V Electrolytic
D1 1N4004
D2 MBR150 Schottky 150V
D3 25 A 200 V Bridge
F1 6.3 A
J1 Key connector
J2 VFO connector
J3 ANTENNA connector
J4 RX antenna connector
K1 SPDT Relay, 12V
L1 54uH Amidon T200-2, 64 turn 1 mm diam. enamelled
L2 54uH Amidon T200-2, 64 turn 1 mm diam. enamelled
Q1 BD136 With heat sink
Q2 2SK1029 With heat sink and fan
R1 2.2k
R2 3.9k
R3 150ohms 2W
R4 33k
R5 100k
R6 680ohms
R7 10ohms 2W
R8 4.7k
R9 10k
R10 100k
R11 10ohms
T1 See text
T2 2 x 30V 200W Power supply transformer
U1a TC4426
U1b TC4426 Not used, ground input
U2a 4013 Not used, ground inputs
U2b 4013
Z1 270Veff MOV or Varistor, not mounted