Technical Description of the HF amps built for VE5RI contest station by VE5FF: The MK-1 and MK-3 (Cyclops)
Check out the Schematics Page!
More Pictures of AMPs
The New 2 Holer!
Casualties of Contesting!
The Mark 1 is a class AB1 high frequency amplifier utilizing the Eimac 4CX1500B power tetrode. It is capable of operating in the HF frequency domain on all amateur bands from 10m to 160m including WARC. Power output is 1200 watts with 50 watts of drive on any band. 3rd and 5th IMD products are -35 db and -40db from carrier.
Ip: 900 ma
Ec1: -60 vdc
Ec2: +325 vdc
Idle current for AB1 operation: 300ma
Plate dissipation: 1500 watts
Circuit Description,RF deck:
The tube is used as a single ended linear amplifier with the screen grid connected directly to the chassis. All supply voltages are referenced to the cathode. This means that the chassis becomes the RF ground reference, with the DC return path floating. RF drive is supplied to the control grid across a 50ohm non-inductive resistor and through a coupling capacitor. The benefit of this configuration is twofold; when the screen grid is operated at RF ground the input gains the greatest isolation from the output. Neutralization is not necessary. Input circuitry is also simplified, there is no requirement for a switched matching network for the control grid. The capacitive reactance (about 80 pf) presented by the tube to the exciter at 28 mhz may be cancelled with an inductive reactance shunted across the 50 ohm terminator when operating on 12/10 meters. This scheme was not incorporated in this particular amplifier because the exciter used is a TS940 with auto-tuner.
The output uses a conventional Pi-network with 10kv vacuum caps for tune/load and a roller inductor/tapped inductor scheme. The roller is used to tune 160-40m, the tapped coil for 20-10m. You may have noticed that the tube is mounted in the horizontal plane. One unique feature of this tube family is that it can be used in any mounting position, while the 8877 or the 3cx1200 can only be used in the vertical plane. A Pi-L network was not considered after initial testing because the harmonic suppression exceeds accepted amateur standards. I felt that the extra complexity wouldnt have offered a significant gain in performance.
The high voltage is fed to the plate via two RF chokes of different values and design (120uh/60uh) placed at right angles to each other. This avoids crazy choke resonances in the amateur bands.
Power Supply and Control:
The HV power supply uses a pair of oil-filled transformers rated at 110/220 v input/2200 v secondary @ 1 amp. A full wave bridge with 8 electrolytic capacitors (450uf/450v) make up the rest of the supply. Voltage output no load is 3000v. Nothing exotic about the design, strictly radio handbook modeling here.
The screen supply consists of a 400v center tapped transformer with a variable transformer on the primary. A full wave bridge is used with capacitor input and a string of HV zener (50 watt) diodes shunt fed. RF bypassing is a must, I discovered that zeners dont like stray RF! This supply will be changed to active shunt regulator ( HV mosfet).
The Control Grid bias supply uses a 90volt transformer with full wave bridge. Output is -20 to-100 vdc. This allows you to adjust the bias for Class C to operate FSK and CW. Because the amp is run in AB-1 for SSB, this supply doesnt have to be rated for high current by virtue of the fact that no grid current is drawn in this mode. The 4CX1500B is rated for AB2 operation, but typically not operated there. Note that the 4CX1000 should not be operated in AB2.
Filament voltage is 6vac/9amps.The cathode is indirectly heated. One downside in using the 1000A or1500B is that the cathode requires warm up of about 4 minutes before you can operate. If you insist on ignoring this fact, you will end up destroying the tube prematurely (shorted grid to cathode). The cathode timer period must expire before screen and plate voltages can be applied to the tube.
The control loop monitors screen current (sometimes negative) in order to protect the tube from an overload condition. An overload can be caused by failure of the anode supply, antenna loading or failure or overdriving the control grid. When an O/L is sensed, the screen and anode supplies are disabled by removing voltage to the mains primary contactor coil. A status indicator led is enabled which tells the operator that there is a problem.
You may ask why use a tetrode when a triode seems to be less complex to use. There were a few considerations that helped me form a design philosophy.
The prime goal was to use a device that offered exceptional operating characteristics in linear amplifier service with a minimum of drive power from the exciter and be able to deliver 15 db of gain. The operating voltages had to be low in order to keep the cost of the power supply within reason. The tube had to be readily available from the surplus market as a pull (replacement cost control). So performance and economy became the design criteria. Of course there is added complexity with the addition of bias, screen and control supply/circuitry, but the time spent in designing and building these was not significant.
Another FAQ is why not use tubes from this family in triode grounded grid configuration. There is one good reason for this; because the 4CX1000A and 4CX1500B were specifically designed for linear amplifier (AB1) service this requirement influenced the design of the grid structures (low electron intercept).The control grid dissipation of the 1500B is rated at 1 watt and the screen is rated at 12 watts. When you connect the grids together and then operate in grounded grid the control grid will draw almost 75% of the total grid current ( do you remember how much grid current is drawn by a pair of 3-500s?). You WILL destroy the tube in this mode. The 4-400 and 4-1000 can be used as triodes (screen and control grids strapped together) in AB2. Because they were originally designed for class C operation, their grid structures use a different design scheme and the dissipation ratings are greater. As well, in order to get any power out of these tubes, you should be operating the plate at 4kv with 100 watts of cathode drive.
The design of this amp incorporates the Eimac 4CX3000A /8169 power tetrode. This tube is designed specifically for linear amplifier service. Essentially the same circuit design is used as the MK-1, with certain exceptions. The control grid input requires a 1:4 un/un transformer terminated with a 200ohm resistor. This is used in order to get RF voltage step-up. RF input voltage has to overcome the dc bias voltage (-160 volts) in order to drive the control grid. The Screen power supply uses an active shunt fed regulator (HV mosfet controlled with an op-amp used as a voltage comparator, courtesy G3SEK). Normally the screen potential is 300 volts . This amp was designed for 100% duty cycle-FSK/CW/SSB modes.
There are some differences between the 4CX1500B and the 4CX3000A. Mounting is vertical only and the filament is instant on. The filament material in the 4CX3000A is thoriated tungsten while the cathode in the 4CX1500B is oxide coated. One interesting point to consider is that large power tubes using thoriated tungsten filaments can be rebuilt for about 60% of new cost.
4CX3000A operating parameters:
Ep 5800 vdc
Ip 800 mA
Ec1-160 vdc (adjustable to -300).
Ec2 300 vdc
Filament: 9 v @ 40 A
Idle current in AB-1: 500 mA
Plate dissipation: 3500 watts
Control Grid Dissipation: 50 Watts
Screen Grid Dissipation: 175 Watts
IMD: -35/-42 db
RF power out with 50 watts drive: 2000 watts.
Gerard Hnatiw VE5FF