-> Projects -> GS23B Amplifier

GS23B Amplifier

For EME and contest operation the need arose to get on the 70cm band power levels around 1 KW. Design and construction should be such that no major de-tuning exists over a contest timeframe and more or less unattended operation is possible. These design criteria where possible by using a GS23B tetrode tube in a cavity design.

gs23pa_400_300 (42K)

Mechanical design

A complete Russian originated cavity was used. This is proven to work very well on the 70cm band (432 MHz) with only some modification to the input circuit and rework on the anode RF capacitor. Modifications designed and done by Nanko pa0v, thanks!.

cavity (27K)

All manual controls (input matching, anode tuning and loading) are available on the front of the cavity. Although a very good efficiency can be obtained a lot of heat needs to be drained off. This is done via the external attached radial air-blower. Blower is attached at the anode end of the cavity and will suck air out off the cavity. In this way ensuring that the cathode/grid components are cooled appropriate and circumventing the need to construct some airtight box around the complete cavity. Blowing air in the anode component is not a good idea in this case, because hot air will enter the cathode/grid component part and would not provide enough cooling.

cavity_and_blower (26K)

Together with g3sek's tetrode contol boards supplying the screen, grid and heater the complete amplifier is housed in a 50 * 30 * 20 cm case. A highvoltage powersupply supplements the amplifier.

Input match modification:

To improve the input matching a small serie inductance and trimmer capacitor was installed together with an UHF N-connector.

Input Coupling

Tubular trimmer capacitor is accessible "through" the serie inductor which consists of an aprox. 30 mm long copper tube 6 mm outside diameter (to be verified....).

Anode ring capacitor modification:

As described below in the electrical design chapter the anode should be very well RF decoupled and will carry of course the DC anode voltage. In the original cavity design this was done by a thin sheet of mica. However, this material very vulnerable for mechanical stress and was already damaged. Therefore, the mica was replaced by a thin sheet of kapton. Kapton is, although not as sturdy as teflon, very much suited for this job. It provides a neat Er (giving a nice capacitance), it can also be very well glued and proving a solid construction of the 2 copper parts.

to be inserted anode ring picture

Inner ring will contact by the fingerstock the anode and the outerring is firmly pushed against the cavity wall. A piece of woven metal is inserted to decrease the RF contact resistance even futher.

Electrical design:

RF design of the tetrode amplifier is based upon grounded anode design. This as opposite of the more common grounded grid designs for tetrode amplifiers, but will offer equal performance and characteristics.

Electrical Diagram GS23b Cavity (12K)

Anode is RF decoupled by the anode(ring) capacitor (measured to be around 1nF) and highvoltage DC (3000V) is directly applied. No DC choke is needed due to the well decoupled anode. Output tuning is done by a movable short and loading by a small flapper attached to the output connector. Input tuning via the tubular capacitor and moveable short.

Heater voltage is supplied via a separate transformer with soft-start circuitry to prevent a too large heater inrush current which would reduce the tube's lifetime. To reckon in voltage spread of this transformer and to set as accurate as possible the heater voltage (should be 12VAC as measured at the tube connections) a number of small valued resistors.

to be inserted heater circuitry diagram

Using g3sek's tetrode boards the screen and grid voltages are supplied, including the trip circuitry. Please refer to the information on Ian's website for detailed information. Control grid standing current in the G1 powersupply has been set a bit higher than suggested (20mA in stead of 10mA) to be on the save side. Also the resistor setting the zenerdiode current (R108, 10K, 1W) was replaced by a 1Kohm 1/4W resistor. Original value is too small to provide enough diodecurrent (100uA). Increasing the zenerdiode current to some mA will improve the resulting control grid voltage stability.

Electrical parameters